More than 100 companies, including Microsoft, Amazon, Visa and Anthropic, put out a joint warning that AI powered hacking is about to get a lot more common. When the industry building the thing is the one waving the flag, that’s worth paying attention to.

Anthropic showed AI agents running actual lab equipment, microscopes, robotic arms, liquid handlers, the whole setup. In one drug discovery test, the AI adjusted the experiment as it went. Another ran three times faster with AI controlling the equipment directly. The lab itself is becoming programmable. Experiments can run overnight, respond to what they find and keep working while the scientists go home and sleep. Yay! This could be a game changer for medicine. Whether that’s thrilling or terrifying depends on which side of the microscope you’re standing on and how well padded your bank account is.

A federal judge also ruled that the Pentagon broke the law when it blacklisted Anthropic for refusing to let Claude be used for mass surveillance and fully autonomous weapons. The Pentagon broke the law. I know. Shocking.

Then there’s the electric bill. Data centers are now the biggest driver of rising electricity demand in the country, and regular households are paying for it through higher power bills, sometimes dramatically higher, while states hand these companies massive tax breaks to move in. So we subsidize the data centers, then pay higher electric bills because they’re there. What a fucking deal.

Finally, for my haters, I brought you snacks.

Go watch Ed Zitron sit down with Steven Bartlett on Diary of a CEO and call the entire AI industry a con to his face. OpenAI lost almost $21 billion last year. Let it nourish and fortify you through Sunday.

Nobody knows where all this goes, including me. Maybe we’re watching one of the most consequential technologies in history, or maybe it’s all a big smelly turd. Probably both.

Wake me when it’s over.

And finally, one word for this week’s cosmic unfairness. Dolly.

Melissa Tart is the founder of Farther North Studio, where she teaches experienced women to use AI with confidence so they can turn decades of expertise into their next act.

Learn more or schedule a Discovery Session at stan.store/MTart

Subscribe at substack.com/@melissatart

I pulled out the six points I think matter most, particularly for anyone trying to figure out what AI means for their career or business.

1. Jobs are going to shift quickly. Gates expects both white collar and blue collar work to be affected within the next decade. Previous workforce transformations unfolded over generations. We may not have that luxury this time.

2. The risk extends far beyond job loss. AI is putting increasingly powerful capabilities into the hands of people with very little training. That changes the scale of what individuals can do, including the damage they can cause.

3. Critical thinking is becoming one of the most important skills we have. Deepfakes, tailored misinformation, and increasingly convincing AI generated content will make determining what’s real, manipulated, or completely fabricated considerably harder.

4. Our institutions weren’t built for this. AI crosses employment, security, markets, education, children, and countless other areas at once. The agencies responsible for those individual pieces aren’t necessarily looking at the whole system.

5. Some jobs may need to remain human on purpose. Gates calls these “Human Reserved” jobs. Caregiving is an obvious example, but the larger question is worth considering. Where does human presence, trust, empathy, and judgment have value that shouldn’t simply be automated away?

6. We may need to rethink how we tax automation. Gates raises the possibility of taxing AI tokens or robots. When employing a human means paying payroll taxes while investing in automation can receive favorable tax treatment, we’ve created an economic incentive that deserves much closer scrutiny.

There’s plenty here to debate, including Gates’ own position in the technology industry and the enormous financial interests surrounding AI.

The larger warning still deserves attention.

The technology is moving fast. The systems around it, and the people whose lives will be affected by it, aren’t moving at the same speed. And some are in complete denial.

If you get an opportunity, read the Newser article. I think it’s important.

McKinsey reports that 80% of workers using AI are more productive, yet company ROI is sitting at 37%.

Why? And more importantly, how does this impact you?

Companies can spend millions on AI and still get mediocre results if people don’t know how to use it effectively. The technology is moving faster than the skills required to get real value from it.

For those of us working, running businesses, consulting, freelancing, or thinking about what the next few years might look like, that’s an opportunity.

Start with what you already know. Take a task you do regularly and use AI to improve it. Pay attention to what saves time and what creates more work. Question the output. Verify it. Use your own experience to recognize what AI misses.

Then move to the next task.

You don’t need to learn everything about AI. Nobody can keep up with everything happening in this field right now. You do need enough experience using it that you understand what it can do for you and where your own judgment remains essential.

The workplace is changing while we’re standing in it.

What are you doing to prepare?

And of course one of the ugliest responses came from someone I knew. She told me she was “about a millimeter away” from declaring me “an enemy of humanity.” Then came “FUCK YOU,” the prediction that my position was going to “kill us all,” and a swipe at my AI-assisted art for good measure.

I get attacked over AI constantly, and I’m tired. I’m tired of being treated as though learning a technology somehow makes me responsible for every billionaire, data center, layoff, copyright fight, environmental consequence and reckless corporate decision attached to it. I don’t pull any of those levers. I’ve spent three years trying to understand what’s happening and the entire past year writing long, detailed articles about the parts that concern me from damn near every angle.

I’ve written about the concentration of power and the people accumulating it, jobs, privacy, surveillance, environmental costs, regulation, education and the spectacular lack of leadership while this technology moves faster than the institutions responsible for protecting us. Last week I devoted an entire series to children and AI, what this could mean for developing minds, what schools need to confront and what parents need to understand. I’ve started an AI curriculum in a school. I’ve never told anyone there’s nothing to fear. Some of this scares the hell out of me.

My conclusion hasn’t changed. I’d rather understand what’s coming than be blindsided by it.

My interest in AI started three years ago because of cancer. My sister died from it. My mother’s sister died of cancer at 34. My father had melanoma. I’ve lost countless friends to cancer and watched families bury their babies. When cancer carves its way through your family and takes the people you love, you look for hope wherever you can find it. So when I began hearing that AI might accelerate medical research, improve diagnostics and help scientists find treatments faster, I paid attention. That’s what brought me to AI three years ago, and that’s why I’m still here. Cancer.

Since then, I’ve spent hundreds of hours and thousands of dollars learning. I’ve shared an enormous amount of what I know for free because I believe being prepared matters, especially now. We have a government and leadership structure that’s nowhere near ready for what’s happening. I don’t believe anyone is coming to neatly protect my family, my friends, the women I teach or me from every consequence of this technology. Nobody’s doing it for us. We’re on our own.

Yesterday, before that comment, I read that Merck and Moderna had reported encouraging Phase III results for an individualized mRNA cancer therapy used with Keytruda after surgery in patients with high-risk melanoma. My father survived melanoma years ago. Of course I cared.

I also know I’m forever chasing sparkly things, particularly when the subject is cancer. So before I let myself get excited, I sent the news to a friend who is a cancer researcher. This is his life’s work. Here’s what I sent him.

“I don’t know if you’ve seen this, but Moderna and Merck just reported encouraging Phase III results for a personalized mRNA cancer vaccine plus Keytruda in high-risk melanoma after surgery. Every advancement in cancer treatment feels hopeful to me. It’s a good day.”

He had already been reading about it. He told me I should be hopeful. He said mRNA vaccines, personalized or not, alone or combined with immune modulators like Keytruda, are one really strong way forward in the fight against many advanced cancers. Then he raised cost, insurance coverage and who gets access.

I thanked him for the feedback. I’d wanted an expert’s perspective before allowing myself to get too excited. I told him my dad had survived melanoma and that who we vote for in November matters on so many levels because no one should be denied treatment for lack of money or insurance.

That was our conversation. Cancer, science, hope, cost and access. AI never came up because I had no idea AI was involved.

Today I read The Neuron’s coverage and learned that AI is part of the design process for this individualized treatment. Moderna says its algorithms analyze sequencing data from a patient’s tumor and blood, evaluate mutations and predict up to 34 neoantigens most likely to trigger an immune response. Those targets are then encoded into an individualized mRNA therapy made for that patient.

Three years after cancer first made me pay attention to AI, I learned that AI was involved in a promising melanoma treatment I’d been talking to a cancer researcher about the day before.

That doesn’t make every use of AI good. Holy hell, I’m exhausted by people who complain, nitpick and act as though acknowledging one useful application means I’ve endorsed the entire industry. Spare me the lecture. A technology can help identify targets in a tumor while corporations use other applications of it in ways that are exploitative, dangerous or grotesque. Medical progress doesn’t erase job loss, surveillance, environmental costs or concentrated corporate power. Those problems don’t erase medical progress either. I’m capable of holding more than one thought in my head at a time. Apparently that’s rare.

The person behind that comment was Mandy, a local résumé writer I’d hired several times. Two months ago, we were talking privately about AI and discussing a possible collaboration. She told me, “I do believe in AI for the important work. That’s probably where we align.” When I told her I’d built a curriculum because I believe women are going to be hit first and hardest, she wanted to talk. I had ideas I thought could help her and her clients, including free access to my curriculum and a tool designed to help people improve their chances of getting hired.

We scheduled a time to talk. She needed to reschedule, which wasn’t a problem. I gave her my number and told her to call anytime. She said she’d text first. I never heard from her.

I’d seen her outrage about AI before and never attacked her, mocked her or said a cross word to her. I never disrespected her right to be concerned or to express it. Her work could be affected. Mine could be affected. All of us could be affected. I have plenty of concerns of my own, and some are the same ones I’ve spent the past year writing about.

I also know that fear gets worse when we don’t understand what’s happening. A little more knowledge gives us a better chance of preparing, and preparation gives us something fear doesn’t. Agency.

Then today my four words, “Don’t panic. Get fluent,” apparently put me within a millimeter of becoming an enemy of humanity.

Mandy did ask one question worth answering. “How the fuck does one ‘get fluent’ in something smarter than you? Something that never stops changing?”

Since you asked, let me tell you exactly what fluency means to me. I need to recognize where these systems are useful, where they fail, how confidently they can lie, what information I should never give them, how to verify what they tell me, where automation is entering my life and when my own judgment needs to override the machine. I don’t need to match AI’s intelligence. I need enough understanding to make decisions instead of having decisions made for me.

And yes, it keeps changing. That makes learning more important, not less.

I’ve been sounding alarms about the leadership vacuum because this technology is arriving at the worst possible time. Congress and regulators are absent, schools are scrambling, and parents are trying to figure it out on their own while their children are already using it. Meanwhile, corporations have billions of dollars riding on moving faster than everybody else. I’ve written about this over and over because I don’t see anyone in authority doing nearly enough to protect ordinary people from what’s coming.

That’s what I mean every time I say, “Don’t panic. Get fluent.” I’m talking about being prepared enough to protect yourself in a world that’s changing whether you approve of it or not. It’s not a love letter to AI. It’s survival advice.

And here’s the part I’m done softening. None of this has ever been in my control. I didn’t create AI. I don’t control the companies building it, the money behind it, the governments struggling to respond to it or the speed at which it’s moving. This already happened. AI is here, and no amount of outrage in my comment section is going to change that. The only thing that has ever been mine to control is how I respond. I made that decision a long time ago. I’m going to learn it, question it, challenge it, use what helps, reject what doesn’t and prepare for what comes next.

If you want nothing to do with AI, don’t. If you want to sit on the sidelines and complain about the fact that it exists, go right ahead. I don’t give a shit anymore. I’m not spending my time trying to drag people toward a reality they’ve decided they don’t want to understand.

I’m interested in people who are paying attention, asking hard questions, learning enough to protect themselves and their families, and understanding the moment we’re in well enough to build, collaborate and figure out what comes next. Those are my people. If that’s you, let’s work.

If all you have for me is outrage, accusation and contempt because I chose to prepare, you’ve rendered yourself completely irrelevant to the work I’m doing. I don’t need to persuade you. I don’t need your permission. And nobody, certainly not someone I hired to work on my résumé, is going to shame or bully me out of work I know is meaningful and urgent.

I’m not going to be the outlet for someone else’s unhinged outrage. I won’t absorb it, answer for it or let it redirect me. I’m never giving anyone that kind of access to me again. No one gets to unload their fear and anger onto me and leave me questioning work I know matters. That ends here.

I refuse to perform ignorance as an act of moral purity. I’m unwilling to leave myself, my family or the women I teach less prepared because somebody has decided that understanding AI is evidence of bad character. I’ll use what helps me, question what concerns me, reject what I think is dangerous and keep learning because I don’t trust the people in charge enough to do otherwise.

I’ve paid for that position in time, money and an astonishing amount of abuse from people who often don’t have a clue what I actually teach or believe and have shown little interest in finding out. Through all of it, I’ve stayed open to questions, shared what I know and respected other people’s right to have concerns without judging them for it. Some of those concerns are the same ones I’ve spent the past year writing about myself.

But this crossed every line.

I’d paid Mandy for work I respected. I was prepared to give her my curriculum and resources for free because I thought they might help her and her clients. I never attacked her for her concerns, besmirched her or said a cross word. She had every right to reject AI, reject my approach and want nothing to do with any of it. She did not have the right to make me the target of her outrage.

Today she publicly told me “FUCK YOU,” said my position would “kill us all,” called my work bullshit and came within “a millimeter” of declaring me an enemy of humanity.

Hey, Mandy. Fuck you right back.

Melissa Tart is the founder of Farther North Studio, where she teaches experienced women to use AI with confidence so they can turn decades of expertise into their next act.

Learn more or schedule a Discovery Session at stan.store/MTart

Subscribe at substack.com/@melissatart

What caught my attention was something Brown described hearing in rooms most of us will never have access to. Parents ask powerful people in technology what their children should study to prepare for what’s coming, and the advice tends to run toward coding, physics, and other technical disciplines. Yet when some of those same people talk about the education and intellectual influences that shaped them, you hear philosophy, history, literature, the Stoics, and deep reading. Brown noticed the disconnect. I did too. If intellectual breadth helped develop the judgment of people now building our technological world, what explains the gap between what shaped them and what they’re recommending for our kids?

Brown also pointed to Steve Jobs, whose own approach at home was considerably more restrained than the mythology of an Apple household might suggest. Jobs told journalist Nick Bilton that his children hadn’t used the newly released iPad and that the family limited technology at home. Walter Isaacson later described evenings around the Jobs kitchen table filled with discussions about books, history, and other subjects, with devices banned from the table.

The clip moved quickly through a group of capacities Brown believes will matter in the years ahead, and I don’t think they mean much if we simply recite the terminology. Anticipatory thinking is the ability to look ahead, consider several possible outcomes, and think about consequences before they arrive. Situational awareness means understanding what’s happening around you right now and recognizing which pieces of information matter. Temporal awareness adds an understanding of time, including what came before, how quickly circumstances are changing, and how much time you have to respond. Systems thinking asks you to notice how one change affects everything connected to it. Pattern recognition helps you recognize when something happening today resembles something you’ve seen before. Brown also talks about attention, focus, history, nuance, and paradox. Taken together, these are ways of teaching a person to observe, question, connect information, and think several moves beyond whatever happens to be directly in front of them.

Then Brown raised the possibility of an emerging “thinking class,” and that opened a much larger door for me. What happens when some children grow up surrounded by serious reading, history, philosophy, art, music, conversation, debate, and adults who continually expose them to a wider world, while other children have far fewer opportunities for any of it?

Think about two kids from the same town. One rarely travels beyond the place where they grew up. The other has visited different countries, heard unfamiliar languages, walked through museums, eaten food they’d never encountered before, and experienced cultures where the assumptions of home no longer apply. By the time they’re eighteen, they’ve been given very different frames of reference.

Travel is expensive, especially international travel, and it simply isn’t available to every family. Expanding a child’s world can begin much closer to home. Reading can transport them into other countries, cultures, histories, lives, and ways of thinking long before they ever board a plane. Libraries, museums, music, documentaries, cooking, meeting people outside your own circle, and time spent in nature can all widen a child’s understanding. Bare feet in the grass, swimming in a neighborhood creek, walking through the woods, watching birds, planting something and waiting for it to grow, or simply being outside without an agenda can teach forms of attention that no device can reproduce. Even boredom has value when a child has enough space to invent something to do.

That gets much closer to the question Brown’s clip raised for me. What are we deliberately putting into children’s lives while technology becomes more powerful around them and more intimate inside their daily lives?

There’s a famous and admittedly rarefied example in the Kennedy family. Robert Kennedy later said he could scarcely remember a meal when Franklin Roosevelt or events in the wider world weren’t dominating the conversation. The Kennedy children grew up with history and current affairs as a regular part of family life, and even the younger children were expected to know what was happening in the world and have something to say about it.

Very few, if any, of us grew up or vacationed in Hyannis Port or anywhere remotely comparable. Nobody needs to recreate the Kennedy dinner table either. A child can be asked what they believe, how they reached that conclusion, what information might change their mind, and what they think could happen next. You can do that over pancakes on Saturday morning or during ten minutes in the car.

Before any of this turns into another impossible assignment for parents already working full time, driving to practices, managing homework, caring for relatives, and trying to get everybody fed, that reality has to be acknowledged. Families shouldn’t have to invent an entire philosophy of childhood around artificial intelligence by themselves. Schools and public leaders should be helping parents understand what’s changing and how to prepare for it. Too much of that responsibility is falling back on parents and grandparents who are still trying to understand the technology themselves.

I also went looking at the Silicon Valley stories because they’ve been repeated so often that parts of them have hardened into folklore. Steve Jobs really did limit technology at home. Bill Gates has said his children didn’t get phones until fourteen and that phones weren’t allowed at family meals. Evan Spiegel said his family limited his seven year old’s screen time to roughly ninety minutes a week, and Peter Thiel described a similarly small weekly allowance for his young children in 2024.

The famous school story needs more care. In 2011, The New York Times reported that the chief technology officer of eBay and employees of Google, Apple, Yahoo, and Hewlett Packard had children at the Waldorf School of the Peninsula, where classrooms were deliberately screen free. That was real. What grew around it afterward was a much broader story suggesting that Silicon Valley’s elite had collectively rejected technology for their own children. Later reporting found that claim had been greatly exaggerated. Some of the people closest to powerful technology have been unusually deliberate about when, where, and how much of it enters their children’s lives.

Sam Altman, CEO of OpenAI, has said that he used ChatGPT constantly after becoming a parent, asking questions about his newborn and later about developmental milestones. An adult using AI to help think through parenting questions is entirely different from putting the technology into a child’s hands. The relevant question is what happens when children and young adults become the users, and whether the adults around them understand enough to offer meaningful guidance.

By late high school and college, that becomes much harder to ignore. A 2025 study of 666 people divided participants into several age groups, with the youngest ranging from seventeen to twenty five. That group reported the heaviest reliance on AI and the greatest amount of cognitive offloading, which is exactly what it sounds like. Instead of doing a mental task yourself, you hand some of that work to the technology. They also scored lower on measures of critical thinking. The study found an association, so it doesn’t prove that AI caused the difference. What gets my attention is the age. These are people making decisions about college, careers, money, relationships, politics, and the beginning of adult life while a technology capable of doing more of their thinking is becoming part of the furniture.

I use AI every day for many different tasks, and I don’t feel intellectually compromised by it. I argue with it, reject answers, fact check it, make it defend a position, and then make it attack the same position from the other side. Sometimes it’s brilliant. Sometimes it’s mediocre. Sometimes it’s confidently full of shit.

I also arrived here after decades of developing judgment, knowledge, skepticism, and a point of view. A twelve year old hasn’t had those decades. Neither has a seventeen year old. That’s why I care so much about preserving the parts of learning that feel inefficient. Reading the whole book. Struggling with the problem. Writing the first ugly draft yourself. Being bored long enough for your own mind to start looking for somewhere to go. Getting something wrong and having to figure out why. Those experiences are doing work inside a developing brain even when there’s nothing impressive to show for it yet.

I have people in my life, so none of this is theoretical to me. AI will open doors for them that didn’t exist for my generation. It could also upend much of what we’ve always told young people they could count on. What infuriates me is that no one can tell us with any confidence which skills will still be in demand ten years from now, which careers will provide enough income and stability to build a comfortable life, or whether taking on enormous debt for a degree is a smart investment or a six figure gamble. Education has value far beyond a paycheck. If someone wants a degree because they want the education itself, I understand that completely. But millions of families are making that financial decision because they believe a degree will lead to a good job and a secure future. No one can tell them with any clarity that it will. These are ordinary decisions people have to make every day. We’re now making them in the middle of a technological upheaval while the institutions with the most information and power still haven’t given us a clear picture of what they’re preparing for.

By college, the stakes widen again. Fifty seven percent of U.S. college students now say they use AI for coursework at least weekly, with roughly one in five using it every day. Sixteen percent say AI has already influenced their choice of major or field of study. These students aren’t simply deciding whether ChatGPT belongs in an essay. They’re deciding what to study, how much debt to take on, which internships to pursue, what skills are worth developing, and whether the career they imagined at eighteen will resemble the same career when they’re twenty five.

Then they have to get hired.

Stanford researchers tracking the American labor market reported in an August 2026 revision that employment among workers ages twenty two to twenty five in highly AI exposed occupations stood about 19 percent below where it would’ve been if employment had kept pace with similarly aged workers in less exposed occupations. The shortfall had been 15 percent in the July 2025 data. By June 2026, it was 19 percent. The researchers don’t claim that AI caused every bit of that gap. The trajectory still deserves attention. For young adults trying to enter the workforce, the opening itself appears to be getting narrower in some of the occupations most exposed to the technology.

A degree may still matter enormously, but it can’t be the entire strategy. Young adults need enough knowledge to think independently, enough AI fluency to use the technology intelligently, the judgment to know when it’s wrong, and something concrete they can point to and say, I made this, I understand it, and I can explain exactly how I did it.

Current American labor projections don’t give parents a neat answer either. The Bureau of Labor Statistics still expects the United States to add about 5.2 million jobs from 2024 to 2034. Nurse practitioner employment is projected to grow 40 percent. Electricians are expected to grow 9 percent. Even software development, a field sitting squarely in AI’s path, is projected to expand.

Those numbers are real. They’re also forecasts built on conditions and assumptions we can measure today. They can’t tell us what happens if AI capabilities accelerate far beyond what those models anticipate.

That tension makes career advice impossible right now. We used to feel reasonably confident telling a young person to become a doctor, lawyer, engineer, nurse, accountant, or programmer and expect a stable professional life on the other side. Who can honestly promise that kind of certainty today?

One credible American data set shows a widening employment shortfall for young workers in highly AI exposed occupations. Another, focused specifically on employers recruiting the Class of 2026, says hiring of new college graduates is expected to rise this year. Both findings can exist at the same time. Different industries are moving at different speeds, and the picture changes depending on where you look. That complexity is precisely why families need better information instead of broad reassurance.

There’s useful information inside the uncertainty. Employers hiring college graduates are increasingly using skills based hiring, and they want evidence. They’re looking for people who can solve problems, communicate well, work with others, think critically, and explain how they’ve used those abilities in the real world. AI fluency is beginning to appear in entry level job requirements too. Internships, apprenticeships, portfolios, independent projects, research, freelance work, businesses, volunteer experience, and other forms of demonstrated ability are becoming increasingly valuable. A diploma still matters. It has more company now.

And then we run straight into forecasts that sound like science fiction.

Elon Musk has publicly predicted that money could become largely irrelevant by 2036 if AI and robotics create goods and services in such extraordinary quantities that scarcity collapses. When pressed on how people would survive if jobs disappeared during that transition, he said the Treasury could issue checks directly to Americans. Daniel Kokotajlo, the former OpenAI researcher behind the AI 2027 project, has described a different model involving a citizens dividend funded by the AI and robotics economy. His scenario begins with payments around $25,000 per person and eventually reaches numbers so enormous that they barely resemble the economy we know now.

Those are forecasts, not promises, and Kokotajlo himself has been explicit that his scenario probably won’t unfold exactly as written. To me, much of it still sounds like an episode of The Jetsons written by economists. Yet these ideas are being discussed publicly by people who have spent years building, studying, financing, and forecasting this technology. Dismissing them because they sound strange would be as careless as accepting them as inevitable.

The larger question begins before any of those predictions come true.

If work eventually occupies a smaller place in human life, or if income becomes less tightly connected to employment, even partially, we’ll have to reconsider what we’re preparing children for. A good life has never consisted entirely of earning a paycheck, although our culture has spent an extraordinary amount of time behaving as though it does. Music, art, reading, friendship, nature, curiosity, movement, craftsmanship, community, humor, love, and the satisfaction of becoming genuinely good at something have value that reaches far beyond employment. If work changes dramatically, those parts of life don’t become less important. They become part of the infrastructure of a human life.

This is where I stop being patient with the way our public leadership is handling the conversation.

The Trump administration does have an AI agenda. Until I started researching this article, I had never heard of it. Have you? Has your representative explained it to you? Have you seen a public campaign telling families what it means? Did anything arrive in your mailbox or inbox? Have you seen an ad on television explaining what the government believes is coming and how Americans should prepare? If this administration believes AI is consequential enough to build federal policy, workforce programs, and retraining efforts around it, what exactly is it preparing for? And where is the information the rest of us are supposed to use?

The 2025 AI Action Plan lays out the administration’s broader strategy. It also created an AI Education Task Force, launched a national student challenge, expanded apprenticeship efforts, and funded retraining programs. In my opinion, the failure to speak plainly and repeatedly to the American public about what all of this means is a dereliction of duty. We are talking about a technology that could change the course of human history, reshape work, education, wealth, and daily life, and force families to make decisions years before we know how any of it will unfold. Why the hell isn’t the government talking about this constantly? Why are ordinary people being left to discover the plans, the risks, and the preparation buried inside policy documents most of them will never read because they don’t even know those documents exist?

One federal Rapid Reskill grant notice issued in 2026 is remarkably direct about why the money exists. Applicants don’t need to document a conventional qualifying event because, in the government’s own words, “the qualifying event for this grant is the broad and accelerating economic transformation driven by AI.”

Read that again.

The government is funding a program built around broad and accelerating economic transformation driven by AI.

The same AI Action Plan tells the public that AI will improve American lives by “complementing their work, not replacing it.”

Those two statements sit side by side. We are owed answers now: what the government knows, what it is preparing for, which scenarios it considers plausible, and what people are supposed to do with that information while there is still time to prepare.

The government and its institutions are leaving families to reverse engineer their future alone.

Where is the plain English explanation for parents wondering whether their child should take on $100,000 in debt for a degree? Where is the serious discussion for a high school student choosing a field that could look entirely different before they finish graduate school? Where is the national conversation with people in their forties and fifties who may have to reinvent careers they spent decades building? What scenarios is the government planning for? What would trigger a larger response? Which occupations are officials most concerned about? How much disruption do they consider plausible over five years, ten years, fifteen?

Perhaps there’s no settled answer inside those rooms either. Say that.

Show us the scenarios. Tell us what you know, what you don’t know, what you’re watching, and what would cause the government to change course. Stop making families piece the future together from research papers, podcasts, labor reports, corporate announcements, grant notices, and whatever a billionaire happened to say into a microphone that week.

Preparing a generation takes years. School systems take years to change curricula. Teachers need training. Universities need time to reconsider programs. Apprenticeships have to be built before people need them. Families make decisions about debt and education years before a child enters the workforce. Young people begin developing habits around attention, effort, learning, and technology long before anybody calls those habits career preparation.

Those years are passing right now.

That’s why the age progression matters so much. Childhood isn’t a series of interchangeable stages where we can circle back later and repair everything we skipped. Habits accumulate.

A seven year old who learns that a machine can sound certain and still be wrong is beginning to build skepticism before the machine becomes an authority in their life. Stay beside young children while they explore AI. Keep books, imagination, nature, play, boredom, conversation, and making things with their hands firmly in the picture. The goal at that age isn’t technological mastery. It’s preserving the enormous amount of human development already underway.

By middle school, privacy begins to matter differently because kids increasingly use these tools away from adults. Some ask questions about bodies, health, friendships, feelings, and identity that they may never bring to a parent or teacher. The child who has already learned to question an answer has an advantage over the one encountering an AI companion as an unquestioned authority for the first time. Parents need enough AI literacy to recognize what their children are encountering and enough trust with them to stay in the conversation.

High school closes another window. At that point, AI can become a collaborator, tutor, shortcut, research assistant, ghostwriter, adviser, or crutch depending on how it’s used. Ask teenagers to show you what they’re doing with it. Let them teach you something. Find out what they’re building, experimenting with, and learning outside the classroom. Talk about ethics, privacy, intellectual ownership, and the line between extending your thinking and surrendering the thinking you still need to learn how to do.

By college, the consequences have become expensive. Major choice, debt, internships, work experience, AI skills, and the employment market start colliding at once. Students should leave school with evidence of what they can do beyond completing coursework. A transcript describes part of an education. A portfolio, project, business, research paper, internship, apprenticeship, or body of work shows what someone can produce, understand, and defend.

By the time they’re applying for that first serious job, the habits are no longer theoretical. They should know how to use AI without becoming intellectually dependent on it. They should be able to communicate with another human being, solve a problem when there isn’t an obvious answer, defend their reasoning, learn something unfamiliar, recognize when the technology in front of them is wrong, and keep functioning when the machine can’t tell them what to do.

You can’t wait until twenty two to begin building those abilities.

And that takes me back to Brené Brown’s two minute clip.

The question that grabbed me was never simply whether kids should learn coding or read Marcus Aurelius. It was about the kind of mind we’re helping a child build before the future arrives. Brown’s list makes much more sense to me in that context. Anticipatory thinking, situational awareness, history, pattern recognition, attention, nuance, and the capacity to live with uncertainty without surrendering your judgment.

Those abilities reach far beyond the workplace. They help a person navigate a complicated world without giving away their attention, agency, or ability to think for themselves.

There’s plenty about the next ten years that nobody can responsibly promise. Some of the forecasts filling podcasts, research papers, government planning documents, and corporate presentations will be spectacularly wrong. A few may be frighteningly right. We may get an employment boom in industries we can barely imagine today. We may get severe displacement. We may somehow get both.

Uncertainty doesn’t relieve us of the responsibility to prepare.

A happy life may look very different twenty years from now. Success may too. Work could remain central, shrink in importance, or change into something we don’t yet have language for. Whatever shape adulthood takes, young people will need a sense of self that reaches far beyond a job title or paycheck. Purpose can come from relationships, creativity, service, mastery, curiosity, community, and knowing how to build a life that feels worth living.

But I refuse to turn that into reassurance.

We’re raising and educating young people in the middle of a technological transition whose architects routinely describe it as historic, while families are still being asked to make some of the biggest decisions of their lives with fragments of information and wildly conflicting forecasts. The institutions with the greatest access to data, expertise, and power may not know exactly where this is going either. That makes openness and preparation more urgent, not less.

We may not know what world these kids are inheriting.

Leaving them to figure it out after it arrives is unacceptable.

Melissa Tart is the founder of Farther North Studio, where she teaches experienced women to use AI with confidence so they can turn decades of expertise into their next act.

Learn more or schedule a Discovery Session at stan.store/MTart

Subscribe at substack.com/@melissatart

Probably more than you think. That’s what makes this worth your attention. You don’t have to use AI, understand it, or want anything to do with it for surveillance technology to affect your life. Most people don’t have time to track every patent, workplace tool, camera network, or monitoring system being built around them. Your approval wasn’t required for any of it. Much of it’s already here.

If you’re on Substack, put this on listen mode and let it run while you make coffee, get dressed, or move throughout your day. This article isn’t about how to prompt or deploy AI agents. It’s about what’s already watching you, what it can collect, and what happens when you don’t know it exists.

You could swear off AI completely, delete every app, go back to a flip phone, and none of it would stop your employer’s software from logging your keystrokes, the camera on your neighbor’s porch from scanning your face, or Ford from holding a patent that could one day decide whether your car even starts. Employers, automakers, and camera companies made that decision for you years ago, often wrapping the technology in explanations so harmless you’d have signed off without a second thought: a tool to find lost dogs, a dashboard to help managers keep tabs on remote teams, a safety feature to catch drunk drivers before they hurt someone.

Call it what it is. If someone tracked your movements, recorded your logins, and pinpointed your location without permission, you’d call it stalking, whether it came from an ex-partner, your employer, or a corporation collecting data on you. A badge or logo doesn’t make it less invasive.

For a lot of people, the fear isn’t falling behind. It’s losing a job, an income, or the stability they’ve spent years building. At first, the questions were simpler: could I learn this, and was I already too late? Now the anxiety has hardened into something closer to anger. People who’ve spent decades learning how to work, build a career, support a family, and plan for the future are watching those assumptions change without their permission. They didn’t ask for any of this, yet they’re expected to adapt while jobs shift, entire industries reorganize, and almost no one in leadership offers a clear picture of what comes next.

People are exhausted by the constant AI conversation, from slop and stolen work to disappearing jobs and data centers consuming power and water. Those concerns are real. Adults who’ve spent decades building knowledge and judgment don’t lose it because they draft an email or use an agent. That’s not the argument here. The missing conversation is covert surveillance: who gets to watch you, what they can do with the information, and whether you’ll recognize the danger before it reaches you, even if you never use AI.

Start with your job, because it’s the closest one to home, and it’s already worse than you think. Researchers at Northeastern University and three other universities examined nine of the most widely used workplace monitoring platforms, tools with names like Hubstaff and Time Doctor marketed as ordinary productivity software. A third of them could track precise location at any time, even while the app sat in the background and you were off the clock. Every single one shared personal data with outside companies, feeding information to more than 145 different domains, including Google, Meta, and Microsoft. A separate survey of 1,500 U.S. employers found nearly three out of four already use some form of online tracking, most of the workforce has no idea it’s happening, and more than half still report real stress and anxiety from being watched.

Then there’s who gets hired. Stanford researchers examined 4 million job applications across 156 employers and 11 industries, all run through automated screening platforms. The software discriminated against 26% of Black applicants and 15% of Asian applicants who applied, at a level that meets the Equal Employment Opportunity Commission’s own legal definition of adverse impact. Had the tool simply recommended people at the same rate across groups, 40,000 more qualified submissions would have advanced. Assessments like these are already embedded in pipelines most people never see, deciding who gets a second look before a human ever does.

Step outside your job and the picture doesn’t improve. Flock Safety built a network of license plate reading cameras that now blankets huge stretches of the country, sold to police departments as a tool for solving crimes. It’s also been used, repeatedly and documented in criminal court, for something else entirely. One Milwaukee police officer ran the system more than 170 times over two months to track a woman he was dating and her former partner. A separate case in Joplin, Missouri involved an officer who queried one woman’s plate nearly 400 times. A police chief in Sedgwick, Kansas used the cameras over 200 times in four months to stalk his former partner and her new boyfriend, and lost his law enforcement certification for it. Flock’s system places no automatic requirement on anyone to justify a search and no flag when the same person gets queried again and again outside an active case. Most of what we know about the abuse didn’t come from Flock’s own oversight. It came from whistleblowers and the people being stalked. This wasn’t simply a public safety tool. Once a camera network can map vehicles, routes, routines, and relationships, anyone with access can reconstruct a person’s life. At that point, the danger isn’t hypothetical.

Your car isn’t exempt either. Ford filed a patent for a system that, after a missed payment, can lock out features like GPS, air conditioning, and the radio, and eventually drive the vehicle itself to an impound lot. Ford says the patent doesn’t confirm product plans and insists it has no intention of building the system. I don’t believe that settles anything. Companies don’t patent technology like this for decoration. Maybe the final version looks different, but I think some form of it is coming. A separate federal requirement, not connected to Ford’s debt-collection patent, calls for new vehicles to include passive impairment detection through eye tracking, steering behavior, or breath sensors, with a rollout target of 2027. Regulators say no system on the market yet fully meets that standard, but some manufacturers say their monitoring tech can already detect intoxication using pulse readings and driving behavior, with the ability to intervene. The systems serve different purposes, but both shift more control from the driver to the vehicle.

And then there’s the story with the softest cover of all. Ring built a feature called Search Party, marketed with a Super Bowl ad about reuniting families with lost dogs, using AI to scan footage from cameras across an entire neighborhood. In a leaked internal email, Ring’s own founder told employees the dog-finding tool would end up becoming one of the most important pieces of technology the company would ever build to unlock the impact of its mission. Ring is currently facing a federal class action lawsuit over capturing and storing people’s images without their consent, and members of Congress have raised concerns directly with the company in the past over its facial recognition ambitions. The dog-finding feature gave millions of already-installed cameras a new, more ambitious job before anyone thought to ask what else that network could see.

Tracking can save a life. If someone you love disappears, camera footage, location pings, and digital traces may become invaluable. That benefit is real, and so is the cost. Those same tools can map an ordinary person’s movements every day, long after any emergency has passed. The question isn’t whether surveillance can ever do good. It’s whether rare moments of rescue justify surrendering privacy, autonomy, and the right to move through the world without being constantly recorded. We should be told that bargain exists before it’s made for us.

You can’t protect yourself from a system you don’t understand, and you can’t challenge one you can’t see. Companies and public agencies already collect more about our movements, habits, work, and relationships than most people can account for. We should know what’s being gathered, who can access it, how long it’s kept, and what recourse exists when the information is wrong or abused. Awareness is only the beginning. The public should be demanding transparency, enforceable privacy rights, stronger consumer protections, and real consequences for misuse.

I started studying AI three years ago and noticed a gap almost immediately. I rarely saw women in mid to late career on the stages or in the conversations shaping this technology. At the same time, many were being hit with jargon that had never been part of their working lives. I didn’t have the solution yet, but I knew the problem was real. Farther North Studio grew from that recognition. I teach experienced women to use AI with confidence so they can turn decades of expertise into their next act.

My work is centered on women in mid to late career, but surveillance has ramifications for everyone. It doesn’t stop with the present or the person standing in front of the camera. Records created today can shape jobs, insurance decisions, investigations, and opportunities years from now. Children are especially exposed because they’re growing up inside systems they didn’t choose and don’t yet understand.

We’ve spent years warning kids that what they post online can follow them. Now the conversation has to include what they never chose to post at all. Cameras, phones, cars, schools, and automated systems can create records of mistakes, routines, friendships, and locations while judgment is still developing. We already know that criminal and disciplinary records can affect college admission, athletic scholarships, employment, and other opportunities. As surveillance expands, it isn’t unreasonable to ask what happens when even more of a young person’s life becomes searchable, stored, and available long after the moment itself has passed.

Parents need a new version of the talk. Alongside sex, drugs, driving, and social media, children need to understand surveillance: when they’re being watched, what their devices reveal, who may have access, and how to protect themselves. Technology marketed as security can also create a record of daily life. They should be able to recognize when safety crosses into surveillance before a mistake or pattern becomes permanent.

What You Can Do

Treat company devices as monitored. Employer-issued laptops, phones, and tablets may log activity, location, audio, or video outside normal working hours. Keep personal conversations and sensitive information off them.

Don’t install work apps on a personal phone unless you understand every permission they require. Ask for a company-owned device when possible.

Keep your own record of completed work and results. Monitoring software measures activity, not value, and those aren’t the same thing.

Read your handbook and state law. Find out what your employer must disclose, which tools are in use, and how the data influences hiring, scheduling, reviews, or discipline.

Compare notes with coworkers or your union. A pattern affecting several people is harder to dismiss than one isolated complaint.

Ask what connected technology collects before you buy or install it. Where the data goes, how long it’s stored, who can access it, and whether any feature can be disabled all matter.

Teach children basic surveillance awareness. They should know that cameras, apps, and connected devices can create records even when they haven’t posted anything themselves.

Knowledge alone won’t dismantle these systems, but it changes your position inside them. You can question what data is being collected, refuse permissions that aren’t necessary, challenge automated decisions, and demand stronger protections from employers, companies, and elected officials. You can also teach your children to recognize surveillance before it becomes invisible to them.

The cameras are up, the software is running, and the patents are filed. The only thing left undecided is whether you’re paying attention.

Melissa Tart is the founder of Farther North Studio, where she teaches experienced women to use AI with confidence so they can turn decades of expertise into their next act.

Learn more or schedule a Discovery Session at stan.store/MTart
Subscribe at substack.com/@melissatart

The share of salaried jobs open to people with no experience has fallen 73% in four years. In 2022, inexperienced candidates were considered for about one in every 15 openings. Now it’s one in 50.

And this isn’t only affecting new graduates. Experienced workers are applying for entry-level jobs because there aren’t enough positions anywhere else. The entire workforce is being pushed downward while employers expect more from every person they hire.

I’ve been in rooms with CEOs who have said this plainly. They are not hiring anyone who doesn’t arrive fully trained, completely AI adaptable, and ready to deploy three AI agents on the company’s behalf from day one. I could name some of them, but I won’t.

That is the language being used behind closed doors.

You don’t have to like AI. You don’t have to celebrate what it’s doing to the job market. But refusing to learn it will not save you.

This is only the beginning. What’s happening now will look small compared with what’s coming, and I know that sounds dark. I’m saying it because I’m trying to help you prepare before the avalanche reaches you.

You can love or hate AI. You may want it regulated into the ground or banned altogether. I don’t care which one you pick. What I care about is whether you understand what’s actually at stake. Most of you don’t. Not deciding is still a decision.

Three years ago, I went through the whole cycle myself, curiosity, dismissal, uncertainty, I’m too old for this, turned into I’ll just learn enough to get by. I wanted to know if this so called magic could do something for cancer research. I spent the next two years learning as much as I could while balancing a full-time job and everything else in between. Picking up complex systems has never been an issue for me. I was an early adopter of technology going back to the cell phone that came in its own carrying case, and a boxy, ugly, clunky old dinosaur of a computer. But this was different. It was a mind fuck.

What that time gave me was insight. I recognized it immediately, because I’d been living it myself. The jobs most exposed to this technology are disproportionately held by women, particularly women mid to late career, carrying a higher salary and benefits package than their younger counterparts. Their excuses were my excuses, word for word. It won’t touch me. I’ll be retired before it matters. We’re years away from that being a problem. Curiosity, denial, fear, all of it. I know blissful ignorance when I see it.

We built our whole professional identity before any of this existed, and now we’re being told to start over or else, and that feels like an engineered threat, because it is. It doesn’t care how close you are to retirement, how many years you’ve put in, or how far along you thought you were. It’s coming anyway.

I watched an entire industry line up in the front row of a presidential inauguration, fawning and preening like courtiers granted an audience with the king. That image is burned into my memory, because it answers a question a lot of us keep coming back to. Can’t they see what it’s costing us? They can see it, and they don’t care. Period. They are singularly focused on their own interests, and they are not going to slow down or self-correct.

Waiting them out isn’t a strategy. It’s not for you. You’re not in the club. You’re going to have work taken from you, whether you want it or not. You have no say. There’s a serious proposal from inside the AI research world that your compensation could land somewhere between twenty five thousand and ten million dollars a year. Don’t sit around waiting on that ten million. You’ve earned it. I hope you get it. I wouldn’t bet your life on it. They stole your life experience, used it to train your replacement, and now they’re selling it back to you. This is about preserving your life and your livelihood, and protecting the people who depend on you. That’s why learning this is no longer an option.

Let me give you two facts, both from this week, both real. Meta just signed a deal to train its AI on Newsmax, which paid out over a hundred million dollars combined settling election fraud defamation claims with Dominion and Smartmatic. Meta already had Fox News in that same training pipeline, the network that paid Dominion seven hundred eighty seven million dollars for spreading the same kind of lies. Add in the Daily Caller, the Washington Examiner, and the New York Post’s parent company, and you’ve got a very clear pattern of who Meta decided was worth paying for content. That same day, over a thousand employees at OpenAI and Anthropic signed a petition asking governments to help humanity keep pace with AI, so it doesn’t outrun our ability to supervise it. Mark Zuckerberg published an op-ed arguing for the opposite. Gas, not brakes. So if you’re getting your news and your opinions about the world through anything owned by Meta right now, you are being fed information from outlets that aren’t news organizations anymore. They’re one-sided entertainment companies with no balance to offset them, and they’ve told you, on the record, that slowing down isn’t their plan. That’s not a conspiracy theory, that’s Tuesday.

Be furious about that. Let the fury turn into insight, then action, instead of looking away.

I’m not telling you to use AI. That’s your call. If you don’t want to touch it, and your competitors do, that’s a risk you’re choosing to take, and I’m not saying it’s wrong to stay raw and analog in a world that’s changing this fast. But I will say this. It is not sustainable.

You can wait as long as you want to actually use AI. What you can’t wait on is understanding it.

Let me give you an example. My daughter rails against this harder than almost anyone I know. She’s one of the most vulnerable people to what’s coming. She’s in denial. I stopped arguing with her a while ago. Now I just show her things I create, a carousel, a meme, something funny, something serious, an app I built, and she approves, and sometimes asks me how I did it. People don’t always follow your advice. Sometimes they follow your example. Another family member had a job loss this year and had to put important life events on hold. It was a combination of factors most families are familiar with by now, rising costs, tariffs, AI stepping in where it can. It’s all connected, and it’s hitting families like mine, and probably yours, if not now, then sometime soon.

This shift doesn’t require a degree in technology or a master’s in engineering, or any of the credentials you’ve been telling yourself you need as a reason to keep avoiding it.

You don’t have to love this technology, and you don’t have to use it, but you do have to stop pretending it isn’t coming, because dabbling a little isn’t the same as understanding it, and understanding it is the only thing that’s actually going to protect you. You cannot bring a pebble to a fight this size.

Once I understood what was coming, a shift in focus and direction was the right call for me. This past year, I built a private community, filled with tools and material for people like you. I kept it private, on purpose, to protect your ideas and mine. At the same time, I became a certified coach so I could teach this properly, and my own learning hasn’t stopped since, it’s only accelerated, because the pace of change doesn’t wait for anyone, including me. This is the clarion call. You’ll have to decide if it’s important enough to act on. Maybe you save this and come back to it when you’re ready, but the truth is the best time to act was a year ago. Right now is the next best time.

Out of everyone in my family, I went first, because I needed to protect mine. Now I’m asking you, who’s going first for yours?

You don’t have to figure this out alone. Reach out to me directly, and I’ll provide you the link to join the Farther North Studio community.

Melissa Tart is the founder of Farther North Studio, a coaching and educational platform for women in mid-to-late career who are navigating change, building what’s next, and learning to use AI with confidence in a rapidly changing world.

To learn more or schedule a Discovery Session, visit: stan.store/MTart

Subscribe: substack.com/@melissatart

Turns out evergreen is a content marketing term. It means something has sustaining value, it sticks around a lot longer than your last boyfriend. AI isn’t evergreen. It can change before sunset, never mind by next month.

So here’s my evergreen opinion about AI. Stop wasting your money on prompt libraries. I’ve got more of them than most people probably do, and I never open them, because I already know they aren’t going to give me what I want or what I need. That doesn’t mean everybody selling prompts is a fraud, or that every prompt inside one is useless. They were written to sell a mountain of them to everybody at once. If you’re starting out, go ahead, buy fifty dollars worth. I’ll sell you some myself.

Unless you’re just playing around, testing the waters, getting a feel for what these tools can even do. Then sure, an off-the-shelf prompt might be useful. It’ll get you moving. But for most people it becomes clear fast that generic is exactly the problem. These prompts don’t know a single thing about you. They’re built to get as many people curious enough to start experimenting. That’s the best use for them, and that’s the most honest thing I can tell you.

Buying somebody else’s prompts is like walking into the best little family restaurant in Italy, sitting down, opening the menu, and ordering buttered noodles. You’ll eat. You’ll leave full. You also skipped the entire reason anybody waited three hours for that table. You could order Nonna’s Sunday gravy off that same menu, plenty of people do, and it’ll be good. But it’s never going to taste like it does when you make it with her, in her kitchen, watching her stir and taste and adjust, getting it wrong yourself a few times before it finally tastes like hers. She stopped measuring thirty years ago, and she couldn’t write the recipe down for you even if she tried.

Let me be transparent with you. I don’t use anybody else’s prompts, custom GPTs, or gems. I’ve bought the courses, spent the money, sat in the rooms, and none of it was wasted. But here’s what three years of that taught me: eventually you have to build your own, because most of what’s out there is built for hordes of people, not your specific life. If you’re just playing around, an off-the-shelf GPT is fine. If AI is actually going to be part of how you work, waiting around for someone else’s tool to get better is time you don’t get back.

I’ve been at this three years, and I still argue with my AI constantly. Some days it’s brilliant and I tell it so out loud like a lunatic. Some days it hands me something so lazy I’m sitting here going, sei un deficiente, che stronzata, how is this still happening, you have three years of my context. One of my chatbots recently told me it wasn’t going to argue with me anymore. That fight is the whole process. Nobody sells you that in a bundle because it isn’t a product. It’s time, spent badly, and then badly again, and then, finally: brava.

Maybe you’re just starting out, without much of a skill set built up yet. Amica mia, I’m trying to save you some money here. The bigger value isn’t the fifty dollars, it’s skipping the part where you realize three weeks in that none of it actually fits your business, and you’re starting over anyway. Put your energy where it’s actually going to matter, and it’s not on a set of five thousand prompts. You can buy them. Nothing’s stopping you. You won’t get the output you’re expecting, because it was built for everybody, and you are not everybody.

So here’s what actually happens when we work together instead. I dig into what you’re trying to accomplish, what your timeline is, who you are and what you do, until you’re a little uncomfortable with how much I’m asking. That’s the AI Strategy Session. We look at what’s already worked, what hasn’t, or whether you’ve tried anything at all, and then we fix it. It’s a bespoke approach built around your business and your voice, not the right words fed into a box a thousand other people are typing into. You stay adaptable while the tools keep changing on you, sometimes before you’ve even finished learning them. And when it finally works, you’ll understand exactly why.

Buttered noodles are what you got today, free, in this article. If you want the sauce, find me at stan.store/MTart.

I use these tools every day. I teach people how to use them. Sometimes the output is excellent, and sometimes it’s absolute shit. That’s just reality. So I’m not criticizing AI from the outside. I’m criticizing the mythology that’s grown up around it.

People keep saying AI is going to make everyone dumb. How exactly?

There are legitimate concerns when you’re talking about kids whose brains are still developing. A kid shouldn’t be allowed to hand every assignment to AI and let it do all the thinking and writing. There has to be a lever on that. That lever is called parenting.

These tools can help kids learn, explore ideas, ask better questions, work through difficult problems, and satisfy their curiosity. They’re going to use AI whether you like it or not. That ship has sailed.

The question isn’t whether they’re going to use AI. The question is whether you’re going to teach them how to use it responsibly or leave them to figure it out on their own.

That’s your job.

Try telling someone who’s 58 that everything they’ve learned up to this point is going to disappear because they asked AI to help streamline a presentation. Come on.

You’ve already lived an entire life. You’ve learned, failed, adapted, reinvented yourself, changed careers, started over, recovered from losses, figured out impossible situations, and taught yourself things because you had to.

If somebody wants to hand AI the keys to their brain and stop thinking altogether, that’s their choice. But don’t stand there and tell me using AI to help with my work suddenly made me stupid. That’s absurd.

And you developed a voice along the way. Wit. Humor. Sarcasm. Timing. Instinct. The ability to know when something is funny, when it’s cruel, when it’s ridiculous, and when it deserves to be called exactly what it is. AI doesn’t have any of that on its own. It borrows from us. It can imitate a tone, but it doesn’t have a life behind it. You do.

And that’s really what nobody wants to talk about or admit. It’s not about whether these tools are useful. They obviously are. It’s about the fact that every word, idea, and scrap of human thought ever put out into the world was absorbed into building these systems. Nobody asked us, and nobody paid us for it.

And here’s the part that’s almost funny if it weren’t so infuriating. Everywhere we look, it’s AI, AI, AI. AI is going to take your job. AI is going to change your industry. AI is going to run your business, manage your life, and apparently become your lover. We’re pushed to use these tools, get comfortable with them, learn the agentic workflows, streamline everything, and adopt AI or get left behind.

Then they turn around and build another AI tool to detect whether you used AI so they can slap a scarlet letter on your forehead and condemn you for doing the exact thing they told you to do.

Rush, rush, rush to use it. Learn it. Adopt it. Build with it. But wait. Now we’re going to shame you for using it.

What the hell do they want? I don’t think they know.

And honestly, fuck them. Whoever them even is.

Use the tools, don’t use the tools, I don’t care. They’re not going anywhere. Choose your poison.

I bring AI my original work. It doesn’t replace the thinking, experience, or judgment that created it.

The tools enter after that, when the editorial process begins.

I never accept the first draft, and I never rely on just one AI. I ask a question, get a response, then take that response to a second AI and give it a far more demanding assignment. I ask it to act as one of the best editors in the world, take out its sharpest scalpel, and cut the work apart. I want it to challenge the reasoning, identify weak spots, find errors, question anything that feels unsupported, and keep going until the piece is strong enough that I’d be willing to put my name on it.

It always finds something. The first answer is a rough draft, and often not a very good one. Left alone, it can be shallow, clumsy, incomplete, or flat-out wrong. The second AI isn’t there to admire it. It’s there to make sure I don’t publish something I’d be embarrassed to claim as mine.

And yes, this process can take hours, days, or even weeks. It’s not unusual for me to go through 25 to 50 iterations before I’m satisfied. I don’t stop because an AI says it’s finished. I stop when the work has earned my confidence. That’s the level of responsibility I believe comes with putting my name on something other people will read, rely on, or act upon.

So my process now looks like this. I ask my question. I get an answer. I hand that answer to a different AI and ask it to tear it apart. Then I take whatever it finds back to the first one and ask if it agrees. If the answer includes quotes, statistics, dates, or other factual claims, I run those through a separate verification step using a Claude Skill connected to Perplexity through MCP. I don’t stop until I’ve challenged the reasoning and verified the facts. Then I make the final decision.

Two AI tools, arguing it out, with me as the judge.

What the review process actually looks like

The doctor analogy was later rejected. Leaving it marked Pending here shows the process as it actually happened: some decisions are immediate, while others need distance before they become clear.

Every AI has blind spots. Some confidently invent facts. Some miss obvious weaknesses. Others simply stop too soon. A second AI isn’t there to agree with the first one. It’s there to expose what the first one missed.

I use this process for client work, content, research, and more. Every project goes through the same quality-control process because that extra verification step helps ensure factual claims are accurate before I publish or rely on them.

Here’s the version you can steal today. Take your last AI response and hand it to a different AI. Don’t ask for feedback. Give it a job. Tell it to become an elite editor. Tell it to challenge every assumption, cut weak writing, question unsupported claims, and keep editing until the work is publication-ready. Then compare the two versions and decide which one earns your name.

That’s the whole technique. Not complicated. Remarkably effective.

Two AIs sparring by default. A third joins the conversation when the facts need verifying.

Melissa Tart is the founder of Farther North Studio, a coaching and educational platform for women in mid-to-late career who are navigating change, building what’s next, and learning to use AI with confidence in a rapidly changing world.

To learn more or schedule a Discovery Session: stan.store/MTart

Subscribe: substack.com/@melissatart

What the hell are they talking about.

That was the whole reaction. Not curiosity or intrigue. A woman who’s spent the better part of three years inside AI every single day, testing it, breaking it, building with it, running headfirst into a sentence that stopped her flat anyway.

So I did what I tell every Starlet inside Farther North Studio to do the moment this happens. I opened Claude and typed one sentence.

Tell it to me like I’m a fifth grader.

Here’s the version that actually made sense to me.

Say I ask an AI to find my granddaughter a birthday present. It searches, hands me three links, and it’s done. That’s a single step, checked once, finished. That’s a loop.

Now say I tell it what’s actually in my head. One of my granddaughters is turning ten. She’s an animal lover down to her bones, her room is a full zoo of stuffed animals collected since the day she was born, and every single one of them provides her comfort, not decoration. She doesn’t travel anywhere, not even to my house, without a stuffy tucked under her arm. She already did the Animal Kingdom at Disney and it was a full triumph. Budget isn’t a factor. I’d rather give her an experience than an object, and if the first idea doesn’t fit her, I want a second one ready without starting the conversation over.

Now the AI isn’t running one step. It’s weighing several paths at once, comfort, meaning, budget, backup plan, and choosing which one to follow depending on what it finds. If the first idea doesn’t fit, it branches to the next one instead of stopping. That’s the basic idea behind what engineers call a graph.

Here’s what actually stopped me, though. Claude pointed out that I’m already doing this. When I research something, then ask it to explain that research, then have it fact check itself, then rewrite, then build an image, then polish before I publish, I’ve been running a workflow with branches and fallbacks built in for a while now. I simply never called it that. Or I never called it anything at all.

I’m not new to this. But every so often a piece of language still catches me off guard, the way loops versus graphs did this morning. When that happens, I don’t wrestle with it. I hand it to the tool and ask it to hand it back to me in plain language.

Nobody gave me that move when I started. I’m giving it to you now.

The AI world has a habit of taking an ordinary idea and dressing it up in language that makes you feel like you’ve lost a fluency you used to have. Most of what sounds intimidating is a familiar idea wearing a costume.

Here’s what I actually want you to walk away with. It isn’t about not understanding something. It’s about having a tool. And the tool is this: use AI to solve AI.

Confused by a term or an idea in an article? Open whatever AI you use and ask it to explain the thing that just confused you. My own preference is to run every answer through two AI tools at once and ask each one how it applies to the exact work I do, but that’s a lesson for another day.

That’s the whole move. Something you can use every single time this world throws language at you that doesn’t make sense yet.

Tell it to me like I’m a fifth grader.

That’s it. That’s the tool.

Melissa Tart is the founder of Farther North Studio, a coaching and educational platform for women in mid-to-late career who are navigating change, building what’s next, and learning to use AI with confidence in a rapidly changing world.

To learn more or schedule a Discovery Session, visit: stan.store/MTart

Subscribe: substack.com/@melissatart

There is one catastrophic problem with that performance.

Donald Trump’s own administration has been dismantling the defenses that protect it.

Nearly 1,000 employees have left or been removed from active service at the Cybersecurity and Infrastructure Security Agency, the country’s primary cyber-defense agency for election infrastructure. That represents nearly one-third of CISA’s workforce. His administration ended federal support for the Elections Infrastructure Information Sharing and Analysis Center, which helped state and local election offices detect threats, share intelligence and defend their systems against cyberattacks.

Attorney General Pam Bondi disbanded the FBI’s Foreign Influence Task Force, the team created to identify and disrupt foreign operations targeting American politics and elections. The task force coordinated expertise across the FBI’s field offices, intelligence agencies, election officials and technology companies. The administration eliminated it while Trump now demands that Americans panic about foreign influence.

Trump has also fired the Democratic commissioners of the independent Election Assistance Commission and allowed its remaining Republican commissioner to resign, leaving the bipartisan body that helps test and certify voting systems with no commissioners at all. His administration spent more than a year scrutinizing the agency, questioning its rejection of Trump’s 2020 election lies and vetting potential replacements aligned with his political agenda.

This is not election security.

It is political arson followed by a televised warning about smoke.

Trump is not merely failing to strengthen the country against the dangers he describes. His administration has systematically weakened the institutions designed to confront those dangers. He cannot dismantle cyber defenses, eliminate foreign-interference investigators, hollow out election oversight and then pretend he has discovered a vulnerability created by someone else.

To the extent that America is now less prepared, Donald Trump helped make it less prepared.

That matters because Trump’s concern about election integrity has always been conditional. He did not question the elections he won in 2016 or 2024. He continues to attack the election he lost in 2020, despite recounts, audits, court decisions and reviews conducted by Republican officials and members of his own administration finding no widespread fraud that changed the result. His July 16 speech offered no credible evidence that a foreign power altered vote totals in 2020, 2022 or 2024.

Now the country is approaching another election that will determine control of Congress, and Trump is doing exactly what he has done before. He is manufacturing suspicion in advance. He is telling his supporters that the system cannot be trusted unless it produces the result he wants.

Only this time, he is doing it after stripping away some of the very protections that could reassure the public.

The danger is not simply that foreign governments may attempt to interfere. That threat is real and should be confronted with more expertise, more coordination and stronger institutions.

The danger is that the president of the United States is weakening those institutions while preparing to use their weakness as evidence against the legitimacy of an election.

That is not the behavior of a leader trying to protect democracy.

It is the behavior of a man preparing an alibi.

The address was 2016 Thomas Street. She died in 2016.

I’ve had ten years to decide what to say about that, and after all this time, I keep coming back to the same conclusion. For years, I gave far too much weight to the coincidence of that address. The address is irrelevant now. Michelle was, and always will be, the story. And if the people who loved her stop telling her stories, the disease gets the last word on who she was. Not on my watch.

If you’ve ever lost someone you love, you know the fear I’m talking about. It isn’t just that they’re gone. It’s that, little by little, people stop saying their name, thinking silence protects you. The opposite is true. Hearing their name is a gift. So is a story you didn’t know, or a photograph you’ve never seen. There will never be another memory made, and whatever’s already here is all there’s ever going to be. I’ll never tire of hearing Michelle’s name.

The Quiet One

Michelle was the third of four, and the quiet one, which in our family counted as a talent. Three girls, then a boy, all of us light blonde except her, who came out strawberry blonde and stayed that way her whole life, and except my brother, who arrived two years after her with hair like a struck match. She sat between us in color the way she sat between us in everything else. Her friends called her Switzerland, because whatever war the rest of us were waging, she declined to join. I teased her for being boring, because in forty-seven years she never once gave our parents a reason to lose sleep. She took the teasing the way she took nearly everything. She didn’t need to win. She just needed everyone to be okay.

She was born on August 13th, my Aunt Frances Irene’s birthday, so my mother named her Michelle Irene and considered the matter settled by heaven. Not that the name got much use. Her nursery was decorated in Winnie the Pooh, and somehow the nickname stuck. My mother started calling her Pooh when she was a baby, the rest of us called her Mishy Pooh, and neither name ever left her. Even now, ten years after her death, my mother still calls her Pooh, just as often as she says Michelle. That name has never gone anywhere.

My older sister sent me two memories this week that I’d never heard, and I’m grateful for both. Bee stings were an ordinary hazard of our yard, ordinary enough that my mother kept baking soda by the porch to mix into paste, and my sister remembers getting stung while she was carrying Michelle, sitting down to dab the paste on her own toes, and Michelle, too little for real sentences, saying “bee boo, bee boo,” patting her hand, kissing it. She also remembers the white pebbles she was made to clear from around the tree out front, how she’d set Michelle down in the grass and worked her way around the ring, and how Michelle would pull the pebbles back out behind her, one by one, squeaking with delight. That was Michelle before she could talk in full sentences: comfort and mischief, both at once, already entirely herself.

For a while, when we were small, the Christmases were enormous. Ten-speeds, motorcycles, and snowmobiles waiting outside, then stereos, jewelry, and beautiful clothes for three girls under the tree, because my mother did nothing halfway. It didn’t stay that way forever, and the family joke about those years belonged to Michelle: “back when we was rich,” grammar wrong on purpose. I still say it. It’s still hers.

We grew up outside, in a family that ran on gasoline and cold water. Motorcycles and dune buggies on my grandparents’ peach orchard, hours deep in the woods with our dad. And in summer, the creek behind the house, where certain pools went deep between the rock formations. Our dad would dive into one of those pools and simply not come up. He could hold his breath longer than seemed survivable, and Michelle and I would scan the surface, half thrilled and half sick, until he broke the water clear across the creek. We fell for it every single time. It never occurred to us that we could lose anybody. That was the childhood: nobody we loved stayed under for long.

Who She Became

She studied wine at Johnson & Wales and cooked the way some people garden, seriously, and for the pleasure of feeding the people she loved. Small dinner parties at the bungalow, good bottles, a table full of friends she’d collected without any visible effort. Her dogs came one at a time, and each ran the house in turn: first George, a wirehaired fox terrier with the manners of a disapproving landlord, and after George died, Henri, the soft one, the comfort dog. Outside the door she kept a feral cat named Sophie fed and accounted for, never allowed in, never missing a meal. She loved every animal she ever met except birds. She hated birds. Nobody ever got an explanation, and I’ve stopped wanting one. A person should get to keep a mystery or two.

She spent her career in insurance, on the underwriting side, reading risk for a living.

Michelle wasn’t denied a beautiful life. She had one. Friends who adored her, dogs she loved, wonderful food and wine, beach houses and mountain houses, a deep love of traveling the world. She built a life full of experience, adventure, and people who loved her.

She didn’t want children of her own. The man she loved already had a son and a daughter, and she was happy to be part of their lives. She spent ten years with him, starting at thirty, and wanted exactly one thing by the time she turned forty: a thoughtful proposal. She didn’t care about an extravagant wedding, an expensive ring, or a bigger house. She just wanted to know that the life they’d built together mattered enough for him to choose her on purpose.

Before she left for Ohio to help care for our grandmother, they went to dinner. He tossed out something like, “Do you want to get married?” It wasn’t a proposal. There was no thought behind it, no planning, no joy in it. It hurt more than if he’d said nothing at all.

Michelle spent several weeks in Ohio helping care for our grandmother. I flew in so she could go home. On her fortieth birthday, she got on a plane expecting to return to the life she’d spent ten years building. Instead, she walked through her own front door and found him standing there with his bags packed, on his way out.

I was still in Ohio when she called me. Instead of the thoughtful proposal she’d quietly hoped for by forty, she spent her birthday mourning the future she thought she was flying home to.

Ten years together. No proposal. Only goodbye.

Our grandmother died shortly after that, and Michelle came to me and said our grandmother had visited her and told her everything was going to be okay. I believed her then, and I believe her now. In the seven years she had left, she truly let that relationship go, which is more than I can say for myself, because she asked for the simplest thing a person can ask for after ten years and didn’t get it, and she didn’t have a partner beside her at the end either. Michelle made her peace with that. I’m still working on mine.

The Augusts

After the breakup, she gave me August. Every year, the whole month, she’d arrive with what looked like everything she owned, set up to work from my place, and for one month a year I had my sister full time. We cooked. We sat on the beach. Her friends drifted down for a few days at a stretch and drifted home again. Her birthday fell in the middle of the month, so in those last years, every age she turned, she turned at my house. There was also the great squirrel caper, which I won’t be explaining. The people who were there know. That’s the point of it.

One of those summers produced the cake. She’d fallen for a dessert at her friend Gloria’s, a plain yellow cake from a box with cream cheese and mashed fresh strawberries, gooey in the middle, and she’d raved about it ever since. So at a beach house in South Carolina we made it together, followed the box exactly, and it came out perfect. Perfect was wrong. “Oh my god, we overcooked it,” she said, and that’s how we learned that Gloria’s masterpiece had been underbaked by accident. We made a second one and underbaked it on purpose. My sister knew exactly what she loved, even when what she loved was technically a mistake, and she saw no reason to let the instructions overrule her.

And every Christmas her friend Katie made her cookies, Ritz crackers with peanut butter in the middle, dipped in white chocolate, sprinkles on top. Michelle raved about those cookies every year and shared exactly none of them. However many Katie gave her was however many she ate. Then came the last Christmas, when she wasn’t well, and she handed me one. I’ve made them every Christmas since, boxes of them, mailed to my parents, pressed on friends. I know an inheritance when I’m holding one.

November

Her last international trip was to Spain, with her friend Laura. They worked the harvest at Daphne Glorian’s vineyard, Clos i Terrasses, home of Clos Erasmus. When Michelle came home, she wasn’t feeling right.

Before another work trip, this one to Boston, she squeezed in a quick doctor’s appointment because she’d been having pain. The doctor wasn’t initially concerned, because Michelle was exceptionally healthy: constantly traveling for work, deep into her yoga practice, and at one point doing a hundred burpees a day. As the doctor was leaving the room, Michelle stopped her.

“But should it wake me up at night?”

The doctor turned around immediately and sent her for imaging that same day.

The next day, Michelle flew to Boston for work. When she landed, her doctor called and told her to get on a plane and come home immediately. Michelle said she’d be back Friday.

“No, Michelle. Now.”

That was the beginning.

In November of 2014, Michelle was diagnosed with gallbladder cancer that had already spread. The doctors gave her three to six months.

Here’s why her job breaks my heart. My sister underwrote risk for a living. She could read a table of numbers the way the rest of us read a room, and no loving conspiracy was going to keep her from reading her own. We conspired anyway. We surrounded her, filtered what came out of the oncologist’s office, kept the air full of plans, and in twenty months we never once discussed her dying. Not one conversation. She let us. She sang along with our hope like a song she didn’t quite believe the words to. But she knew the numbers. She always knew the numbers.

She tried everything, and so much of it came from our research together. Acupuncture, meditation, a rebounder she’d bounce on, fresh frozen aloe vera juice shipped up from Florida at great cost, which she drank down and absolutely hated. And yoga, which she loved. Near the end, when she was too sick and in too much pain to practice, her instructor would welcome her into class anyway, and Michelle would lie in child’s pose in the middle of the room and just breathe with them. She went anyway. At the end, that was the whole practice. Being in the room.

Every Appointment, Everyone

No one ever let her go to an appointment alone. Not a scan, not a treatment, not a single procedure, not once in twenty months. Kristin, Karen, Susan, Nancy, Donna, Laura, and more friends than I can name here rearranged their work and their travel and their own families to build a rotation so tight that Michelle was never the only person in a waiting room. My father helped pay for things quietly, the way he pays for things. Daphne Glorian, the same woman from that vineyard in Spain, donated the wine for one of our silent auctions, including bottles of Clos Erasmus that were extremely valuable, and the auction raised a significant amount to help with Michelle’s care. It came from the same vineyard Michelle had walked through months before any of this had a name.

Somewhere in there, someone applied for her disability, and it was approved on compassionate grounds, fast-tracked, everyone said. Fast-tracked still meant a six-month wait before a woman who had paid into that system her whole working life could receive a cent of it. A terminally ill woman, told that if she stayed alive long enough, she’d earn back what was already hers. The community covered the gap because the system wouldn’t.

At one point she had a liver-directed procedure, her artery blocked and radiation delivered straight into the tumor, somewhere in South Carolina. I still don’t know the exact name for what they did to her that day, and I won’t guess at it here. I know it was aggressive. I know she agreed to it without much hesitation. Later, her oncologist told us she’d done a great deal to further their research, because she tried everything anyone put in front of her.

Our entire lives became research. I didn’t know yet how long that would last. It never ended.

The Sofa

Her friends kept her covered in New York, when treatment took her there, and a childhood friend opened her home for the length of it. I flew to my mother’s house in Charlotte, where Michelle was living by then, and moved onto the sofa outside her bedroom door.

People assume I slept there so she could wake me if she needed something. That’s backward. I’d already lived through the morning we woke up and found her bright yellow, jaundiced, because her liver had stopped functioning. We rushed her to the hospital and they inserted a stent. After that, I wasn’t afraid she’d wake me. I was afraid she wouldn’t.

I slept facing her door for months. Some nights I woke up gasping from a dream that this was happening, and then woke up further and understood that it was. Her light would be on at three in the morning, a thin line under the door, and I’d lie there knowing she was awake too, and neither of us ever knocked. I couldn’t imagine what was happening on her side of that door, and I never found the words to fix any of it, because there were no words. I stayed anyway. It was the only thing I could do.

I cried plenty, just never where she could see me. You can’t cry in front of someone you’re trying to convince is going to beat cancer.

What I did instead, in those hours, and I’ve never told anyone this, is negotiate. I’m not religious. I’m Buddhist, and at three in the morning even that got very simple. If I have twenty years left, I said to whoever was listening, take ten of them and give them to her. I did the math. It meant we’d go out together, and I was fine with that. I’d had the marriage, the children, the whole ride, and I felt complete, and she hadn’t gotten her turn. It seemed like a fair trade. I picked the date it would have to happen and put it in my phone, not as a metaphor. It sat there for years, a private deadline nobody else knew about.

After she died, I kept the books on that bargain. I watched the calendar quietly, and when that stored date finally arrived, part of me expected not to wake up the next morning. I woke up fine. I’m still here, ten years later, healthy in body and mind, and I understand now that all my negotiating in the dark moved nothing at all except me.

Hold onto that. Later on, I finally went looking for what does move.

What Grief Does to a Body

By the end of that first year, I couldn’t cry without something else happening first. I held everything in until I was alone, and even then my body found a way to keep more of it than I could safely carry.

By the end of 2015, doctors diagnosed me with COPD, despite the fact that I was a non-smoker. I carried an inhaler in my pocket and used it several times a day, because I could lose my breath without warning. My mother and Michelle once had to get me to the hospital during an episode. My sister, dying of cancer, rushing me to the emergency room. Doctors told me it could eventually take my life if it didn’t resolve.

It resolved completely, roughly three years after Michelle died, and it never came back. Looking back, I don’t think it was COPD at all. I think it was grief. I’d held everything in for so long that my body finally just stopped letting me breathe.

What the Sick Are Owed

Chemo took most of her hair, and one night her friend Nancy and my mom shaved off the rest. If anyone was ever built to be bald, it was Michelle. She had a perfectly shaped head and she looked good, and none of that mattered, because it was never about how she looked to us. She was so thin by then. “I don’t know what to do with my skin,” she told me once, and I said honey, you look beautiful, and we tried everything, and some things can’t be fixed by trying.

Then there were the comments. A man she’d gone to school with reshared something she’d posted and announced to his audience that she was dying. She saw it. There were others, that she looked like a man now, that she looked eighty years old. She saw those too. She was forty-six, fighting for her life, grieving her own reflection, and people she barely knew were narrating her decline for sport.

So take one instruction from this if you take nothing else. If you know someone who is sick and dying, their body is not your material and their prognosis is not your announcement. It is not your right. Keep your mouth shut.

Folly Beach

Michelle found her peace near the sea. She always had. That July, Victoria was house-sitting in Charleston for her son and daughter-in-law, and Michelle went down anyway, chemo on Friday and the drive on Saturday, because she’d been looking forward to that trip for weeks. She cried almost the whole way down with Lori beside her. She already knew she was dying.

Lori photographed her at Folly Beach that week, and in those photographs you can see something I still can’t look at for very long: the moment she came to terms with where she was going. Not defeat. Surrender, which is a different thing entirely.

By then Michelle was actively dying. Her kidneys were shutting down, her abdomen was swollen, and there was no mistaking what stage this was. Victoria raced her back to Charlotte, the dogs still in the car, on speakerphone with emergency personnel the whole way. If Michelle’s heart stopped before they made it, the dispatcher said, pull over and call 911 immediately. When they finally reached the hospital, one of those dogs, Earl, bolted from the car, as if he’d read the chart.

She died on the morning of July 15, 2016, at 7:39 AM. She was forty-seven, and she’d been given three to six months and made it twenty, outlasting the outer edge of that sentence by more than a year, which, if you knew Michelle, was the least surprising thing she ever did.

Afterward, the family sat down with her oncologist, and he told us she had done a lot to further their studies, because she tried everything. I hold onto that sentence like a railing. Through all of it, the pain, what must have been fear, she never complained. Not once, not ever. She fought like she meant to win, every day, until the last one.

I wrote her obituary. Later, I commissioned the painter Carrie Jadus to make a large work from Lori’s photographs of that day at the beach, a tribute for my mother, with a print made for my dad. On the back I wrote that I hope everyone who sees it feels her extraordinary presence and remembers that love never dies. The artist worked three words into the canvas itself, words Michelle believed were powerful. You’ll have them before we’re done here.

The Neighborhood

Michelle lived at 2016 Thomas Street, and her presence was part of what made that block feel like a neighborhood instead of a row of houses. After she died, people up and down Thomas Street and Mimosa eventually moved away, one at a time, because being there without her cost more than staying was worth.

Donna and her husband Alvin were two of Michelle’s closest friends, not just her neighbors. They vacationed together, spent ordinary days together, and when Michelle got sick, Donna was there for nearly all of it, with Alvin right beside her the whole way, just as kind and steady as she was. Donna had her own cancer running underneath all of it the whole time, and she died on July 15 too, several years later. The same date, years apart. Some of us like to think of them together now, at some big table with good wine, their feet in the sand, waiting on the rest of us to show up.

For years, seeing the Charlotte skyline meant I was almost to her. I know that drive well enough to do it half asleep. Now it’s one of the hardest drives I make, because she’s not waiting there anymore.

What I Went Looking For

Our entire lives became research while Michelle was alive. Mine never stopped.

That’s how I became fascinated with artificial intelligence, of all things. I kept reading, kept asking people who understood it better than I did whether any of it could actually change how fast we find answers for people like Michelle. It became my work, and it was never about making my own life easier. It was about the chance that it might save someone else’s.

I kept reading about cancer. I found other families going through it and offered what I could, sometimes money, sometimes just time, and I trained as a grief coach so I’d have more than sympathy to hand someone, and I never charged for it, because I remembered exactly what I’d needed and hadn’t gotten.

Along the way I ran into people convinced that cures already exist and are being hidden from the public, and I ran into people inside the industry who reject that outright. I don’t want a conspiracy, and I don’t want false hope either. I wanted the truth, even an unflattering one, so this year, ten years out, I finally sat down to find it.

Here’s the honest version. As of late 2025, exactly three AI tools carry FDA clearance for cancer pathology, and all three are for prostate cancer. Not one touches the disease that took my sister. The AI-designed drugs, the digital twins that would model your treatment before you ever receive it, those are real research programs, and they are not real for patients yet. If someone you love is sick right now, AI is not the cavalry, and if Michelle were diagnosed today, artificial intelligence still wouldn’t change what happened to her.

But something else moved while I was watching the machines. In 2014 there wasn’t a single FDA-approved targeted therapy for biliary tract cancer, the family her disease belongs to. The map had nothing on it. Since 2020, that same disease has gained five approvals, treatments matched to the specific biology of a specific tumor. The gains are honest and modest, measured in months and not miracles, and the American Cancer Society still puts five-year survival for distant-stage gallbladder cancer at four percent. Four. That number should make you angry. It makes me angry.

And yet. In 2014 my sister’s tumor was a closed book. Today it would be sequenced, read, and matched against a shelf of options that didn’t exist when she needed them, and her oncologist told us her trying had helped build that shelf. I went looking for artificial intelligence and found something slower and more human waiting behind it: twenty years of unglamorous science finally reaching a disease that nobody funds.

So here’s the one practical thing I can hand you, and it isn’t a piece of technology. Say this out loud in an oncologist’s office: we want comprehensive genomic and biomarker testing. Insist on it. That’s where the new options live, and it doesn’t happen automatically everywhere. My bargaining in the dark moved nothing. Science moved. Slower than love needed it to, but it moved, and it’s still moving.

Perspective

I know families right now saying goodbye to children who haven’t reached double digits, and I know some of those parents would trade anything for forty-seven years. Grief isn’t a contest, and knowing that doesn’t make it any less of a shock to lose a healthy younger sister at forty-seven, with no warning and no history of illness before the pain that wouldn’t let her sleep.

Michelle’s forty-seven years were full. They were also nowhere near enough.

Nine Months

Michelle didn’t want children of her own, but she was a godmother, to my daughter Hilary. Nine months after Michelle died, on April 2, 2017, Hilary gave birth to a girl and named her Aubri Michelle.

Aubri came into the world very ill and spent the first six months of her life in the NICU, and her troubles were centered in the same part of the body that took Michelle. I believe Michelle sent her to us. Nine months, almost to the day. Grief is love with nowhere left to go, and here was this baby, given no good prognosis, who needed every bit of love we had left to give. We poured it all into her, all the love we still had for Michelle and nowhere to put. I know Michelle’s hand was on her the whole time, because the baby got the healing miracle Michelle never did. Babies with what Aubri had don’t always survive it. She did. Nobody expected the degree to which she’d heal, and I don’t think that was luck. Like Michelle, the baby was never alone either. We ran it in shifts, the same way we had for her aunt, and sat with her around the clock. We played her classical music. I’d talk to her tiny body, part by part, thanking whatever needed healing, telling them everything was going to be okay. And some nights, when it was just the two of us in that room, machines keeping her alive, I’d rock her and ask her, half joking and dead serious at once: is that you, Michelle? Did you have to come back with all these problems too?

On August 13th, Michelle’s birthday, with Aubri still in the hospital, I made three of those ugly strawberry cakes, underbaked on purpose, the way Michelle ruled was correct, and carried them in to the nurses on Aubri’s floor. It was the only place her birthday made sense that year. The nurses ate gooey cake at their station while a baby carrying my sister’s name fought on down the hall, and I understood that this is what remembering someone actually looks like. Not a candle. A recipe done wrong on purpose, fed to the people keeping watch.

We’ve never called her anything but the baby. Not Aubri. My parents say it, everybody says it, nine years old now and still the baby, and everyone in this family knows exactly what that name means. It means the gift Michelle sent us when we didn’t know how we’d survive losing her.

The Reading

A couple of days before what would have been Michelle’s fiftieth birthday, I scheduled a reading with a medium named Justin Terry, because there was one question I wanted to ask her. The first night, I made the mistake of writing down two initials, J and M, and a question for each, instead of focusing only on her. J came through immediately and took over the whole session, dominating it exactly the way he would have in life. Michelle just waited quietly. That was so like her that I should have expected it. I left devestated. I’d come there for her.

The next day, a cancellation opened up another appointment, and I went back. This time I wrote one thing on a folded piece of paper.

M

How do I honor you best on your fiftieth birthday?

The paper was never opened.

Justin struggled. He kept talking about Mickey Mouse, then stopped himself.

“No… Winnie the Pooh.”

He couldn’t have known that. Michelle’s nursery had been decorated in Winnie the Pooh from the day she was born, and my mother started calling her Pooh before she could walk. The rest of us called her Mishy Pooh. Ten years after her death, my mother still calls her Pooh.

Then came the answer to the only question I’d written down, in his exact words:

“Hold me in your heart. None of the other stuff matters.”

I’ve spent the years since trying to do exactly that.

What I Hope for Her Now

I don’t know what to make of moments like that, and I’ve stopped needing to. I’m Buddhist. I believe we come back, more than once, in different forms. Michelle kept small Buddhas around her house and did yoga most of her life, though I can’t say she believed exactly what I believe. I don’t know if she’s already come back. If she has, I hope it’s as a person, not as a puppy, and I hope somebody finally gives her the proposal she asked for and never got. I hope her next life runs long enough to make up for how short this one was, and I hope whatever comes next in medicine and in machines, the parts that couldn’t save her this time, get to protect her the next.

If I’m honest about what I’m really hoping for, it’s specific. I hope she gets the partner she deserved the first time, the one who proposes because he means it. I hope she gets a long life this time, not one cut off at forty-seven. I hope the medicine arrives before she needs it instead of five years after, and I hope whatever we eventually build with these machines is fast enough and good enough to help someone exactly like her, so some other family gets to keep their sister.

Michelle signed off on everything she wrote the same way. Every post, every update, all the way to the end. On the back of her painting, I wrote that I hope everyone who sees it feels her extraordinary presence and remembers that love never dies. The artist worked her own three words into the canvas itself, the same three she used to close everything she ever sent into the world.

Miracles, peace, and healing.

For every family sitting beside a hospital bed tonight, every caregiver sleeping on a sofa outside a bedroom door, and everyone who still smiles when someone says the name of the person they miss most.

About the Author

Melissa Tart is the founder of Farther North Studio. After losing her sister Michelle to gallbladder cancer in 2016, she spent the next decade studying grief, cancer research, and artificial intelligence, driven by one question: could the future spare another family what hers endured?

Today she helps women in mid-to-late career understand and use AI with confidence while continuing to write about grief, resilience, caregiving, and the human side of technological change.

If this essay resonated with you because you’re grieving, caring for someone you love, or trying to make sense of life’s hardest transitions, I hope you’ll know you don’t have to navigate them alone.

Read more essays: substack.com/@melissatart

Learn more or book a session: stan.store/MTart

For years, this family built its entire political identity on one claim: Joe Biden was too old, too slow, too far gone to lead. They demanded proof. They demanded cognitive tests. They ran ads about it. Now a sitting member of Congress has formally demanded a cognitive and neurological evaluation of Trump himself. Nearly a decade of psychiatrists, former staff, and a man who knew him personally for fifteen years have all said some version of the same thing. And the family that spent four years diagnosing Biden from a distance has gone quiet about the man they actually live with.

Except one of them hasn’t.

Mary Trump is not in the circle. She’s the daughter of Fred Trump Jr., the brother Donald spent decades making sure everyone forgot. She’s sued her uncle. She gets nothing from the name Trump except the weight of carrying it. She’s a clinical psychologist who watched her own grandfather disappear into Alzheimer’s, and she’s spent months now, on the record, on camera, saying she sees the same pattern in Donald. The deer in the headlights look. The loss of control over his own body. The decline she calls obvious.

The people who actually share his roof, his bank accounts, and his Cabinet table have said nothing. Melania hasn’t spoken. His sons haven’t either, while their net worth climbs alongside his, from roughly $2.3 billion to over $6 billion since he took office again, built largely on a cryptocurrency venture that pulled in $1.4 billion in a single year, on licensing deals in Saudi Arabia and Qatar, on an investment firm and a drone company chasing the same government contracts their father controls. The Cabinet sees him every day and knows exactly what the rest of us only catch in clips. They’ve stayed silent too.

Here’s the question a journalist is supposed to ask. What’s the payoff? What does silence buy you when you know the man running the country isn’t well? Proximity to power isn’t a neutral position. It’s a transaction, and every person around him who stays quiet while he threatens other nations, while he manufactures crisis after crisis, while his family’s fortune multiplies off the very office he holds, is choosing that transaction over the truth.

This isn’t a new playbook. It’s one of the oldest ones in the history of authoritarian power, and pretending otherwise is its own kind of denial. Hitler didn’t rise to power by convincing Germany he was strong. He rose by convincing a frightened, humiliated country that its problems had a face, and that face belonged to someone else. The Jews. The outsider. The group that didn’t really belong, no matter how long they’d lived there, no matter what they’d built. Once people believe that, they stop asking why their own leader is failing them, because they’ve already been handed someone else to blame.

Trump has run that exact playbook for a decade, and the target has always been the same. Immigrants aren’t a policy position for him. They’re a pressure valve. Every time his own competence, his own health, his own conduct needs a distraction, the enemy of the week gets rolled back out. Rapists. Invaders. Poison in the blood of the country, his words, not mine. That language isn’t accidental. It’s the language every scapegoating regime in history has used right before it stopped being rhetoric and started being policy.

Here’s the part that should scare people more than the rhetoric itself. This isn’t a leader in command of his faculties running a calculated, cynical strategy. This is a man whose own family history includes Alzheimer’s, whose own niece is a psychologist telling the country she sees the disease again, whose own former friend privately worried about his memory back in 2017, whose own former chief strategist believed he was already in cognitive decline years ago. A sick man pointing at outsiders to explain away a sickness nobody around him will name isn’t less dangerous than a healthy one doing it on purpose. It’s more dangerous, because nobody is steering.

That’s the comparison. Not that the atrocities are identical. Not that history repeats itself in exact costume. The mechanism is identical, a frightened public handed a scapegoat by a leader nobody in his own house will hold accountable.

Half the country is already saying this out loud. The people who haven’t said a word are the ones standing closest to him, the family that inherited his name and is currently getting rich off his office while he falls apart in front of the entire world. Melania. Don Jr. Eric. The Cabinet cashing paychecks tied to his signature. They know. They’ve always known. And every day they stay quiet is a day they choose their own comfort over the country they claim to love. No one owes him that silence less than the family wearing his last name and spending his fortune.

Melissa Tart is the founder of Farther North Studio, a coaching and educational platform for women in mid-to-late career who are navigating change, building what’s next, and learning to use AI with confidence in a rapidly changing world.

To learn more or schedule a Discovery Session, visit:
stan.store/MTart

Subscribe:
substack.com/@melissatart

A decade ago, that reaction would’ve seemed absurd. A family releases a photograph of an elected official, the public accepts that it documents the moment it claims to show, and nobody inspects facial contours, lighting patterns, shirt folds, shadows, fingers, or reflections before deciding whether a senator was actually in the room. We do now, and pretending otherwise would be foolish.

Common sense used to mean trusting your eyes. In 2026 it means knowing exactly how far your eyes can take you and refusing to travel one step past that point without evidence. That’s the skill this image demanded, and it’s the skill I want to walk you through, because I put it to work on this one myself.

First, the context that makes the skepticism rational rather than paranoid. We’re living in the age of Trump, when fabricated imagery is no longer confined to anonymous trolls or obscure corners of the internet. The current president persistently, consistently, and belligerently circulates memes depicting himself as Jesus, God, a conquering hero, or a physically fit thirty five year old bodybuilder. Those images aren’t presented as documentary photography, but their constant presence has normalized a political culture in which reality, spectacle, propaganda, parody, and fantasy are deliberately blended.

That has consequences. When a government, a political movement, or a public official treats manipulated imagery as a legitimate communication tool, the public eventually learns not to trust what it sees. Distrust stops being an irrational response and becomes a necessary part of processing information. That doesn’t mean every unusual photograph is fake. It means every important image now deserves scrutiny.

I hate conspiracy theories. I’ve never enjoyed constructing elaborate explanations around fragments of information, and I have no interest in deciding something sinister must be true because it feels suspicious. I prefer science, documentation, corroboration, and facts that can withstand examination. The problem is that we’re living in an upside down information environment where finding those facts increasingly requires ordinary people to conduct investigations that once would’ve seemed ridiculous.

So I conducted one.

Look at the two close ups side by side. The man in the hospital image looks strikingly different from the person photographed in a wheelchair not long before his hospitalization. His face appears narrower, his smile changes the entire shape of his mouth and cheeks, and the loose skin once visible beneath his chin seems dramatically reduced. Serious illness can transform a person quickly. A month in a hospital can take twenty pounds off a body, particularly when eating becomes difficult or the system is under enormous stress. Lighting, posture, expression, camera angle, and picture quality create their own distortions.

Even so, weight loss doesn’t erase skin. That doesn’t prove cosmetic surgery, digital alteration, or artificial generation. It creates a reasonable question, and a reasonable question attached to the health of a sitting senator deserves a better answer than one carefully framed picture.

The public shouldn’t have to perform this kind of examination to establish whether a senator is alive and capable of serving. A public servant isn’t a private celebrity whose representatives can release a single image and expect every question to disappear. When a person holds national office, the public has a legitimate interest in clear, current, and credible information about that person’s ability to do the job. Instead, Americans received one photograph and were left arguing over pixels.

Jimmy Kimmel made the problem impossible to ignore when he placed his own face onto the body in the hospital image. His version preserved the same room, clothing, posture, bedding, shadows, and surrounding detail. Only the face changed, yet the altered picture remained instantly recognizable as the same scene. He didn’t prove the McConnell image was fabricated. He proved how easily one element of a convincing photograph can now be replaced while everything around it stays visually coherent.

That’s why appearance alone no longer settles anything. The old advice about spotting artificial images focused on malformed hands, extra fingers, mismatched earrings, warped teeth, unreadable signs, or objects melting into one another. Those clues still surface occasionally, but the technology has moved past them. A picture can contain natural lighting, plausible shadows, correct anatomy, realistic fabric, and a believable background while still carrying a generated or replaced face.

So I stopped studying faces and went looking for a detail I could verify.

The newspaper beside McConnell was the only element in the frame that existed independently of him. I wanted to know whether it was a genuine edition, whether the visible material matched an actual publication, and whether the date corresponded with July 12, 2026, the day the photograph was released. I ran the question through three deep research systems and compared their findings rather than trusting a single answer: Claude’s Fable 5, Perplexity, and Gemini’s Deep Research.

All three reached the same conclusion: the visible edition matches a genuine newspaper published on July 12, 2026, the same day the image went public.

That matters, and it doesn’t settle the larger question. A genuine newspaper can appear inside an altered picture. It can be photographed separately, inserted digitally, or sit in an authentic scene where another element has been changed. Verifying one detail strengthens the case without proving that every part of the frame is real. That distinction is where common sense earns its keep.

Due diligence isn’t the search for an answer you like. It’s identifying what can be confirmed, separating it from what remains unknown, and refusing to close the gap with speculation.

Sometimes the responsible conclusion is incomplete. Mine is.

Here’s what I can say. The newspaper checks out. The setting contains no visual defect that definitively exposes generation. McConnell’s appearance differs sharply from his recent public appearances, and illness, weight, lighting, and expression could explain some of that change, possibly even all of it. Kimmel’s demonstration shows how little visual coherence now tells us about the authenticity of a face. What I can’t say, with any confidence, is that the original photograph is genuine. There isn’t enough independently verifiable information to reach that verdict, and there isn’t enough evidence to declare it fake either. Anyone claiming certainty from these images alone is claiming more than the facts support.

That’s the larger lesson. AI fluency isn’t blind faith in technology, reflexive suspicion of every image, or the belief that a chatbot can deliver a final ruling. It’s the ability to slow down, compare sources, test specific details, recognize the limits of the evidence, and stay comfortable with an unresolved result. In the age of Trump, with a president who’s made synthetic political imagery routine, an internet that rewards outrage before verification, and tools capable of manufacturing convincing evidence in seconds, that discipline is no longer optional.

Common sense now comes with homework. I did mine, and the honest answer at the end of it is that I still don’t know.

Melissa Tart is the founder of Farther North Studio, a coaching and educational platform for women in mid to late career who are navigating change, building what’s next, and learning to use AI with confidence in a rapidly changing world.

To learn more or schedule a Discovery Session, visit:

stan.store/MTart

Subscribe:

substack.com/@melissatart

Nobody saw it.

That’s not unusual. I spend hours writing about AI, coaching, politics, cancer research, and other subjects that matter to me, and the result is usually the same.

So why would I keep bending myself into a pretzel for an algorithm?

People aren’t algorithms. We’re collections of interests, memories, skills, and curiosities. We can spend one hour talking about artificial intelligence and the next admiring a full moon, baking sourdough, swimming a mile, restoring an old chair, or carving a tiki from a fallen tree. None of those things compete with each other. Together, they tell the story of a life.

I’m done trying to make myself easier for an algorithm to categorize. I’m a multidimensional human being, and that’s what you’ll get from me. If the algorithm doesn’t know what to do with that, it’ll just have to stay confused.

Nobody signed us up for that, and four or five hours on a Monday morning can’t undo years of habits, no matter how hard we act like they should. We treat one square on the calendar like it owes us a complete overhaul, then we’re stunned when Monday shows up looking exactly like Sunday did, only with more emails.

Here’s what nobody wants to admit: Monday isn’t doing this to us. We are. We spend Sunday night writing an impossible script in our heads, casting ourselves as the disciplined, put together person we wish we were. By nine the next morning, we’re grading ourselves against a character we invented the night before. That’s not a scheduling problem, it’s a target problem.

Reinvention is too big a goal for a Monday, let alone the first four hours of a week.

Confidence doesn’t come from an overhaul. It comes from a finished task, a habit you actually kept, an uncomfortable choice you finally made instead of putting it off again.

This is where I tell my clients to start small enough that it feels almost silly. Don’t promise yourself you’ll master AI this week. Open one tool you’ve been avoiding and ask it one real question. Hand it the email you’ve been putting off, or the calendar mess you’ve been ignoring, and see what comes back. That’s movement, not mastery, and movement is what actually compounds.

Stack enough of those small moves and the whole picture shifts. You stop white knuckling through Monday mornings and start handing off the parts you used to dread, one delegation at a time, while you get your life back.

Monday doesn’t need to be your personal revolution. It can be Monday, nothing more than that. One step taken on purpose beats a heroic plan written at eleven o’clock on a Sunday night, because that plan never survives contact with an alarm clock.

Melissa Tart is the founder of Farther North Studio, a coaching and educational platform for women in mid-to-late career who are navigating change, building what’s next, and learning to use AI with confidence in a rapidly changing world.

To learn more or schedule a Discovery Session, visit: stan.store/MTart

Subscribe: substack.com/@melissatart

For the first time in my life, the Fourth of July doesn’t feel like a holiday. It feels like another reminder of how far we’ve drifted from the country I grew up believing in.

White nationalists are comfortable enough to march through the nation’s capital with their faces covered. Convicted environmental offenders walk away with presidential forgiveness while the planet keeps getting hotter. Americans are expected to nod along as obvious excuses replace obvious truths.

The biggest story isn’t even the parade or the pardons. Americans are being robbed blind while they’re told everything is going great. I’d argue 70 percent of the country already sees it. The other 30 percent probably never will.

I spent twenty years living near Washington. I loved watching families gather on the National Mall. Tourists came from every corner of the world because the place represented something larger than politics. Looking at those same streets now, I barely recognize what they’ve become.

My childhood memories haven’t changed. Swimming all afternoon. My dad making homemade peanut butter ice cream. Fireworks lighting up the sky. Those moments belonged to an America that believed character mattered.

That’s what I miss.

Real patriotism has nothing to do with wrapping yourself in a flag while excusing corruption, rewarding extremists, or treating the Constitution like an inconvenience. Loving this country means expecting better from the people entrusted to lead it.

I’m not giving up on America.

I’m giving up on pretending this is normal.

More often, it shows up as a joke that lands wrong, a casual swipe at “those people,” or a flash of contempt dressed as common sense. It passes quickly enough to ignore, especially when a working relationship looks promising.

And in business, there’s always an incentive to ignore what unsettles you.

The client is ready. The money is real. The project looks good on paper. A partnership could open doors. The referral came through someone you trust. The conversation has been pleasant enough. So you start negotiating with your discomfort before the contract is even signed.

I’ve done that more times than I’d like to admit.

Somewhere in the last several years, I stopped.

I run a coaching practice for women rebuilding their lives in mid-career and beyond. My work isn’t partisan, and it was never designed to be. I help women navigate change, confidence, reinvention, technology, identity, and whatever comes next when the old life no longer fits.

Still, there’s one category of person I’ve found myself unable to work with, regardless of how good the fit appears at first glance or how much revenue sits on the table.

If someone is still standing behind MAGA in 2026, actively and proudly, I can’t get past it.

I’ve tried.

I’ve argued with myself about whether that makes me intolerant, rigid, unfair, politically narrow, or bad at business. I’ve reminded myself that a coach is supposed to meet the person in front of her. Then I sit across from someone whose politics are not merely “politics” in any abstract sense, and my body registers the truth before my business brain can finish its defense.

Something in me closes.

That response used to embarrass me. Now I understand it as information.

For a long time, many of us were taught to treat politics and character as separate rooms in the same house. A person could vote one way, behave another, and still be fundamentally safe to trust. Business relationships, we were told, could float above the moral weather.

I don’t believe that anymore.

This isn’t about tax policy, zoning debates, or a normal disagreement over how to solve a shared problem. I know how to work with difference. I have no interest in building a life where everyone thinks exactly like I do.

But political disagreement and moral incompatibility are not the same thing.

A disagreement asks, “Which solution do you believe will work?”

An incompatibility asks, “What have you decided you can excuse?”

That’s the line I keep returning to.

When someone tells me they’re still MAGA, especially now, I hear more than a voting preference. I hear tolerance for cruelty toward people with less power, indifference to truth when reality becomes inconvenient, and permission for grievance, domination, humiliation, and denial to masquerade as strength.

Those aren’t merely political positions to me. They’re character signals clustered around a political identity.

Once that information enters the room, I can’t pretend the foundation is solid enough for trust.

This is usually where someone says, “Isn’t that divisive?”

Maybe. But pretending values don’t matter becomes its own form of division when the cost lands on your peace, labor, judgment, and self-respect.

The harder question for me is no longer whether this line is fair. I’ve stopped trying to prove my discomfort to people who are committed to misunderstanding it.

The question is practical: how do you keep from getting entangled before the truth becomes obvious?

Because the clean refusal rarely does the most damage. The real cost comes when the misalignment reveals itself too late.

You’re three weeks into a project, two months into a partnership, already visible in someone’s network, or emotionally invested in a client’s growth. Then a post surfaces. A cruel joke slips out. A conspiracy theory enters the conversation as if it’s harmless. Someone complains about “wokeness” and suddenly reveals contempt for the exact kind of person your work exists to protect.

At that point, you’re not declining a relationship. You’re extracting yourself from one you’ve already entered.

That kind of exit costs more than money. It disrupts momentum, strains your nervous system, and weakens your trust in your own discernment. Worst of all, it reminds you that part of the damage came from overriding what you already sensed.

So I built a filter.

Not a political test. Not a demand for perfect ideological alignment. Not a purity ritual dressed up as strategy.

A values audit, early enough to matter.

Before the invoice, contract, calendar hold, or emotional labor begins, I pay attention to the signals people usually reveal without being asked.

The way someone talks about people who disagree with them tells me how conflict will feel later. Contempt has a sound, and it almost always appears in language before it shows up in behavior.

I notice how they describe people with less power: employees, service workers, strangers online, anyone who can’t benefit them. Character is least rehearsed when there’s nothing to gain.

Humor matters too. If the joke depends on someone vulnerable being humiliated, I don’t treat that as a personality quirk. I treat it as a preview.

I also watch how a person handles correction. Defensiveness under mild pressure becomes a forecast for what they’ll do when the stakes are higher.

Admiration gives people away as well. Some are drawn to courage, honesty, and repair. Others respond to dominance, spectacle, and control. Those are very different moral ecosystems.

None of this requires an interrogation about voting records. I’m not asking clients to submit a political résumé before we begin. I’m asking myself to stop discarding the evidence already in front of me.

That’s the part I care about now.

The danger isn’t only the slogan, candidate, or public performance of allegiance. It’s the private bargain I used to make with myself when I wanted the project, income, approval, ease, or validation of being chosen.

The minute I start negotiating against my own discernment, I’m already paying for the relationship.

Some people will call that closed-minded. Others will insist business should remain open to everyone. Plenty still believe professionalism means sanding yourself down until nothing you believe creates friction.

I don’t accept that definition.

Professionalism isn’t self-abandonment. Inclusivity doesn’t require unlimited access to me. Discernment isn’t cruelty, and a boundary doesn’t become a tantrum because someone dislikes where it sits.

I believe every woman deserves support, dignity, and a coach who sees her fully. I also believe I’m allowed to protect the room I’ve built.

Those beliefs don’t cancel each other. They make the work possible.

The room has a charge. Its assumptions shape the work. Its safety determines what people are able to say. My ability to coach honestly depends on not betraying myself at the door.

That means my values need to be visible enough for the wrong people to self-select out.

I don’t need to plaster my business with political declarations or turn every post into a manifesto. But I do need to be clear about what my work serves: women’s autonomy, truth-telling, reinvention, dignity, courage, and a future that doesn’t require us to shrink.

That clarity will repel some people. Good.

A business that attracts everyone eventually attracts people you’ll have to recover from.

I’m curious how many other coaches, consultants, founders, creatives, and business owners are doing this same moral math in private. I suspect there are more of us than anyone admits, because saying it out loud invites the usual accusations: divisive, difficult, unprofessional, too political.

But underneath all that noise, the real question is simple.

Could I work closely with someone whose worldview requires my silence in order to keep the peace? Would taking their money mean ignoring the contempt they carry for people I care about? How long can anyone call it “just business” when the body already knows better?

For me, the answer is no.

I don’t need everyone to agree with me. I do need to trust what someone has chosen not to excuse.

That’s the moral math I can’t make work anymore.

If you’ve felt that same cold, quiet closing and wondered whether it made you difficult, maybe that was the wrong question.

Perhaps you were paying attention. Perhaps your body recognized what your business brain wanted to negotiate. Perhaps moral clarity begins the moment you stop confusing access with generosity, tolerance with integrity, or silence with professionalism.

Draw your own conclusions.

I’ve drawn mine.

Melissa Tart is the founder of Farther North Studio, a coaching and educational platform for women in mid-to-late career who are navigating change, building what’s next, and learning to use AI with confidence in a rapidly changing world.
To learn more or schedule a Discovery Session, visit: stan.store/MTart
Subscribe: substack.com/melissatart