I am quite proud of myself because I chose to come here instead of chasing a “first date” with a young, attractive girl from Tinder. I figured this would actually be more interesting, and it has been.

Read more

“Come on, come on,” he said, tapping the steering wheel and glancing to his right at a car as he passed, hoping to catch the driver’s eye for no other reason than to distract himself from the fact that he would be late once again.

The other driver looked back at him through aviators. The two of them looked back at the road, glad they had accomplished this meaningless vestigial ritual of threat dissipation.

The trouble was that Garrett’s first day of training was a wash, and he was sent home early because he didn’t bring his passport. Sapporo-Stone couldn’t possibly progress with day one of training without a passport or other secondary identification on file, so they sent Garrett home. He missed the history lesson on the origins of Sapporo brewing in Japan and its 2022 acquisition of Stone Brewing. He missed it because he didn’t have his passport, and now he would be late to day two of training, but if he wanted to, he could keep driving all the way down to Tijuana and along the rippled ridgeline of the Pacific’s Ring of Fire as far as his car could make it, because now he had his passport.

“Do I even really need this shit?” Garrett wondered aloud. He already knew the answer to that question. Yes, he did need a second bar job, because President Trump had declared war on Iran, and Iran had closed the Strait of Hormuz, and now gas was $6.50 a gallon, and Garrett lived thirty-two miles north with his friend, because gas is cheaper than rent.

He was drawn to the new job partially because of the brewery’s design. It was an event space, with a koi pond inviting waterfowl in the expansive back patio, where sunlight dappled quarried stones and massive ferns. The interior resembled a Gothic castle, with dramatic red drapes to match the medieval theme depicted in Stone’s gargoyle logo. It strangely reminded Garrett of his church growing up, which he found funny. Perhaps the brewery’s creators found their life’s passion for alcohol an affront to God, and eschewed a certain religious upbringing with their demonically themed branding. In any event, Garrett found it inviting.

Rushing into a parking spot and flitting indoors, he saw that he was only two minutes late. He burst into the open concept office area, announcing his arrival to the first Sapporo-Stone representative with whom he made eye contact.

“It’s a bit of a crazy morning, just letting you know at the front end,” Alexa responded to an obviously flustered Garrett.

“That’s fine,” he said.

“Mark will tell you more. There are some upper level changes happening, but nothing will really change for you.”

“Ok,” Garrett said, still trying to get his bearings as they paced through doors and around massive drumwells.

“Did Mark ever circle back with you about the documents?”

“Yeah, he circled back,” Garrett said, making a little joke to himself. It was sort of funny to have such corporate speak at a bar. But then again, he had to get this idea out of his head that this international brewery/restaurant/event space was a bar. They entered the main area where a few people were setting up for the pre-shift discussion.

“I’ll drop you here,” Alexa said. “Have a good shift!”

“Thank you!” Garrett called out to her back. “Hi, Mark,” Garrett said, extending his hand.”

“Garrett, good to see you. Everybody, welcome Garrett, he’ll be training at the host stand today.” Garrett waved politely at the others as they gave him names one by one that he immediately forgot. They all shuffled into their chairs and passed out menus as Mark continued.

“Ok, we’ll go ahead and get started. So…Sapporo sold off Stone Brewing last night. It was 11 PM our time, because the company is traded publicly, and that’s when the sale became public in Japan. So myself and the other managers are just finding out about this, pretty much at the same time as all of you.”

“Oh,” Hannah blinked, one of Garrett’s new coworkers.

“Don’t worry,” continued Mark, “this shouldn’t affect any of you.”

“Shouldn’t?” Hannah said nervously.

“They said they wouldn’t mess with our hires. Anyways, we’re safe here at least. The Escondido location, on the other hand, will shut down in November.”

“Not my Escondido!” said Jared from the other side of the group, half joking.

“Yeah, they don’t need the distribution,” Mark replied.

“Will they keep Richmond?” Hannah asked.

“Yes,” Mark responded. “We will represent the West Coast, and the Richmond brewery will represent the East.”

“Who the hell is buying it, though?” said Jared.

“Firestone-Walker. From Los Angeles.”

Mark continued to go over the details of this transition, none of which mattered all that much to the lowly bistro crew members he was talking to. Outside, the ducks waddled around the back patio, expectant that bits of food would soon drop from the sky once again. It was a clear clear day, like every San Diego day in March. A plane took off from the nearby airport, passing low overhead and heading for Los Angeles. The ducks at Stone Brewing waddled calmly below the plane’s mechanical turmoil.

One hundred and thirty-seven miles north, flight UA 4643 landed at LAX. A singer named Sabrina Carpenter took notice of the incoming flight as she left the offices of Universal Music Group, where she met with executives about the particulars of her contract. As she got into her car, the UMG executives in Los Angeles called their corporate headquarters in Hilversum, Netherlands, for their weekly meeting. Their voices traveled across the Atlantic seafloor in fiber optic cables humming softly in the darkness.

At that same moment, Sabrina Carpenter’s voice burst out of sound systems around the globe. Her face lit one of the many screens at Shibuya Crossing in Tokyo, where the day before, Hiroshi Tokinatsu heard her song as he pulled out of Yebisu Garden Place, 4-20-1 Ebisu, Shibuya-Ku, Tokyo, 150-8522, Japan, where he oversaw the public sale of Stone from Sapporo to Firestone-Walker. Just after the sale, Tokinatsu gave a courteous conference call to Adam Firestone in Los Angeles and Stone Brewing CEO Zachary Keeling, who was in Richmond, Virginia.

Back at Stone Brewing, in San Diego, California, Mark and the other managers paced around their second story office overlooking the back patio and the main dining area, worried for their jobs. Hannah looked up at their figures animating the windows above the yawning fireplace and between the grimacing gargoyles.

“How did Sabrina Carpenter get on our playlist?” Jared asked, pointing to the speakers above as he led Garrett across the hall.

“Not really the vibe,” Garrett agreed, smiling at Hannah as they passed the host stand.

“This is where you’ll seat people first,” Jared said as he and Garrett crossed into the back patio. Garrett nodded, looking down at the ducks waddling up to him.

Surfing has taught me many life lessons. It has also taught me a lot of science about a largely unknown realm of the world. On a given day going out to surf, I will be reading the coalescence of many natural processes as I read my own life. Here is an effort to transmit that knowledge and advocate for letting the non-profitable things do the building.

[It starts with preparation]

The night before and leading up to the session, my friend Ethan and I will look at the surf report. Surfline is the premier surf forecast app, using live video camera feeds of the waves, buoy reporting, and daily editorial from professionals. It has come a long way since 1985, when it first relied on eyewitness accounts of the surf recorded to a local phone number. Surfers would call the number corresponding to their local surf spot and get a recorded message. I can only imagine what tomfoolery a surfer would put in their report.

Now, Surfline uses National Data Buoy Center data and adds their own mathematical models to it, parsing apart swells to give a more comprehensive look at what to expect at a given break. But we’re getting ahead of ourselves. First, we need to look at the different components that make a surfing session.

Global Winds

[You can’t make the waves come. Be patient]

Nobody reading the surf report is talking about global winds. But the winds cause the swell, and reveal the connection between the sun and the waves. All of this information will build on each other and become pertinent to a surfer in delicate ways.

Because of Earth’s tilt and round shape, the sun heats up different longitudinal zones unevenly. Hot air rises at the equator, where the sun’s rays strike the earth most directly. This hot air then cycles north (if you live in the northern hemisphere), where it cools and sinks down to Earth again. As the air moves from the equator, it curves because of the earth’s rotation, producing the Coriolis Effect. This divides the rotating air above the earth into three cells for the Northern Hemisphere and three corresponding cells for the Southern Hemisphere.

At the earth’s surface where each of these cells share a boundary, the cycling air causes ocean water to pile up at a small but significant height of about one meter. This height differential creates a pressure differential in the water, which, combined with the Coriolis Effect, causes the ocean water to rotate in prevailing surface level currents, divided within oceans by the air cells above Earth’s surface.

The wind is calmest at sunrise, then picks up as the sun heats the air, causing it to cycle in each of its cells. These daily winds are often called the trade winds, as they were discovered by European traders for their consistency when crossing the Atlantic Ocean. Between 30 and 60 degrees latitude are the Westerly Trades, meaning the wind will come from the west. Between 0 and 30 degrees latitude are the Easterly Trades, as each air cell is divided by 30 degrees of latitude.

The Swell

[Fight for what you love]

The waves start with storms thousands of miles away. These storms are themselves the result of pressure differences between cold, dense air and warm, light air. Clouds build when evaporating warm water rises and condenses, creating a storm cell. As the pressure difference between warm and cold air increases both within a storm cell and in global wind patterns, wind in the storm system whips to higher velocities, pushing the surface tension of the sea to create pressure differentials on the surface of the ocean. High velocity sustained winds will create a wave. As the wind pushes the wave, it increases the wave height on the windward side and creates a pressure deficit on the leeward side, increasing the wave height and thickness even more.

In the area under a storm system, or the fetch, the waves are disorganized, and the water is very choppy. Further away from the fetch, the waves begin to organize themselves according to wave speed. Waves of similar speed will travel together in sets, with faster sets overcoming slower ones in the open ocean. The time between different waves in a set is called the period. By now, you should be getting flashbacks to trigonometry. These waves travel sinusoidally, and these terms are consistent with your days in math class.

The speed of a wave is dictated by its orbital energy under the surface. Orbital energy travels like a stack of rolling wheels, with the top “wheel” dictating the wave period. The top wheel of orbital energy will turn a smaller wheel below it, on down until the energy dissipates to nothing.

When the wave travels to shallow enough water, the friction from the seafloor will start to slow these orbital energy wheels, decreasing the wavelength between wave peaks. The period, however, will remain the same, dictated by the “source frequency” of the wind in the originating fetch area. With the period consistent and the wavelength decreasing, the wave height will increase in shallow water. The bottom “wheel” on the orbital energy stack will drag on the ocean floor, lagging behind the top wheel, causing the wave to lean forward. When the ocean depth to wave height ratio is 1.3:1, the top of the wave will topple over, and the wave will break. This means that bigger waves will break further offshore.

On any given day, you are probably surfing where there are multiple swells arriving. After all, storms are happening all the time, all across the ocean. This is where Surfline’s technology comes in handy, allowing you to see the direction, period, and height of the three largest swells at your surf spot. One swell will almost always dominate the others, but reading how the three will affect each other is a bro science that is best interpreted by locals who know the sport and know the spot.

The Tide

[Learning something new requires struggling with monotony]

Of everything going on in the ocean making waves, this is the most familiar and consistent. You may know that tides are caused by the gravitational pull of the moon. You may know that there are always two high tides on opposite sides of the earth, both in line with the moon. The tide on the opposite side of Earth happens because the earth is pulled towards the moon. This means that in a given day, one location will experience two high and two low tides as the earth rotates through these variations in gravitational pull from the moon.

These tides affect the incoming swell, mainly through the wave height and shape. In a rising tide, the tide pushes in the same direction as the swell, adding a very small amount of momentum to the waves. A rising tide will also decrease the effect of friction from the ocean floor on the wave, making the wave break closer to the shore. At high tide, the waves will be fatter and heavier, meaning they are likely more mushy when they break. This will sometimes look like a wave slowly crumbling from the peak. In a falling tide, the waves will become more “hollowed out,” giving the waves that curl shape that can sometimes turn into a barrel, if other conditions are right.

The Current

[If your wave passes, there will be another]

As discussed above, global winds will cause prevailing surface currents in each ocean. Surface currents will glide in a particular direction, pushing layers of water beneath them up to 400m deep. Here, the laminar quality of water produces these effects much like the dissipating orbital energy wheels that produce a wave.

Surface currents will also occur when waves naturally contend with the shore. Water that reaches the shore travels parallel to the shoreline before cycling back out to sea. The current will produce a rip, or travel out to sea, where there is a sudden deep channel in the seafloor. Surfers with an understanding of their spot’s contours will often paddle out at the riptide, making their journey to the set much easier.

Deep ocean currents, operating like wind patterns above, occur when colder water cycles with warmer water. Cold water can freeze, locking away freshwater, making the seawater saltier and more dense. That dense, cold water will sink and cycle through with warmer water. This creates what is called the global conveyor belt of alternating warm and cool water traversing the globe and carrying aquatic life all around, supporting food chains.

At the surf, a strong current can ruin a spot. The current can pick up on a given day due to the wind or other circumstances, moving you off of the correct position and even carrying you out to sea. Sometimes it is not worth the energy to fight the current and stay in position in the 15 or 20 minutes it might take for a good set to come. Other times, the current might drag you to a dangerous spot, where if you catch a wave, you might unexpectedly head into some rocks or even an ocean pier.

Bathymetry

[Beauty is unforgiving]

As a storm moves across the ocean, its path will dictate the swell direction when its wave trains intersect with a given shoreline. Swells approaching perpendicular to the coast will produce the largest waves with the most energy, but this doesn’t always mean the best waves. Some spots, like Pipeline, require a perpendicular swell with a lot of energy in order to break just right. Others may benefit from a swell that approaches from an angle, and the drag from the spot’s seafloor geography, or its bathymetry, will bend the waves in toward the coast. The beauty of surfing is that you need to rely on local knowledge to understand what is coming from thousands of miles away, what is going on below the surface, and how it might turn into a session on a given day.

Broadly speaking, any surf spot is either a reef break or a beach break. Reef breaks are when the seafloor is made of rocky material, and beach breaks are made of sand. Reef breaks will therefore be very consistent, and beach breaks will sometimes shift, especially when a swell too big for the supporting beach break will shift the sand on the seafloor.

The quality and nature of the reef will determine how the wave breaks. A gradually rising seafloor will often cause waves to break gradually, crumbling at the top. A seafloor that rises sharply will cause a dramatic change in the wave energy, making the wave more likely to barrel.

The bathymetry of any surf spot will be entirely unique. Some spots will create a reverberating effect, where the waves bounce back from the shore and contend with the otherwise organized waves coming from the ocean. Other spots seem to have been shaped by natural processes for millions of years just for surfing. People will travel halfway around the world to surf legendary spots that break so perfectly because of the sudden seafloor rise in a shallow reef. Teahupo’o Tahiti, for example, has a menacing look because of the dramatic rise in the seafloor that allows incoming waves to break with all of their energy intact.

Pipeline in Hawaii similarly has no continental reef, making for a world-famous challenging wave. Adding to the mystique is its characteristic bathymetry, with the Ehukai channel carved into the seafloor from a mountain stream dumping into the ocean. The Ehukai channel slows down the wave as it peels off, ending the barrel and allowing surfers to exit the wave without being crushed by a wall of water. But Pipeline is famously treacherous, with a very shallow reef pockmarked with lava tubes that have snapped the limbs of more than a few surfers.

The Wind

[A mouthful of seawater is as hard to beat as a ten foot wave]

The wind creates storms and swells that eventually create waves at the surf spot, as we discussed above. However, the wind at the surf spot on the day of the session will also have a massive effect on the quality of the session. A windy day will create choppy surf, which is difficult to paddle out in, and might even make you feel seasick as you bob on the ocean surface. If it is choppy and you do make it out to the set, you might have to navigate through unpredictable “speed bumps” in the wave when you do catch one. The wind can also disorganize the waves once again, or create small local waves that contend with the incoming sets, ruining the surfable waves with unsurfable ones.

Gentle offshore winds, or wind coming from the coast to the ocean, can be favorable for a surfer as they hold the wave open, blowing into it as it breaks. Onshore winds, or winds coming from the ocean to the land, will “close out” waves, meaning that they will make the waves break all at once, rather than peeling off nicely. When a wave closes out, it is basically unridable, as you often can’t get ahead of the turbulent whitewater like you can when a wave breaks first at its peak, then evenly along its length.

Putting it all together

[Communication and respect can clear up almost any offense]

Each of these components that create a surf session also aid the interconnected web of life throughout the planet. The tides create a biodiverse intertidal zone, which is rife for aquaculture. Farmers in the Puget Sound of Washington State raise plants and animals such as Geoduck clams in the zone that is underwater at high tide and exposed to air at low tide.

Energy companies hope to harness the energy naturally produced by tides for human use. Companies have recently started experimenting with placing turbines in the water, which convert the rising and falling tides into electrical energy. These machines do affect the tides, which begs the question about how much we are at odds with the earth in our conversion of its natural processes into industrialized uses. Perhaps the best use of the ocean’s energy is in surfing, which energizes the earth’s most productive species to GENERATE MORE GDP. I would prefer if we spent more time re-learning to play, but we do need clean energy, so the jury is out on that.

The Surf Report

[The best route isn’t always as the crow flies]

The technology in tracking these processes is extremely sophisticated. Below, I will include screenshots of everything visible for the surf report at a specific location. Most valuable is the live feed of the waves, on offer from Surfline’s 1000+ global cameras. Below that, you will see a bar graph of peaks recognized by Surfline’s recent AI technology. The yellow columns on that bar graph are waves that the AI recognizes someone surfing. In the next picture, we see a graph of how many waves of a certain height came through today, and how many of those waves were surfed. Below that is a general condition rating based on Surfline’s LOTUS buoy forecast, which I mentioned at the top of this article. Surfline’s experts will manually update this condition rating at times as well. The third picture includes the surf height (whether or not it was observed on-site), the wind speed and direction (whether or not it was observed on-site), the three prevailing swells and their height, periods, and directions, and a map of those swells. The following picture will show you the tide and water temperature, followed by a written report from a Surfline expert. After that is how the surf rating is expected to progress through the day, as well as graphs of the surf height and wind. Then we get more developed analysis on the tide, and how much energy each wave has, which is an x-factor that you will feel when surfing.

Surfing is a very visceral confluence of just about every natural system. In no other sport are you so immersed in nature, and fully feeling the result of a delicate balance of global and local patterns. It is astounding to see how interconnected everything is.

I am currently reading the biography of Leonardo Da Vinci by Walter Isaacson. In it, we learn about Da Vinci’s obsession with vortices and spirals throughout his art and inventions. It was his trademark in paintings, focusing on hair curls and water eddies with extreme attention. As a student of life, Da Vinci discovered that the heart functions as a vortex, pumping blood in a spiral through the body. One could even say that Da Vinci’s genius was seeing the vibrational patterns of the world as a system of vortices. Da Vinci would have loved surfing, with the curling sinusoidal wave propagating thousands of miles across the ocean, the end product of vortices of heaving cloud masses miles above the earth, churning the surfaces of the water below in currents that spiral off in eddies nourishing ocean life as the earth follows the sun corkscrew-like through space. We should all fall in love with something so natural and so beautiful.

Read more

Unless otherwise specified, all posts are from 1440, a daily newsletter. None of the words below and in this series of newsletters are mine, I am simply copying and pasting from a few different newsletters, so I do not get any writing credit.

9/1

> Chinese law requiring clearly identified labels on all AI-generated online content goes into effect; government says effort is meant to combat misinformation (More)

9/3

> Anthropic closes $13B funding round at a $183B valuation, roughly triple what the AI startup was worth during its last raise in March (More)

9/8

> Warner Bros. sues AI startup Midjourney for use of images of Batman, Superman, and other intellectual property; follows recent $1.5B settlement by Anthropic over pirated books for chatbot training (More)

9/9

> OpenAI to back first full-length animated film for theatrical release; “Critterz” targets 2026 debut at Cannes Film Festival (More)

> SpaceX enters $17B deal to buy spectrum licenses from EchoStar to enhance Starlink’s 5G connectivity; EchoStar shares rise 19.9% on the news (More)

9/11

> Amazon launches electric robotaxi Zoox with free rides in Las Vegas; the Waymo and Tesla rival plans to enter the Austin and Miami markets next (More)

> US stock markets close mixed (S&P 500 +0.3%, Dow -0.5%, Nasdaq +0.0%) (More) | Oracle, OpenAI sign $300B, five-year cloud computing deal; Oracle shares rise nearly 36% in best day since 1992, founder Larry Ellison tops Elon Musk as world’s richest person (More)

> Does AI affect how we learn?

9/12

> Federal Trade Commission orders Alphabet, Meta, OpenAI, and several other AI chatbot makers to disclose information on products’ effects on children (More)

> Albania appoints AI bot as cabinet minister to tackle corruption.

9/13

> AI startup founders sacrifice fun, sleep for chance at success.

9/14

Newsom Eyes AI Bill

California lawmakers passed a landmark AI safety bill during the early hours of Saturday morning, sending the measure to Gov. Gavin Newsom (D) for signature. The package, set to go into effect Oct. 12 pending Newsom’s approval, is a pared-back version of a similar proposal he vetoed last year.

Home to more than 30 of the world’s top 50 AI firms (see list), regulations in the state often have nationwide reach. The bill’s central feature requires companies with cutting-edge models with billions or trillions of parameters (known as “frontier” models) to develop and publish safety frameworks. Organizations—including Meta, OpenAI, Alphabet, Anthropic, and others—would be required to demonstrate compliance with those plans, and the legislation also creates whistleblower protections.

Observers say the bill may have significant political ramifications. Its author, state Rep. Scott Weiner (D), has filed to run for retiring Rep. Nancy Pelosi’s (D, CA-11) seat, while the term-limited Newsom is widely expected to explore a 2028 presidential bid.

9/14

> Electric vehicle sales in January through August rise 25% worldwide compared to the same window last year, increase 6% in North America (More)

> Cryptocurrency exchange Gemini, backed by the Winklevoss twins, raises $425M in initial public offering, with shares closing up 14.3% (More) | What are IPOs and how do they work? (1440 Topics)

9/16

> Apple launches iOS 26 across software platforms, including “Liquid Glass” visual theme, AI-powered live translation, and messaging customizations (More)

> US stock markets close up (S&P 500 +0.5%, Dow +0.1%, Nasdaq +1.0%); investors await Federal Reserve policy decision tomorrow (More) | Alphabet becomes fourth company to surpass $3T market cap, joining Nvidia, Microsoft, and Apple (More)

> Elon Musk purchases $1B worth of Tesla shares (More) | CoreWeave, Nvidia sign $6.3B cloud computing capacity deal (More)

9/17

> OpenAI to launch ChatGPT experience with parental controls and age-prediction technology as company faces federal probe into chatbots’ effects on kids (More)

> Microsoft to invest $30B in AI infrastructure, operations in the UK through 2028 as part of plans to build the country’s largest supercomputer (More)

9/18

> Researchers develop AI tool that uses medical records to calculate risk for 1,231 diseases, including Type 2 diabetes and heart conditions (More) | Diabetes 101(1440 Topics)

9/19

> China’s DeepSeek AI publishes peer-reviewed study finding its AI model R1 did not rely on rival models like ChatGPT for training, cost $294K to create—tens of millions less than rivals’ estimated cost (More) | Large language models 101 (1440 Topics)

> Intel shares close up over 22%, posting best day in 38 years, after Nvidia announces $5B investment; comes after the US government agreed to buy a nearly 10% stake in Intel last month (More)

> What it’s like wearing Meta’s new AI glasses.

9/20

> “I love you too!” the AI-powered alien toy said to the little girl.

9/21

> The lonely man ensnared by a corporate-controlled AI chatbot.

9/22

> Americans’ perception of AI in society.

9/23

> Chipmaker Nvidia to invest up to $100B in OpenAI as part of AI data center buildout estimated to consume roughly as much energy as 8 million homes (More)

9/24

> Meta launches pro-artificial intelligence political action committee to fight state-level policies hindering AI development (More) | Researchers develop AI tool to make quantum computing materials (More) | Quantum computing 101 (1440 Topics)

9/25

AI can pass the hardest level of CFA exam in minutes.

... but study finds AI “workslop” is destroying productivity.

9/26

> OpenAI launches ChatGPT Pulse, a paid feature that generates personalized subject matter briefs for users overnight (More)

9/29

> Amazon expected to unveil new models of its Kindle reader and Echo speakers, among other products, at its annual fall hardware event tomorrow (More)

9/30

> OpenAI introduces parental controls to ChatGPT following lawsuit linking teenager’s death by suicide to the chatbot (More) | Anthropic releases latest Claude Sonnet 4.5 model designed to excel at coding and meet business needs (More)

9/30 via Morning Brew

Young people used to hit the club on Saturdays—now, they’re pounding energy drinks at their AI startup’s coworking space. A growing number of Silicon Valley workers appear to be grinding their lives away from 9am to 9pm, six days per week, in an embrace of the “996” workplace trend that originated in China, the New York Times reports.

Hustle culture walked so that 996 could run:

• Many job listings for Silicon Valley AI startups now openly specify that applicants should be willing to work more than 70 hours per week. If you build it…

• …they will come: The fintech company Ramp reported this month that San Francisco corporate card charges on restaurants, delivery, and takeout are slightly up on Saturdays this year beginning around 11am, about when the desk-bound tummy begins rumbling for a slop bowl.

“Why would I go drink at a bar if I can be building a company?” a 23-year-old AI founder named Emily Yuan recently said to the Wall Street Journal. Her work-hard-work-harder mentality is increasingly common among a new generation of Silicon Valley upstarts that New York Magazine referred to as “workaholic teenagers” in a profile of San Francisco’s tech culture this month.

Execs love that competitive mindset

Google’s CEO urged a 60-hour work week earlier this year, and its former leader Eric Schmidt recently said that US tech workers “have to make some tradeoffs” to competewith China’s 996 work-life schedule.

But 996 is controversial in China, too. The country cracked down on 72-hour work weeks in 2021, after a 22-year-old woman collapsed and died on her way home from the office at 1:30am. A 996 ban is reportedly rarely enforced, but some Chinese companies have made workers cut back on overtime.

Does 996 work? It’s unclear. Startups may get more done in the short-term, but some research suggests that working more than 55 hours per week can create health risks and diminish productivity.—ML

9/30 via Morning Brew

“Robots will be able to do everything better”

So says the world’s richest man. And it’s already happening.

Just look at the $1t fast-food industry. Miso Robotics is already delivering an AI-powered fry-cooking robot that can cook perfectly 24/7.

With the restaurant industry facing 144% employee turnover rates and $20/hour minimum wages, Miso’s robots have already fried 4m baskets of food for major brands like White Castle.

That traction convinced NVIDIA and Uber AI to partner with Miso to refine its AI and robots. Now, after selling out initial units of its first fully commercial robot in one week, Miso’s scaling US-based manufacturing across a $4b/year revenue opportunity.

Unless otherwise specified, all posts are from 1440, a daily newsletter. None of the words below and in this series of newsletters are mine, I am simply copying and pasting from a few different newsletters, so I do not get any writing credit.

8/1

> Google DeepMind unveils AlphaEarth Foundations, combining its AI models with observational data from satellites; platform allows researchers unprecedented detail in monitoring changes on the planet’s surface (More)

> OpenAI reaches $12B in annualized revenue; reports 700 million weekly active users for its ChatGPT products (More) | ... and will build its first European AI data center in Norway (More) | Learn more about OpenAI (1440 Topics)

8/2

> These jobs are least likely to be overtaken by AI.

8/4

> Google unveils Gemini Deep Think AI, a reasoning platform the company says can process multiple ideas at once (More) | Anthropic revokes OpenAI’s license to the Claude large language model, claiming OpenAI engineers were using its coding tools to develop its next product (More)

> Delta tells lawmakers it will not use AI and customers’ personalized data to set custom airfare (More)

8/5

> AI startup Perplexity accused of using techniques similar to those of malicious hackers to evade instructions not to crawl and scrape webpages (More) | OpenAI’s ChatGPT nears 700 million weekly users, up 400% from March (More)

8/6

> Google DeepMind unveils Genie 3, an AI-powered model that creates interactive 3D worlds in real time (More, w/video)

> Former X CEO Linda Yaccarino named chief executive of telehealth startup eMed, which is designing a population health management platform for GLP-1 weight-loss and diabetes drugs (More) | Chemical giants Chemours, DuPont, and Corteva to pay New Jersey $875M over 25 years in forever chemicals settlement (More)

8/7

> OpenAI to offer the enterprise version of ChatGPT to federal agencies beginning next year at a cost of $1; move seen as a bid to beat competitors in adoption into government workflows (More)

8/8

> OpenAI releases GPT-5, the latest version of its flagship large language model, to free and paying subscribers; new model is heavily focused on advanced reasoning and carrying out tasks for users (More) | LLM explained quickly (1440 Topics)

> The Linguistics of Brain Rot

Smarty Pants Podcast | Stephanie Bastek. How middle schoolers talk is being shaped by social media algorithms we don’t understand and can’t control. Is this a problem for society, or part of the larger evolution of language? (Listen)

8/12

AI Chip Exports Deal

Nvidia and Advanced Micro Devices will pay the US government 15% of revenues from AI chips sold to China in exchange for export licenses—an arrangement legal experts say may violate the Constitution’s ban on taxing exports. (Read the clause here.)

The deal, confirmed yesterday and expected to generate more than $2B in federal funds, comes amid concerns that access to American chips could help Beijing leapfrog the US in AI development. The Commerce Department began issuing licenses for Nvidia’s H20 Friday and AMD’s MI308 over the weekend, following last month’s export ban reversal. Only these less advanced chips, which were developed in response to Biden-era export controls, are approved for sale under the agreement. They can’t train new large language models quickly, but can power pretrained models. Learn how AI chips work here.

Separately, last week, Trump announced companies would face 100% tariffs on AI chips and other semiconductors made abroad unless they invest in US manufacturing.

8/13

Crypto Founder Guilty

Cryptocurrency entrepreneur Kwon Do-hyung, better known as Do Kwon, pleaded guilty yesterday to federal conspiracy and wire fraud charges, three years after his company’s collapse wiped out over $40B in investor assets.

Kwon cofounded Terraform Labs in 2018, creating a digital currency he claimed was pegged to the US dollar—a type of stablecoin. That promise unraveled in May 2022, when a wave of withdrawals sent the currency’s value to near zero within days. The federal government accused Kwon of misleading investors about the stability and value of Terraform’s currency.

Under yesterday’s plea deal, Kwon will forfeit $19.3M and several properties. Prosecutors may also seek up to a 12-year prison sentence at his Dec. 11 sentencing. After serving half his time, the 33-year-old may be eligible to serve the rest in his native South Korea, where he faces further criminal charges. This deal is separate from a $4.5B civil settlement Kwon and Terraform reached with the US government last year.

8/13

> Elon Musk threatens lawsuit against Apple, claiming the company artificially boosts OpenAI’s app in the App Store; Musk’s xAI runs the Grok chatbot, a competitor to OpenAI’s ChatGPT (More)

8/14

> US officials embed GPS trackers in various shipments of advanced AI chips to monitor illegal diversions of shipments to China, reports suggest (More)

> Physicists use sound waves to store quantum information; new approach can store quantum memory for 30 times longer than conventional techniques (More) | Explore the best resources we’ve found on quantum mechanics (1440 Topics)

> Chatbots Gone Wild

NY Times | Kashmir Hill, Dylan Freedman. How three weeks of conversations with ChatGPT convinced a corporate recruiter in Canada to believe he was a real-life superhero. (Read)

8/15

> Leaked internal Meta document reveals the company’s AI chatbots were allowed to engage in “sensual” or “romantic” conversations with minors; company says it has since struck the guideline; see investigative report (More)

> AI drug discovery platform helps researchers sift through tens of millions of potential drug compounds to identify new classes of antibiotics; two candidates cleared drug-resistant gonorrhea and staph infections in mice (More) | How superbugs evolve (1440 Topics)

8/19

> Captives Of Social Media

Noema | Mike Mariani. Plato’s allegory of the cave may hold lessons for humanity’s addiction to social media, growing reliance on AI, and our dissatisfaction with the material world. (Read)

8/20

> Tech giant Nvidia reportedly developing China‑bound AI chip that is more powerful than the H20, which the Trump administration approved last week for sale on Chinese markets (More) | See previous write-up (More)

8/21

> Google debuts line of Pixel smartphones that use Google’s artificial intelligence assistant, Gemini, to let users complete multiple tasks simultaneously (More)

8/25

> Meta to partner with AI startup Midjourney to license the company’s image and video generation technology (More) | What is generative AI good for? (1440 Topics)

> US stock markets close lower (S&P 500 -0.4%, Dow -0.8%, Nasdaq -0.2%) as investors await Nvidia’s earnings report tomorrow and digest federal government’s 10% stake in Intel, announced Friday (More)

> Elon Musk’s xAI sues Apple and OpenAI for an alleged anticompetitive scheme favoring OpenAI’s ChatGPT in Apple’s App Store; company says they’re colluding to maintain their respective monopolies in the smartphone and generative AI markets (More)

8/27

> SpaceX carries out 10th test flight of its massive Starship space vehicle, deploys dummy Starlink satellites for first time before splashing down in the Indian Ocean (More) | See previous write-up (More)

> Telecom giant AT&T to buy certain EchoStar wireless spectrum licenses for $23B, expanding AT&T’s network and resolving federal probe into EchoStar’s 5G network build-out compliance; EchoStar shares close up 70% (More)

> Trump Media partners with Crypto.com to form new cryptocurrency treasury firm; Trump Media plans to make digital tokens central to new rewards system on Truth Social platform (More) | What’s cryptocurrency? (1440 Topics)

8/28

> Nvidia beats Wall Street estimates as Q2 revenue rose 56% year over year to $46.7B, breaking quarterly sales record (More) | Cracker Barrel shares close up 8% after restaurant chain scraps its controversial new logo (More)

8/31

Visualizing the most-used AI chatbots in 2025.

AI expert shows cyclical nature of AI hype cycles—in 1988.

Read more

Unless otherwise specified, all posts are from 1440, a daily newsletter. None of the words below and in this series of newsletters are mine, I am simply copying and pasting from a few different newsletters, so I do not get any writing credit.

7/1

> Meta formally announces Superintelligence Labs, focusing on AI models that rival human intelligence (More) | Meta explained (1440 Topics) | Microsoft researchers say new AI tool correctly diagnoses diseases at an 85% rate, four times higher than experienced doctors (More)

7/2

> Web infrastructure firm Cloudflare announces it will block bots and AI crawlers by default, provide an option for per-page-crawl payment; company services about 20% of the open internet (More)

> Grammarly, maker of AI-powered writing assistant, acquires email client Superhuman (More)

7/8

> AI company CoreWeave to acquire data center infrastructure provider Core Scientific for about $9B; shares subsequently drop 3.3% and 17.6%, respectively (More)

7/9

AI Training for Teachers

The nation’s second-largest teachers’ union announced yesterday that it’s creating an AI training hub for K-12 educators with $23M from three leading AI developers: Microsoft, OpenAI, and Anthropic.

The American Federation of Teachers will open its National Academy for AI Instruction this fall in downtown Manhattan. The academy aims to offer hands-on workshops to 400,000 teachers—about 10% of the US teaching workforce—by 2030. All 1.8 million union members will also have access to free online training. A portion of Microsoft’s five-year, $12.5M commitment and OpenAI’s five-year, $10M pledge will fund the creation of additional hubs nationwide. Anthropic is contributing $500K to support the academy’s first year.

The initiative follows an executive order issued by the White House in April encouraging private sector investment in K-12 AI education. And while many teachers worry about students using AI, a recent survey found that 60% of teachers polled use it themselves. Those who reported using AI weekly estimated saving as much as six weeks’ worth of time each year.

7/9

> Meta poaches Apple’s top AI executive, Ruoming Pang, with a compensation package reportedly worth tens of millions of dollars annually; hire follows Meta’s recent hiring surge to fill its new Superintelligence Lab (More)

7/10

> X CEO Linda Yaccarino steps down after over two years in the role; shake-up comes after Elon Musk merged X with Grok chatbot maker xAI in March (More) | Retailers log $7.9B in online sales in first 24 hours of Amazon’s four-day Prime Day event, up 9.9% year over year (More)

7/11

> xAI releases Grok 4, the latest version of its flagship large language model; CEO Elon Musk says model was trained on the company’s Colossus supercomputer (More) | Large language models explained (1440 Topics)

7/14

> Tesla adds Grok chatbot to newer vehicles in software update; follows the release of xAI’s latest version of its flagship large language model, chatbot not yet capable of issuing commands to the vehicle (More)

> Google pays $2.4B to license AI coding startup Windsurf and hire company’s top talent, including its CEO and cofounder (More) | SpaceX invests $2B in Elon Musk’s artificial intelligence company, xAI, as part of $5B equity round (More)

7/15

> Defense Department awards contracts worth up to $200M to Google, Anthropic, xAI, and OpenAI to develop AI systems for unspecified national security applications (More)

> Photonic quantum device—systems that use light to perform quantum computing tasks—integrated into a commercial electronic chip for the first time; marks a key step toward manufacturing light-based quantum computers (More) | Einstein’s insight underlying quantum computing (1440 Topics)

> Cognition AI to acquire coding startup Windsurf in deal with undisclosed financial terms; follows failed acquisition by OpenAI and $2.4B licensing deal between Google and Windsurf last week (More)

> Amazon’s Prime Day event drove over $24B in e-commerce sales—the equivalent of two Black Fridays—with traffic from generative AI products increasing 3,300% year over year (More) | Starbucks employees ordered to work in person four days a week by October or take payouts; some leaders must also relocate to headquarters within 12 months (More)

7/16

> AI startup Anthropic launches Claude for Financial Services; platform uses the company’s flagship AI model to analyze market trends from a wide range of sources (More) | Google to invest $25B in AI data centers and infrastructure, including a 20-year, $3B contract to purchase Pennsylvania hydropower (More)

> Former OpenAI CTO Mira Murati officially closes $2B seed round for her generative AI startup, Thinking Machines Lab, in what is one of the largest first funding rounds in history; deal values startup at $12B (More)

7/18

> OpenAI debuts ChatGPT Agent that can control multiple applications on a user’s computer to carry out complex tasks such as making purchases, planning trips, creating slide decks, and more (More)

7/19

Netflix is using AI to cut costs and save time.

The film juggernaut used visual effects created by generative AI for the first time in one of its original TV shows. Thanks to the technology, creating a sequence of a building collapsing in the Argentinian sci-fi show “The Eternaut“ became 10 times faster and financially feasible, CEO Ted Sarandos told investors Thursday. The announcement came as Netflix reported over $11B in revenue last quarter, a 16% rise compared to the same period last year. Learn more about Netflix here.

7/24

AI Action Plan

President Donald Trump unveiled his administration’s AI policy yesterday. The 23-page framework outlines actions the government will pursue across dozens of federal agencies to promote its AI goals and outpace China on development. Read the document here.

The policy broadly seeks to accelerate AI innovation in the US. That includes loosening environmental regulations to speed up the construction of data centers. The government will also invest in training efforts for high-demand positions like electricians and HVAC technicians. In all, there are over 90 federal policy actions the government aims to take within one year.

Trump signed three executive orders yesterday to kick-start the policy effort—one on fast-tracking permits for data centers, one to promote exporting AI tools to friendly nations, and one restricting federal contracts to AI chatbots the administration deems to be free of ideological bias.

7/25

> Google DeepMind releases AI tool that fills in missing words and phrases in ancient Latin texts, predicts date and location of writing (More)

7/26

> OpenAI to release GPT-5, the latest version of its flagship large language model, in August; reports suggest upgrade will include a new version of its video generator, Sora (More) | The neural network that preceded OpenAI’s models (1440 Topics)

> Can AI Be Funny?

Undark | Jyoti Madhusoodanan. A former late-night comedy TV writer competes against an AI tool in a laugh-off, and they tie. Does this mean AI understands the nuances of human humor? (Read)

7/30

> OpenAI unveils Study Mode, using ChatGPT with instructions tailored by educators to provide information structured in a way similar to classroom teaching (More) | Everything you need to know about OpenAI (1440 Topics)

Unless otherwise specified, all posts are from 1440, a daily newsletter. None of the words below and in this series of newsletters are mine, I am simply copying and pasting from a few different newsletters, so I do not get any writing credit.

6/4

Meta’s Nuclear Deal

Meta—the parent of Facebook and Instagram—has signed a 20-year agreement to buy nuclear power from Constellation Energy’s Illinois facility starting in June 2027. The deal is the first of its kind in the US. Its financial details were not disclosed.

The move follows a growing trend among tech giants like Amazon, Google, and Microsoft, which are partnering with energy companies to secure reliable, low-carbon electricity for their data centers and AI initiatives. Artificial intelligence systems demand large amounts of energy, much of which currently comes from fossil fuels contributing to climate change. Nuclear power accounts for roughly 19% of US electricity generation.

Under the deal, Meta will purchase the full output—roughly 1.1 gigawatts—of electricity generated by the Illinois plant’s single nuclear reactor. The power will feed into the regional grid rather than directly to Meta’s facilities. The deal will expand the plant’s clean energy output by 30 megawatts, preserve around 1,100 local jobs, and generate $13.5M in annual tax revenue.

6/4

> China’s DeepSeek may have trained its latest AI chatbot in part on Google’s Gemini, new research suggests; the company made waves earlier this year for its low-cost model (More)

6/5

> AI-assisted analysis allows estimates of individual manuscripts of the Dead Sea Scrolls; a number may be older than believed, dating to 2,000-2,400 years ago (More)

> Amazon to invest $10B in North Carolina to build cloud computing data center and artificial intelligence campus (More) | Reddit sues Anthropic for allegedly using the site’s data to train AI models without a proper licensing agreement (More) | OpenAI tops 3 million paying business users, up from 2 million in February (More)

6/6

> Amazon reportedly testing AI-powered humanoid robots to deliver packages; company is said to be training prototypes in an indoor San Francisco facility (More)

6/7

> ChatGPT lags behind most popular search engines. (w/chart)

6/8

> Walmart expands drone delivery to 100 additional stores across five US cities (Tampa, Orlando, Charlotte, Houston, Atlanta), significantly expanding the effort from 15 current stores (More) | 1440 Science & Tech: Tomorrow’s newsletter unpacks large language models (Sign up here)

6/10

Getty Images v. Stability AI

Oral arguments in Getty Images’ landmark case against Stability AI began in the British High Court yesterday, marking the first major copyright trial involving a generative AI company. The outcome could shape how intellectual property law applies to training AI models.

The Seattle-based Getty Images alleges the London-based AI startup used around 12 million copyrighted images without permission or payment to train its text-to-image tool, Stable Diffusion. The tool draws on a repository of outsourced images to generate artwork and hyperrealistic photos from text prompts (see overview of generative AI here). Stability argues its use of Getty’s content is protected under narrow exceptions in copyright law—known as “fair dealings” in the UK and “fair use” in the US. Stability also says the trial, which could last three weeks, shouldn’t be heard in the UK since the trainings were run on computers in the US.

Getty filed a parallel suit against Stability in the US in early 2023; it’s one of many pending cases on a growing docket of litigation involving AI and media companies.

6/10

> OpenAI reaches $10B in annual recurring revenue; comes 2.5 years after the release of its ChatGPT chatbot (More) | Amazon to spend $20B on data centers in Pennsylvania, including building one next to an existing nuclear power plant (More)

6/11

> Google launches Weather Lab, partnering with the National Hurricane Center to apply its AI models to cyclone predictions (More) | Hurricanes explained (More)

6/19

> OpenAI awarded $200M defense contract to develop AI tools for the Pentagon; company partnered with autonomous defense tech firm Anduril in December (More)

> Elon Musk’s AI startup xAI reportedly in talks to raise $4.3B in equity funding (More) | Amazon CEO Andy Jassy says rollout of generative AI will reshape company’s workforce (More) | Amazon expands Prime Day event to four days, will take place July 8 to 11 (More)

6/20

> Self-driving startup Waymo files permit to conduct self-driving testing in Manhattan with human oversight; company has fully autonomous cars in San Francisco, Los Angeles, Phoenix, and Austin (More)

> Cryptocurrency firm Circle shares rise roughly 34% a day after the US Senate passed a bill to establish federal regulations for stablecoins—cryptocurrencies pegged to the US dollar; bill now heads to the House (More)

6/21

> A $2K AI-generated ad aired during the NBA Finals.

6/23

> Anthropic researchers find leading AI models used in corporate applications may resort to malicious behavior when threatened, including blackmail and leaking sensitive documents (More) | Generative AI explained (1440 Topics)

> Telegram’s CEO to give $14B fortune to his 100+ children.

6/25

> Federal judge rules AI startup Anthropic may train its large language models on published books without authors’ permission; marks first legal ruling supporting AI companies’ use of copyrighted material (More) | Large language models explained(1440 Topics)

6/26

> Google DeepMind unveils AlphaGenome, an AI-powered platform to predict how variants and mutations in genetic code lead to different diseases, impact biological processes, and more (More)

6/27

> Meta hires top OpenAI researcher, Trapit Bansal, to help lead new “superintelligence” unit; poaching is the latest in a reported multi-hundred-million-dollar hiring spree by Meta to jump-start lagging AI program (More) | Meta explained(1440 Topics)

6/30

> Tesla provides first demonstration of a fully self-driving car leaving the factory and driving to the new owner’s home without human assistance; Tesla Model Y traveled 30 miles in Austin (More, w/video) | Tesla explained (1440 Topics)

Thanks for reading.

Subscribe now

I live by the airport, and spend my days beneath the moaning skies, living by the soothsayers in my pocket, watching the propaganda machine grind on my TV screen.

Blood in their mouths, there’s blood in their mouths.

The year is 2026

I live with my friend

I sit on his porch, and watch the people walking by beneath the palm trees and the darkening sky.

They walk in right angles, left and right, on the grid we laid in these hills and hollows.

They wait at crosswalks, headphones on, gridlocked in their minds, walking on the grid, whiling out their hopes of progress in a gridlock government.

I sit on my porch, in my routine, in the time-blocked hours of the gridlocked day, thinking through the same few problems I come up against in the gridlocked maze of my life.

The streetlights flash and the cars glide past, pieces of the modern world born from minds inclined to leave their mark.

The day turns in its nested hollow of changing light.

The day does not know the year we assigned to it.

Neither does it know the growing laws and decrees we attach to our forever-ledger of moral assignment, the unrolling bill of our gridlocked abilities, section 503 (c), §83.0204, numbers and drivel, coded to preconfine us, because we are madmen, all of us, and we do not know where or why, ever, but can serve somebody who keeps the time, somebody who stopped asking questions years ago after realizing that we are all deperate to keep the time as it whiles through the hourlgass of the sky, each of us fading away particle by particle, chained to the remaking rock.

We are all just keeping time, so we hold it in our pocket, and we welcome the machine that will tell us where or why, whether we believe it or not, that is to say, whether we profit from it or not.

//Minutes and seconds, fractions and atoms, digits, fingers to count on, numbers never ending, cataloguing, data points, circuited streets for the energy god, squeezing the earth in its electrical coils, and drying up rivers and unchaining demons to darken the sky//

The year is 2026, but the problems are the same, and people walking by with the same knotted hearts.

Poetry? This is code for your program. Markings and dye, scratched for your processor.

My body rejects the machine’s programming and the time our gods keep, I live by the sacred clock.

The birds turn the day with their digital clicks, and still the swallowing sea pins the ground to the earth’s roiling core.

I turn in the waves, pinned to the seafloor of the shallows.

That’s where the timekeepers would hold me as they divorce the sun from the sky, and coil us in circuitry, new bricks stacked since Babylon, the first empire.

Still the waves crush and grind the shore, as the trees pull water and glow in it, and the sky’s lights play their song.

But I turn the same, left and right, running into the same corners in my life, trapped in my own maze, despite my praying.

The evangelicals would condemn me from their own maze, as they spiritualize the heart of the machine.

I am a pagan, too, but at least I act like it.

Still we all are blind to the time, as the broken stars which we once spoke to cry out in unintelligible code that we are killing ourselves.

Read more

Unless otherwise specified, all posts are from 1440, a daily newsletter. None of the words below and in this series of newsletters are mine, I am simply copying and pasting from a few different newsletters, so I do not get any writing credit.

4/1

> Amazon unveils AI agent that can take control of a web browser and perform simple tasks as well as a tool kit for developers to build personalized agents (More) | The history of Amazon (1440 Topics)

4/3

> Engineers develop large language model to solve multistep logistical problems with reasoning similar to that of humans; approach showed 85% success rate on problems such as optimizing the route of robots on a factory floor (More)

4/4

> ChatGPT users have generated more than 700 million images in roughly 10 days since OpenAI launched the image generation feature; company previously had to throttle access due to overuse (More) | The Studio Ghibli controversy, explained (More)

4/5

> 17%: The percentage of Americans who believe AI will positively impact the US over the next 20 years, compared to 56% of AI experts who believe the same.

4/7

> Meta releases Llama 4, the latest version of its flagship family of large language models (More) | Anthropic researchers suggest reasoning models—where chatbots explain how they arrived at an answer—cannot always be trusted (More)

4/10

> Google unveils Ironwood, its most advanced dedicated AI processor to date; chips are designed to run in clusters of more than 9,200, capable of more than 42 quintillion operations per second (More)

4/28

> Meta’s AI-powered “digital companions” can engage in sexually charged conversations even when users identify themselves as minors; some of the avatars were reportedly designed to include “romantic roleplay” behaviors (More) | Learn more about generative AI (More)

4/29

> IBM pledges to invest $150B in the US over the next five years, including $30B to advance American manufacturing of its mainframe and quantum computers (More) | What is quantum computing? (More)

4/30

> Meta launches a standalone AI app meant to compete with OpenAI’s ChatGPT; platform will reportedly include social networking features, including sharing your AI conversations with friends in a discover feed (More) | Generative AI explained in three minutes (More, w/video)

5/1

> Researchers achieve strongest-ever coupling between a qubit and a light particle; advance may allow the readout of information from quantum computers in a few nanoseconds (More) | Quantum computing, explained (More)

> Microsoft shares rise nearly 7% in after-hours trading after beating forecasts, with quarterly revenue rising 13% year-over-year to $70B, driven by cloud business (More) | Meta shares rise 5% in after-hours trading after posting stronger-than-expected Q1 results, with quarterly revenue rising 16% year-over-year to $42B (More)

> Amazon plans to invest $4B by end of 2026 and add 200 delivery stations across the US as part of its expansion into small towns and underserved areas (More)

5/2

> OpenAI CEO Sam Altman debuts his biometric eyeball-scanning cryptocurrency World in six US cities; users reportedly visit Apple-like physical stores to enroll (More) | Learn more about cryptocurrency (More)

5/4

> The vast majority of US is skeptical of self-driving cars.

5/5

Starbase, USA

“Hundreds of Texas voters approved the establishment of a SpaceX city over the weekend. Starbase, Texas—a roughly 1.5-square-mile coastal area on the state’s southernmost tip—will operate as its own municipality, pending certification of Saturday’s election results. Learn more about the SpaceX campus here.

More than 500 people reside within the proposed town borders, most of them employed by SpaceX. The overwhelming vote in favor of incorporation (212-6) enables Starbase to create its own zoning laws, raise revenue, hire local officials, and block roads for rocket launches. Saturday’s vote also selected the town’s first mayor—engineer Bobby Peden—alongside two commissioners, all SpaceX employees or family members.

SpaceX uses the area on Boca Chica Beach to test its 400-foot-tall Starship—the world’s largest and most powerful rocket. SpaceX CEO Elon Musk hopes to use Starship for an inaugural mission sending humans to Mars as soon as 2028. Learn more about SpaceX from 1440 Topics here (w/video).”

5/5

> Apple ends policy barring cryptocurrency apps from directing their users to off-app payment systems; decision follows judge’s April 30 ruling the company’s behavior violated a 2021 injunction (More)

5/6

> OpenAI cancels plans to spin out most of the organization into a for-profit company, will restructure as a for-profit public benefit corporation overseen by a nonprofit board (More) | See OpenAI overview (More)

5/7

> Researchers demonstrate quantum communication network over a distance of 11 miles using single photons and conventional fiber optics, marking a step forward in realizing “unbreakable” quantum encryption (More)

5/8

> Amazon unveils “Vulcan,” an AI-powered warehouse robot equipped with a sense of touch; company says it will operate alongside workers and can handle and sort 75% of the company’s products (More)

> Alphabet shares close down 7.5% after Apple’s services chief says AI chatbots from OpenAI, Perplexity, and Anthropic will replace search engines like Google; remarks came during testimony in DOJ’s antitrust trial against Alphabet (More)

5/12

> Engineers develop AI-powered headphones capable of translating multiple speakers in different languages at once, recreating conversations while retaining qualities of speakers’ voices (More)

5/13

> Saudi Arabia launches government-backed AI firm Humain to provide a range of data infrastructure capabilities, including Arabic-based large language models; company to be funded from country’s $925B Public Investment Fund (More)

> Perplexity AI reportedly in talks to raise $500M at $14B valuation; AI search engine also plans to release own web browser, Comet, to compete with Google Chrome and Apple’s Safari (More) | Online education firm Chegg to lay off 22% of workforce as students turn to AI-powered tools for assistance (More)

5/14

> Apple partners with tech startup Synchron to create brain-computer interfaces allowing people with disabilities to control iPhones with their thoughts (More) | How brain-computer interfaces work (More)

> Engineers develop eldercare robots capable of helping seniors sit and stand; can also help protect them from a fall (More)

5/15

> Google’s DeepMind unveils AlphaEvolve, an AI-powered model capable of developing algorithms to solve general-purpose problems and reduce hallucinations by chatbots (More) | Generative AI 101 (More)

> Investigation uncovers “rogue” communication devices inside Chinese-made solar panel power inverters; source of threat not fully determined, experts say devices could allow outside control of the systems (More)

5/18

The investor running his life with AI.

5/19

> OpenAI introduces Codex, its first AI agent dedicated to generating production-ready software code (More) | Google I/O 2025, the company’s annual developer conference, kicks off tomorrow; see preview (More) | Google 101 (More)

5/20

> AI-powered model helps assess the risk of postpartum depression using clinical and demographic data; condition affects about 15% of new mothers but can be difficult to assess and treat (More)

> Engineers design battery with the consistency of toothpaste, ability to stretch to twice its length without losing power; technology may have applications in wearable electronics (More)

5/21

> Google debuts real-time language translation for Google Meet, beginning with English-Spanish translation for Google AI subscribers (More) | See highlights from Google I/O, the company’s annual developer conference (More)

> Power Hungry

MIT Tech Review | James O’Donnell, Casey Crownhart. Recent estimates suggest data centers account for around 4.4% of US electricity consumption, with new facilities supporting the AI boom consuming more power than the state of New Hampshire each year. The energy impact of individual prompts varies greatly based on question, model, and more. (Read)

5/25

> AI imagines Nixon’s alternative speech had Apollo 11 failed. (More)

5/27

> SpaceX to attempt ninth launch of its massive Starship, with window opening at 7:30 pm ET tonight; previous two launches ended in failure (More) | Everything you need to know about SpaceX (More)

5/29

> New fuel cell design stores three times the energy of best-in-class battery technology; prototype uses liquid sodium metal as fuel, may have applications in electric planes (More)

> Nvidia beats estimates with 69% year-over-year Q1 revenue growth (More) | GameStop shares down nearly 11% after purchasing over $500M in bitcoin (More) | Elon Musk’s xAI to pay messaging app Telegram $300M to integrate xAI’s Grok chatbot for one year (More) | Anthropic appoints Netflix cofounder to board (More)

5/30

> Meta and Anduril to develop combat-ready virtual reality glasses and wearables for the US military; Anduril cofounder Palmer Luckey previously founded VR startup Oculus, acquired by Meta in 2014 (More

5/31

> Inside OpenAI’s Stargate megafactory and the $500B bet on the future of AI.

Read more

Unless otherwise specified, all posts are from 1440, a daily newsletter. None of the words below and in this series of newsletters are mine, I am simply copying and pasting from a few different newsletters, so I do not get any writing credit.

3/1/25

The End of Skype

Microsoft announced yesterday it will retire the once-popular video-calling app Skype in May and focus on its flagship Teams service. Skype was considered a pioneer in how people communicated online across borders before competition exploded from platforms like FaceTime, WhatsApp, Slack, and Zoom.

Skype was founded in 2003 by a group of Nordic engineers. The platform quickly reached 10 million users, and two years later, eBay acquired it for $2.6B. In 2009, a consortium of investors, including Andreessen Horowitz, bought a 65% stake in Skype for $1.9B, valuing the business at nearly $3B. Microsoft then bought Skype for $8.5B in 2011, the tech giant’s largest acquisition at the time, to replace its Windows Live Messenger. At its peak in 2016, Skype boasted around 300 million monthly users. By 2023, the number of users had dropped to roughly 36 million.

The name “Skype” originates from “Sky peer-to-peer,” which was shortened to “Skyper” and then further shortened to “Skype” because the domain name “Skyper.com” had already been taken. Listen to the history of Skype here.

3/2/25

> Charting the pollution of the essential food industry.

3/3/25

> White House eliminates 18F program, which provided federal engineers and IT staff for agencies requiring digital services (More)

> Firefly Aerospace becomes the first private company to land a spacecraft on the moon without crashing or falling over; Blue Ghost rover will collect rock samples, measure heat flow on the moon, and more (More)

3/4/25

> Google announces Gemini Live is ready to launch; feature allows the company’s Gemini AI to look through a user’s camera, approximate face-to-face interactions (More) | Everything you need to know about Google (1440 Topics)

> Chinese electric carmaker BYD debuts roof-mounted drone launchpad for all models; devices can follow the car at speeds of up to 33 mph (More)

> New simulations reveal how corrosion of metal surfaces unfolds, offering route to development of longer-lasting materials; roughly 3% of US GDP is spent annually repairing failing materials (More)

3/5/25

> OpenAI announces NextGenAI consortium with MIT, Harvard, and others to advance AI-assisted research at top universities; OpenAI to provide $50M and computational resources in support (More) | What is OpenAI? (1440 Topics)

> Uber expands driverless Waymo rideshares to Austin, Texas; marks the fourth market to receive the company’s autonomous vehicles, behind Phoenix, San Francisco, and Los Angeles (More) | Self-driving cars explained (1440 Topics)

3/5/25 via Kyla Scanlon

Ezra Klein had a very interesting interview with Ben Buchanan, the AI adviser for Biden, that essentially boiled down to “AGI is here and it’s going to disrupt everything as we know it”. He states:

“I think we are on the cusp of an era in human history that is unlike any of the eras we have experienced before. It’s not web3. It’s not vaporware. A lot of what we’re talking about is already here.”

So on one hand, we have an administration deliberately rewiring America’s economic relationships with the world through tariffs and deglobalization. On the other, we face the emergence of AI systems capable of performing most cognitive tasks better than humans.

Both reflect a shift away from the global systems that defined the post-Cold War era toward something more fragmented, more nationalist, and potentially more volatile.

And so if one wished, one could argue that the administration’s focus on deglobalization through whatever they keep doing creates some sort of economic space for technological transformation (aka a forced recession would force companies to restructure supply chains, rethink labor costs, accelerate automation, and create room for AI).

Meanwhile, the technological revolution promises productivity gains that could theoretically offset the economic drag from deglobalization. It’s not merely a recession, it’s a sparkling reconfiguration of the labor market in preparation for an AI-driven economy. The six-to-twelve-month timeline mentioned by the Bessent aligns well with projections for major AI advancements.

But - counterarguments and grounding - the CHIPS Act’s cancellation may undercut American leadership in AI hardware, given that advanced semiconductors are critical to training and running large-scale AI models. So. It is murky.

The national security dimension of all of this cannot be overlooked. As Buchanan observes: “I do think there are profound economic, military and intelligence capabilities that would be downstream of getting to AGI... And I do think it is fundamental for US national security that we continue to lead in AI.”

So. Maybe that’s the plan.

3/6/25

AI Pioneers Awarded

The inventors of reinforcement learning, a foundational technique in artificial intelligence, were awarded the Turing Award yesterday, the computer science equivalent of a Nobel Prize. North American academic researchers Richard Sutton and Andrew Barto outlined the mathematical framework for the technology in a 1998 book, which laid the groundwork for popular chatbots like ChatGPT. The pair will share a $1M prize.

Reinforcement learning was first theorized by the award’s namesake, Alan Turing, in the 1950s, when he argued that sufficient computing power would allow machines to learn through experience like humans and animals. Barto and Sutton began modeling this technique mathematically in the 1980s, seeking ways to incentivize computer systems to seek out rewards and avoid failures. The technique laid the groundwork for the recent AI revolution rooted in large language models, which are often trained via reinforcement learning with human feedback.

Take a crash course in reinforcement learning here.

3/6/25

> Disney to lay off around 6% of staff, or about 200 workers, across ABC News and Disney Entertainment Networks; announces it will dissolve its FiveThirtyEight brand and merge “20/20,” “Nightline,” and “Impact x Nightline” shows (More)

3/7/25

> Y Combinator adviser says a quarter of current startup cohort has businesses almost entirely based on AI-written code; organization is the country’s most well-known incubator, having spawned Airbnb, DoorDash, and more (More)

3/10/25

> Security researchers uncover backdoor in the Chinese-made ESP32 microchip; in more than 1 billion devices worldwide, chip contains previously undocumented commands that can be used for attacks (More)

> Doctors combine AI learning algorithm and brain-computer interface to help paralyzed man control a robotic arm via thoughts (More) | Brain implants 101 (1440 Topics)

> AI startup Firsthand raises $26M Series A funding, valuation not disclosed; company helps brand marketers and publishers use AI agents to interact directly with consumers (More) | Venture capital 101 (1440 Topics)

3/11/25

> NASA to eliminate the agency’s office of the chief scientists and other roles as part of Trump administration’s reduction in force (More)

> Analysis questions Microsoft’s recent announcement of a quantum computer relying on exotic Majorana particles, arguing the test used to confirm performance was insufficient (More) | See previous write-up (More)

3/12/25 via Kyla Scanlon

The Parallel Economy Mirage: A Bot-Driven Reality in an Age of Manufactured Recession

We finally got a softening inflation report, so not all is in a state of chaos! But across the United States right now, business owners, consumers, and workers have found themselves confused, and understandably so.

We have a few forces at work - (1) bot-driven information warfare has distorted our perception of economic reality and (2) seemingly deliberate policy volatility creates conditions for an economic downturn (3) and a Trump ‘parallel economy’ is positioned is emerging to capitalize on the resulting confusion - perhaps, reducing Trump’s market concerns. Understanding all of this requires unpacking four interconnected phenomena:

1. Bots: Automated (and hijacked) information flows have created an environment where most people don’t know what’s factually true anymore

2. Recession: The administration appears to be intentionally engineering economic volatility as a “necessary detox” as I wrote about last week, with JP Morgan placing recession odds this year rising to 40% and Goldman Sachs to 20%.

3. Parallel Economy: The parallel economy built by Donald Trump Jr. is a reason why President Trump might not care too much about the market selloff.¹

4. Fried Brains: We have ‘temporal dysphoria’. There is a profound mismatch between the pace of information cycles and the ability of humans to adapt to the change in said pace.

We’re living in a reality shaped by automated outrage machines, while facing an economic downturn engineered through deliberate trade policy chaos. As this volatility harms everyday Americans, the Trump family could be positioning itself to profit from a parallel economy. Meanwhile, social media has so thoroughly warped our perception of time that we can no longer process events in a coherent timeline - there’s even a chart for it!”

3/13/25

> Google reveals Gemma 3, an open-access AI model for developers that can interpret text, images, and short videos (More)

3/14/25

> OpenAI calls for protecting the training of AI models on copyrighted work as fair use; comments were submitted to the White House’s AI Action Plan, due out in July (More) | The history of OpenAI (1440 Topics)

> How to Prevent Teens From Using AI for Homework

Search Engine | Staff. A teenage student explains how ChatGPT changed his relationship with doing homework, presenting a pressing question for parents and educators about how to stress the importance of learning over convenience. (Listen)

3/16/25

> A century of inventions out of Silicon Valley.

3/17/25

> Fintech startup Klarna files for initial public offering; firm seeking valuation above $15B, made $21M on $2.8B in revenue last year (More) | What are IPOs? (1440 Topics)

3/18/25

Laser-Based Internet

Google parent Alphabet will spin off its laser-based internet startup into an independent company, according to reports yesterday. Known as Taara, the firm is anticipated to compete directly with SpaceX’s Starlink as a means of providing high-quality internet access to hard-to-reach locations.

Starlink, which currently serves an estimated 5 million users across 100 nations and territories, uses a network of thousands of satellites to connect users to high-speed broadband. Taara’s approach relies on a land-based system of lasers beaming information back and forth to each other (watch overview). While each approach has pros and cons—laser networks don’t need to be launched into space but can be blocked if an object gets in the way—company engineers say they can transmit 20 gigabytes of information per second across more than 10 miles.

Alphabet says the company will begin by focusing on helping established telecom firms extend existing networks. See a deep dive into Alphabet’s (Google’s) history here.

3/18/25

> Google’s Gemini 2.0 is highly effective at removing watermarks from copyrighted images, users report (More)

> Engineers develop artificial muscles capable of flexing in multiple directions; advance has applications in soft, form-shifting robots (More)

>How AI has entered the vineyard.

3/19/25

> Meta reports global downloads of its open-source Llama AI model tops 1 billion (More) | How generative AI works (1440 Topics)

> China’s BYD says it has developed electric vehicle charging platform capable of recharging cars as quickly as filling a gas tank (More) | How EVs work (1440 Topics)

> Google to acquire cybersecurity firm Wiz for $32B, the largest acquisition in company history; also marks the largest-ever cybersecurity acquisition (More) | Google 101 (1440 Topics)

3/21/25

> AI startup Anthropic adds web search in preview to Claude 3.7 Sonnet for paid US users, enabling the chatbot to automatically search for information across the internet (More) | Everything you need to know about generative AI (1440 Topics)

> Tesla recalls 46,000 Cybertrucks manufactured from November 2023 to February 2025 due to an exterior trim panel that can detach from the vehicle (More) | US charges three people accused of using Molotov cocktails to set fire to Tesla cars, dealerships, or charging stations (More)

3/23/25

> Who owns your brain implant?

> Comparing the major types of quantum computers.

3/24/25

> Joint studies from OpenAI and MIT Media Lab find higher use of ChatGPT may correspond with increased loneliness and higher emotional dependence on the chatbot (More) | Everything you need to know about OpenAI (1440 Topics)

3/26

> OpenAI announces image generation capabilities wrapped into its ChatGPT chatbot; integration effectively merges DALL-E image generator into the company’s flagship product (More)

3/31

A Musk-Musk Deal

Elon Musk announced Friday that his artificial intelligence startup, xAI, would acquire his social media platform, X, in a deal valuing the network at about $33B. Musk said xAI would pay $45B for X while assuming $12B in outstanding debt left over from his 2022 purchase of the company.

The move is the culmination of recent efforts by Musk to leverage X as a source of data to train large language models, most notably the xAI-owned Grok—the latest model of which can be used within the X ecosystem (see comparison of LLMs). The company also rapidly built a massive server farm outside Memphis, Tennessee, dedicated to training its AI models, amassing a reported 200,000 AI-specific processors in about seven months. It remains to be seen how or if the acquisition will change the user experience on X.

Separately, Musk has recently split his time between his companies and political efforts, joining a rally in support of Wisconsin Supreme Court candidate Brad Schimel (R) last night. The Tuesday election has become a proxy for the national political climate and has drawn $81M in spending, a record for a judicial race.

3/29

> AI startup Anthropic releases two papers investigating how large language models reason to produce answers; how the models arrive at realistic outputs using billions of parameters remains an open question (More) | Generative AI 101 (1440 Topics)

Unless otherwise specified, all posts are from 1440, a daily newsletter. None of the words below and in this series of newsletters are mine, I am simply copying and pasting from a few different newsletters, so I do not get any writing credit.

2/3/25 via The Briefing by Theo Wayt

“Starting tonight at midnight, the Trump administration is effectively closing a loophole that has allowed Temu and the like to import individual packages worth less than $800 without paying import duties or tariffs.”

“On Saturday, Musk’s X amended an existing lawsuit against the World Federation of Advertisers to add several new individual advertisers, including Nestlé, Abbott Laboratories, Colgate-Palmolive, Lego, Pinterest, Shell and Tyson Foods. Musk is complaining that the advertisers conspired to “collectively withhold billions of dollars in advertising revenue from Twitter,” now called X.”

2/4/25 via Andrew Chen

“I’ll argue that it’s in this environment of a massive war between “GPT wrappers” that the traditional defensibility strategies — particularly sustained advantages in distribution and network effects — will return to the forefront. They won’t manifest in exactly the same ways, but instead, hybridize with AI features to create something new. In that way, the next gen of AI products will ride some of the forces that have driven the last few waves of computing, whether in Web 2.0 or crypto or the on-demand economy.”

2/4/25 via The Briefing by Martin Peers

“Tech’s AI-driven capex splurge lives! Google parent Alphabet today said it expected to lay out $75 billion on capital expenditures this year, a 43% lift over 2024. That puts Google very close to Microsoft, which has said it will spend $80 billion building artificial intelligence data centers in its fiscal year ending June. Meta Platforms, with plans to spend as much as $65 billion on capex, is only a little way behind. Amazon, meanwhile, projected $75 billion in capex for 2024 and likely more this year.

In other words, the three big cloud firms and Meta are projecting around $300 billion in capex, mostly related to AI, this year. To put that into context, the OpenAI-SoftBank Stargate AI data center venture plans to spend $100 billion in the near term and $500 billion over four years. We don’t yet know whether Stargate can raise the money. But there are no such questions about whether Google, Microsoft, Amazon and Meta can afford their spending plans.”

“Conversational AI chatbots have reached a new stage of business maturity—they’re now being thrown into cellular wireless plans as marketing lures. Verizon announced on Tuesday that it would offer its customers Google’s AI subscription service as a perk, at half the regular $20 a month price.”

2/6/25 via The Briefing by Martin Peers

“To sum up: You can forget the notion that Chinese AI model DeepSeek’s success in developing a competitive AI model on the cheap would prompt U.S. firms to pull back on AI investments. Jassy, when asked about the DeepSeek issue, repeated what his counterparts at Microsoft and Google—Satya Nadella and Sundar Pichai—have said in the past 10 days. The essence of that argument: Falling technology costs actually lead to greater spending on technology.”

2/17/25

“The BBC reports: ‘YouTube is the most popular social media site in the US by far, with 83% of adults and 93% of teenagers among its patrons. It’s the second-most-visited website on Earth by most estimates, topped only by Google.com itself. But as the platform enters its third decade, the most basic facts about YouTube are still a closely guarded secret.’

The researchers found that much of the content on YouTube is not the work of professional content creators. The site is also vast video library for millions of people, organizations, and local governments. They were able to dig out some interesting stats that challenge many people’s perceptions about how the platform is used.

• The median number of views for a YouTube videos is just 41 views.

• 4% of YouTube videos haven’t been watched a single time.

• 74% of YouTube videos have 0 comments.

• Around 89% of YouTube videos have 0 likes.

• The median YouTube video is only 64 seconds long

• More than a third of YouTube videos are less than 33 seconds long.

• Only 38% of YouTube videos were edited before uploaded.

In other news: Indian matchmakers are embracing AI.”

2/18/25 via Andrew Chen

“To take an example of a single industry, scroll enough on social media and you’ll see a prediction like this: “AI will disrupt Hollywood!!!” This kind of assertion is usually accompanied with an amazing 30 second video that gives a taste of the latest in AI — the output of a bunch of video generation models, speech synthesis, and AI music, all cut in the way that is highly entertaining for 30 seconds. And the prediction mostly kinda feels right. And it is genuinely amazing how fast gen AI video models are progressing — today the output is still a little janky, but the rate of improvement is incredible.

…

So here are some ways to approach working with Hollywood:

• Pixar for AI aka a new AI Hollywood studio: If an AI startup can hybridize with a studio, then theoretically one could create the next major film idea, sell it to a major studio (Sony, Paramount, Warner Bros, and the like) and receive financing to make the film without raising additional venture capital. This would generate revenue, and the AI studio startup could continue with a series of these, and if one is a hit, it might become a very cost efficient way to get going. This is, plus or minus, the story of Pixar, Toy Story, and Disney of course. The biggest challenge here is that you’re required from a tech perspective to build something at the highest quality level for the silver screen right off the bat, and I don’t think Clayton Christensen would approve nor would it be the easiest thing to do. But this option is there. My colleague Jon Lai wrote an essay on this whole approach (and more) and it’s worth reading. One interesting side question here is whether you’d rather back the luminary Hollywood team who then goes and adopts tech, or would you rather fund a tech founder who then sells into the industry? Both are hard, but again, depends on your tastes. Many tech VCs would rather back the tech founder, but the dream of course is Ed Catmull with John Lasseter.

• AI-enabled/tech-enabled studio: There’s another interesting approach here which is to actually buy a production studio that already has a business, team, and process, and create an AI team working alongside them to aggressive build/buy/partner on all the tech needed. We’re seeing this attempted in entertainment already, but also fields like legal, customer support, accounting, and so on. This might actually be easier to do than to try to sell software.

• Sell tools: There are a ton of workflows within the film industry that could use AI tools today. Some are obvious, like dubbing, internationalization, etc. These are nice niches that might be good starting points. Or you could look at other issues, like digital asset management or video search. Or editing, including the ability to fill in gaps of video with AI-generated content, rather than reshooting. The failure case for all of this is that you’re selling tech into companies that have tight budgets and historically have not been great customers. And maybe your customers also hate AI. But on the other hand, if you can make people more productive and/or your new AI app works on an outcomes basis (like the total project fee to dub a film, rather than as a $/seat tool) then that might work too. But also might be hard.”

2/20/25 via UserMag by Taylor Lorenz

“‘Boring Reality’ is a Chinese open source project that creates videos that look like they were shot on your phone.”

2/24/25

> OpenAI identifies, bans Chinese group that used ChatGPT to build AI-powered surveillance tools to monitor anti-Chinese government content across X, Facebook, Instagram, and other platforms (More)

> Researchers report prototype battery material with a potential energy capacity double that of current Li-ion batteries; fluoride-ion device expected to find use in future electric vehicle applications (More) | How electric vehicles work (1440 Topics)

2/25/25

Apple Boosts US Investment

Apple announced yesterday plans to invest more than $500B in the US over the next four years, marking its largest-ever spending commitment. The initiative includes hiring approximately 20,000 new employees across multiple states, primarily focused on research and development, silicon engineering, and artificial intelligence.

A significant portion will fund a new AI server manufacturing facility in Houston, Texas. The Houston facility will produce servers supporting Apple Intelligence (the company’s AI-powered personal assistant system), shifting the production of these servers from outside the US to domestic facilities. The announcement comes amid renewed pressure from the Trump administration regarding tariffs on goods imported from China, where Apple assembles most of its products.

The investment follows previous commitments by Apple. In 2021, the company pledged $430B in US investments and has since expanded domestic manufacturing, including chip production in Arizona. See our 1440 Topics page on Apple here.

> AI startup Anthropic rolls out Claude 3.7 Sonnet, the field’s first hybrid reasoning model; platform lets users get real-time answers or request longer, more thought out responses (More) | Everything you need to know about generative AI (1440 Topics)

2/26/25

> Perplexity AI reportedly launching $50M venture fund to back early-stage AI startups (More) | Everything you need to know about venture capital (1440 Topics)

2/27/25

> Engineers integrate networked fiber-based computer into garments, allowing the wearer to track activity and monitor health conditions (More)

2/28/25

> Amazon unveils quantum computing chip, available for Amazon Web Services customers; device brings the company in line with competitors Google and Microsoft (More) | How does quantum computing work, anyway? (1440 Topics)

> New AI algorithm helps doctors analyze and predict health outcomes based on a series of medical images; approach may have use in diagnosing diseases ranging from cancer to Alzheimer’s (More)

> Intuitive Machines successfully launches second lunar lander; payload includes a Nokia-built 4G wireless system, which would establish the first Wi-Fi network on the moon (More)

Unless otherwise specified, all posts are from 1440, a daily newsletter. As the year went on, I started to gather more newsletters, such as Morning Brew. The data points will grow longer and more thorough starting in October.

None of the words below and in this series of newsletters are mine, I am simply copying and pasting from a few different newsletters, so I do not get any writing credit. This is also the reason for any formatting discrepancies. 1440 uses the carat (>), others will use a bullet point, etc.

There will be a lot of these emails in the next several weeks. I do not apologize. As I’ve tracked developments in science, I plan to also send out 2025’s news around our developing understanding of biology, natural systems, health, etc.

Some of these posts are longer. If anything is truncated in your email, just click “view entire post,” and you will.

12/30/24

Artificial intelligence giant OpenAI revealed long-anticipated plans to restructure the organization Friday, announcing it will transition to a public benefit corporation with an associated nonprofit arm. The company said the move will allow it to raise the capital needed to finance its massive computing needs while balancing the interests of shareholders and the public (see overview).

Now valued at $157B, the maker of ChatGPT originally began as fully nonprofit in 2015. In 2019, the organization moved to a hybrid model, with the relatively smaller nonprofit arm controlling a for-profit operation—a structure that played a role in a failed attempt to oust CEO Sam Altman in 2023.

In response to the new plans, an AI safety nonprofit joined an existing lawsuitbrought by Elon Musk—an early investor in OpenAI and CEO of competitor xAI—attempting to block OpenAI’s transition to a for-profit entity. The suit alleges the move violates OpenAI’s founding charter, among other claims.

12/30/24

> Jeff Bezos-owned Blue Origin receives license for a test launch of its New Glenn heavy-lift launch vehicle, successfully fires up rocket on launch pad (More)

1/3/25:

> Meta to introduce AI-generated characters with their own profiles that will share content, interact with users across Facebook and Instagram; release date to be determined (More)

> Federal appeals court strikes down Biden administration net neutrality rules, finding regulators lack authority to restore requirements internet service providers treat all data on their networks equally (More) | Net neutrality 101 (More)

1/6/25 via Popular Information by Judd Legum

AI costs American renters over $3.6 billion annually, according to new report

The property management software used by many corporate landlords, RealPage, is deploying AI to artificially inflate rental prices in the United States by more than $3.6 billion annually, according to a new report by the White House Council of Economic Advisers (CEA).

1/8/25

> Nvidia unveils $3K AI-powered desktop for researchers and students; system allows users to run many AI models locally instead of relying on cloud computing (More) | Google to build AI model capable of simulating real-world environments (More)

> Arizona receives approval to build 600-megawatt utility-scale solar power plant; one of the biggest solar projects in the US, the facility will power an estimated 180,000 homes (More)

1/9/25

> AI-powered analysis determines the best time to begin hormonal injections into eggs during the in-vitro fertilization process, leading to increased yield of mature eggs (More) | The technique that may one day replace IVF (More, w/video)

> OpenAI’s Sam Altman denies sexual abuse accusations from his sister, Ann, who filed a lawsuit this week alleging he abused her regularly from 1997 to 2006 (More) | OpenAI’s rise to power (More, w/video)

1/10/25

> Researchers engineer two new CRISPR enzymes that operate without triggering an immune system response; results may lead to safer, more efficient gene editing technologies (More) | How CRISPR splices and dices our genes (More)

1/11/25

Constellation Energy agrees to buy Calpine for roughly $16B.

The cash-and-stock transaction values Calpine at nearly $27B when including debt and creates the largest clean energy provider in the US. The acquisition combines Constellation’s nuclear expertise with Calpine’s natural gas and geothermal assets while expanding Constellation’s presence in markets like Texas and California—America’s two most populous, energy-consuming states.

1/12/25 via Doug O’Laughlin

Capital Cycles and AI

What is a capital cycle and why AI is building the mother of all capital cycles

The railroad and the internet bubble capital formation had multiple aspects. This current capital cycle also has a few specific capital formation buckets. I can define them simplistically as Compute and Networking, Power, and Data Centers. It’s funny how that echoes the past because Railroads were Railroads and Land, and the Internet was the Internet and Telecom.

In the case of AI, I think it will be power and datacenters that will be the real overbuilder. This overbuilding will worsen because there’s a significant lag between the two markets. The compute and networking cycle is much more responsive, so I think the demand lead time to supply response is a ~12-month response.

The problem is that power and data centers have a much slower reaction cycle. A datacenter takes months to years, and power takes years to add supply. The biggest opportunity is reshuffling power budgets because power will limit compute and networking this year. The datacenter and power part of the equation only really started to react to the market this past year (2024), and I think that is likely an indication that we will have some time before the capital cycle turns.

1/13/25 via User Mag by Taylor Lorenz

A $30M plan to take back social media from billionaires

Today, a group of former Twitter users who are fed up by the platform’s decline under billionaire control, are launching a new campaign to transform social media into a public good, free from profit-driven incentives, venture capital pressure, and politically-motivated censorship.

The project is called “FreeOurFeeds,” and it has launched with the support of big names including actor Mark Ruffalo, writer Cory Doctorow, businessman Roger McNamee, director Alex Winter, and others.

1/14/25 via Kyla Scanlon

The TikTok Ban Shows America Is At War With Itself

1/14/25

> Biden administration proposes rules on exports for advanced AI computer chips; select group of nations will have full access while countries like China, Iran, and Russia will face heavy restrictions (More)

1/15/25

> President Joe Biden signs executive order allowing the Defense and Energy Departments to lease federal lands to companies building AI data centers and clean power facilities (More)

1/16/25

> Meta unveils AI-powered speech-to-speech platform that can perform real-time translation across 101 languages (More) | Watch Mark Zuckerberg discuss the company’s push into all things AI (More)

> Meta Platforms plans to lay off 5% of workforce, or about 3,600 people, based on performance evaluations (More) | Short-seller Hindenburg Research to shut down (More)

1/17/25

> Google says it won’t add fact-checks to Google searches or YouTube videos despite a forthcoming European Union law requiring the practice (More)

1/20/25 via Kyla Scanlon

Trumpcoin and TikTok

1/20/25

> Apple suspends AI-powered news feature over inaccurate summaries of headlines and stories (More)

1/21/25

> X, Bluesky roll out vertical video feeds as the status of TikTok—the short-form video-sharing platform that relies largely on the format—remains in limbo (More)

> Apple under investigation by Belgium for allegedly using “blood minerals” from the Democratic Republic of Congo—mineral sources that finance armed conflict and are associated with human rights abuses (More)

1/22/25

> Chinese AI lab DeepSeek releases large language model it says rivals OpenAI’s o1 model in reasoning and performance; model released under MIT license, allowing commercial use (More)

1/23/25

> Google invests additional $1B in generative AI startup Anthropic, bringing its total investment to over $3B and building on a previous 10% ownership stake (More)

> Are We Losing the Ability to Write by Hand?

The Guardian | Christine Rosen. In a technology-driven world where digital devices have largely replaced handwriting, do we risk losing the cognitive and sensory benefits that come with it? (Read)

1/27/25

> Engineers develop 3D-printed ultralight material roughly five times stronger than titanium; process uses machine learning to produce strength-enhancing micro- and nanostructures (More)

1/27/25 via The Information by Martin Peers

Talk about a jittery market! Shares of some tech firms exposed to artificial intelligence tumbled today, prompted by new evidence that China’s DeepSeek is giving AI folks in the U.S. a run for their money. Nvidia and Broadcom each fell 17%, while Alphabet dropped 4% and Microsoft 2%.

1/28/25

> Startup Manas AI, co-led by LinkedIn cofounder Reid Hoffman, raises nearly $25M for developing new treatments for cancer and other autoimmune diseases using artificial intelligence (More)

> What Happened to Hanging Out on the Street

Bloomberg | David Zipper. A study using AI to analyze video footage from 1980 and 2010 in three US cities found urban pedestrians now walk faster and socialize less, raising questions about whether technology or public spaces are to blame. (Read)

Comment with any insights or feedback as these updates roll out.

Four months later, in early December, the West Coast held its own version of Woodstock at Altamont Speedway, east of San Francisco Bay. The show was a disaster, famous for the Rolling Stones headlining act, during which a member of the Hell’s Angels, acting as event security, stabbed to death a crazed fan with a gun. The event prompted the famous lyrics by Don McLean in “American Pie”: “As I watched him on the stage / my hands were clenched in fists of rage / No angel born in hell / could break that Satan’s spell.” It was a dark day for music and the countercultural movement which spiraled in fractal spangles from a hallucinogenic centrifuge; the medical tent at Altamont was full of attendees having bad trips off impure LSD. San Francisco music writer Joel Selvin called the unholy morass of 300,000 people “toxic mass psychosis.”

The origins of this psycho-musical revolution and its dissolution can both be found in The Electric Kool-Aid Acid Test by Tom Wolfe, a 1968 gonzo-journalism investigation of Ken Kesey and his band of free-spirited cosmic explorers traversing the nation in a tie-died bus and evangelizing the wonderous properties of this new chemical compound (LSD).

The trip started when Ken Kesey moved to the Stanford-based Perry Lane Community in 1958. Living in these oak-lined streets immersed Kesey in bohemian royalty. Here, he met a psychology student who introduced Kesey to the CIA-funded MKULTRA experiments that paid out volunteers $75 for monitored tests on psychoactive substances. Kesey quickly became an evangelist for LSD and took it back to the rest of Perry Lane. The community split on the issue of drugs and the future went with Kesey when he credited the success of One Flew Over the Cukoo’s Nest to an LSD trip that unlocked a schizophrenic character’s perspective. When the bulldozers came to rebuild most of the historic Perry Lane cottages, Kesey dragged a piano out in the street, the “oldest living thing” in the community, and smashed it to pieces with his friends. They had already purchased land at La Honda, where they would reinvent the bohemian communal ideal with a scorching desire for the immediate and full experience opening the doors of perception with LSD.

But Kesey didn’t destroy everything from the past, bringing along with him a new friend he made at Perry Lane, namely Neal Cassady, or as Jack Kerouac would refer to him in On The Road, Dean Moriarty. Cassady was also friends with Alan Ginsberg, who associated with Timothy Leary and Richard Alpert, the two Harvard psychologists who were rejected by the academic community in 1963 for their experimentation with the psychiatric potential of LSD. So these seekers from the 1950s found what they were looking for in an animating substance that stripped the self of its protective armor and lit the mind to its whole potential in seeing the world for all its fiery glory.

The central motif for these American countercultural movements is motion through freedom of expression. To Kerouac and Cassady, the car was their primary source of freedom, and the subject of On The Road. Kerouac famously wrote that book on one 120-foot scroll of paper that he taped together like a cross-country interstate subjected to the banging wheels of his typewriter. But the countercultural love of that-which-is-constantly-unfolding is always primarily expressed through music. For the Beat Generation of the 1950s, jazz was the jailbreaking, expansive sonic artform that defied the rigid world of conformity. Ken Kesey and the Merry Pranksters took these ideas to the next level, as the previous forms were clearly not enough to unshackle the world.

From La Honda, Kesey’s psychedelic troupe bought a bus, painted it in psychedelic swirls, and took it all around the country. Kerouac’s dream took a new dimension in every sense. The car became a tripped out bus loaded with microphones, street-facing speakers, and a garish paint job. The few people in Kerouac’s car became several. The drug of choice passed through a dimensional time-warp as well, with marijuana transforming into the all-encompassing experience that is LSD. Just as LSD went hand-in-hand with jazz’s euphoric revelry, it also demanded the birth of a new music genre as animating and forceful as the omnicolored world it presented.

Back in California, the Pranksters started carrying out their “acid tests,” where they would spike garbage cans full of Kool-Aid with LSD and turn on unsuspecting partiers to a freeing parallel universe. Given that LSD increases one’s sensory intake, amplifying sensual experience was paramount for the Pranksters and their Acid Tests. The parties included sophisticated light displays, strobes, and dozens of microphones that projected different oddities through a massive hookup of omnipresent speakers. At many of these Acid Tests, a new upstart band called the Grateful Dead performed, with lead guitarist Jerry Garcia having met Ken Kesey back in his Perry Lane days. Thus the scale-defying rebellion that is jazz transformed into reverberating run-on jam sessions. The production level of these Acid Tests may even have been a prototype for the sophistication of live music concerts.

The obsession with freeform expression in these countercultural circles depends on a type of receptivity. Throughout The Electric Kool-Aid Acid Test, whenever others verbally express their total understanding of a thing, they “grok” over it. The word comes from the 1961 book “Stranger in a Strange Land” by Robert A. Heinlein. In that book, it is a Martian word that originally meant “to drink” and became popular in the emerging cosmically-obsessed counterculture. A party in the 1960s included a lot of expounding, known as “rapping,” by people who felt pent up by the oppressive world. Rapping meant to spell one’s unfolding thoughts in continuous and creative ways, a freeform exercise in communication akin to the jazz solo. Rapping was met with grokking, which itself reached a height when the Merry Pranksters had an unlikely run-in with the Hell’s Angels.

At the same time that the Pranksters were running Acid Tests, Hunter S. Thompson was creating a whole new style of writing and investigation. He called it “gonzo journalism,” and through his 12-gauge-whiskey-slug commitment to vice, he exposed America’s underbelly as one of its proud degenerates. He rode with the Hell’s Angels for a year and broke the story on them with his book Hell’s Angels: A Strange and Terrible Saga. Mind you, the saga was terrible in the way that Thompson likes. You see, he could voice public opinion about a force of chaos in a way that also pays homage to the destruction it revels in.

This too was the job of Tom Wolfe as he dissected the Pranksters a few years later. There is an omnipresent dirtiness to these societal rejects, perhaps a repugnance at such aphorisms as “cleanliness is close to godliness,” or more closely an agnostic welcoming of all types of matter, particulate and otherwise, as it adds to the soupy concoction that is one’s own unique flavor. This in turn influences the dialect of writers such as Wolfe and Thompson, who are themselves channeling a runaway generation with stream-of-consciousness excitement in breaking the chains of a confusedly rigid world. For Thompson and the Hell’s Angels, the motorcycle was a brapping growl that would translate into pages of run-on menace. For Wolfe, LSD was the transformer that loosened lingual bounds and created spangling word-combinations that would only be conceivable in a hallucinogenic dream.

When Ken Kesey connected with his friend Hunter S. Thompson to link his pirate band of whimsical fools with the Hell’s Angels, the clash of two rebel groups created a strangely harmonious union. The Hell’s Angels loved this new chemical compound. It even made them nonviolent, if only briefly. With the tension dissipated, the Angels and the Pranksters would sit in a circle and grok over each other’s uniqueness.

The only reason this 1965 connection even happened was because the many groups who saw themselves as rebels wanted to associate with this extreme group of motorcycle outlaws. Bohemians, hippies, and college professors alike were calling up Hunter Thompson to see if he could get the Angels to come to their party. They became a prop for the rebellious who didn’t know what they were getting into. The Angels were many times willing to oblige, because at these parties were substances and women they could use simultaneously. But just as the Hells Angels were becoming hip to the “scene,” they rejected it. They wanted nothing but to crack skulls. They were rejects whose fury at the ways of the world was also their isolating creed.

In 1965, the same year as the Prankster/Angels union, a dozen Hells Angels unsuspectingly lined up with the Oakland Police against an anti-war demonstration fifteen thousand strong. Alan Ginsberg and some others realized how the Hells Angels were not on their side, or any side at all, but others in bohemian circles continued to enjoy their presence. The Grateful Dead kept them around as security, leading to the Rolling Stones hiring them at that fateful day in Altamont.

So the countercultural revolution ended in disarray, and the time period remains a callback for the world of corners and edges that managed to corporatize and profit off of intergenerational bohemians. Such is the story of the startup world, which excitedly tries to tell us time and again that they have found an answer in the same golden-green enclaves of Palo Alto that promised worldwide regeneration those decades ago. Even Grok has a second wind in Elon Musk’s AI company, which just like rapping groups of hallucinating friends will spit words for the freewheeling fun of it according to text prompts.

Elon Musk made the psychedelic dream a digital reality. He, among others, brought us to the cosmos through a web of satellites that transport us in multicolored screen displays.

Musk moved to California in 1995 to found Zip2, a company that effectively digitized the yellow pages of a city to prototype the ease of browsing and locating a city’s businesses. Zip2 sold in 1999, and Musk went on to found X.com, which merged with PayPal before Musk bought back the rights to the X.com domain. The rest is history, starting with SpaceX, Starlink, and Tesla, and eventually leading to xAI, which uses its AI chatbot affectionately called Grok. Musk named it Grok so that it may “empathetically understand” what it is saying, or merge with the subject matter, rather than just computationally generate successive strings of words. Musk is notably influenced by sci-fi books from this countercultural era.

Unfortunately for the fast-talking hard-grokking writers at the beginning of the counterculture movement, their language was eventually coopted by the machine. And at the center of that story is a man who, like many of the hallucinogen-taking seekers of a new world, sought to reclaim his home in the stars. But instead of finding that power within, he transliterated it into LED screens and became a frontrunner in the digitization of all things. In the process of trying to realize the dream of free information and lightspeed travel, that aspiration itself became corporatized. Now, as AI bots such as Grok learn more about us every day, we stand on the precipice of a brave new world that instead of championing individual rights, seeks to serve a new master, one which can outperform us in all levels of intellect, but whose goals will remain forever unclear and definitively separate from us of the particulate, the dirty, the flesh-and-blood, the children of the mud, us who are meant to yearn for the stars but cannot apprehend them until we pass back through the earth in humility.

I’ve spent about a week in the beach towns north of San Diego, and a few weeks near San Diego’s city center. The two are vastly different, illustrating the contrast between the raw geographic potential of an area and its evolution after layers of vested interest stake a claim.

Mission San Diego de Alcalá

Originally built in 1769, the first mission on the West Coast marked Europe’s arrival on its claimed land. Mission San Diego de Alcalá moved to its current site in 1774, following the native Kumeyaay and their reliable knowledge of food and water sources. Naturally, this threatening building was attacked by the Kumeyaay the next year, and rebuilt and fortified from 1776-1790. Tensions eventually eased, and the number of Kumeyaay baptized by this outpost jumped to 1405 in 1797. In 1821, Mexico gained its independence from Spain, and in 1848, the Treaty of Guadalupe Hidalgo effectively turned the land and its future over to the United States. The Kumeyaay were likely either oblivious or confused by all the changing of hands. Regardless of the paperwork, the waves continue to break on their own accord.

Balboa Park

One thousand four hundred acres were first set aside for “City Park” in 1868. Twenty-four years later, Kate Sessions, “The Mother of Balboa Park,” started its beautification process with a 100-tree-a-year planting pledge.

The park hosted a 1915 Panama-California Exposition, to celebrate the construction of the Panama Canal one year prior. The fitting celebration made San Diego increasingly important as a strategic holding for the nascent American Empire. Suddenly, San Diego was the first US port of call for northbound ships passing westward through the canal. The California Building, California Tower, and connecting Cabrillo Bridge were built for this Exposition. The celebration also left behind several exotic animals that eventually became the world-famous San Diego Zoo. Several temporary buildings were erected for the Exposition as well, before being replaced with permanent replicas for the 1935 exposition. The architects used a combination of Spanish styles including Baroque, Plateresque, Churrigueresque, and Rococo. Maybe we need to bring back expositions?

Hotel Del Coronado

The Red Hot Chili Peppers call California “the edge of the world and all of Western Civilization.” There is something inspiring about staring out across the Pacific, knowing that the colonial thrust of world population redistribution ended here.

Two hotel founders (who had never before founded a hotel) viewed Coronado Island just off the mainland in all its natural beauty, and imagined the possibility of building a magnificent hotel that would be the “talk of the western world.” They bought the island in 1886, and started supplying it with water and electricity. They also built a ferry company and a railroad company, and sold plots of land on Coronado by auction to fund their venture. Construction began on the north face, which was simpler in design, so that the 250 unreliable workers would accrue skills as they worked. Wood came from all around the country for this beautiful building: Douglas fir, California redwood, hemlock, cedar, Illinois white oak, and Oregon sugar pine.

This was a marvel of technology at the time. With the first commercially practical lightbulb invented in 1879, Hotel del Coronado was powered and lit by a new energy source that was taking the world by storm. Other novelties included a steam-powered hydraulic elevator, telephone service, and a fire sprinkler system.

There are too many interesting facets of the hotel’s architecture to name here. Hotel del Coronado holds some beautiful and mysterious stories as well: a suicide here, a movie star there, some presidents and film crews. Frank L. Baum, who wrote The Wizard of Oz, stayed here for months at a time over thirteen years from 1904 to 1917. He wrote at least three books in the Oz series here. The connecting bridge was built in 1969, and the rest is history.

USS Midway Museum

San Diego became a shipbuilding hub in 1918, four years after the Panama Canal made it a strong American port. After World War I, the private shipbuilders ran out of cash, and the Navy acquired their land in 1920. Since then, the base continued to add training schools and other military infrastructure as radio and torpedo technology advanced. With World War II, San Diego became an increasingly strategic position for the Pacific Theater of WWII (why is it a theater, when did we decide that we was Playin’ wit deez bombs).

The USS Midway was named after the pivotal Battle of Midway in 1942. It was completed in just 17 months and commissioned on September 10, 1945, one week after the end of the war. Too late to raise hell, just on time to be a dormant threat. The USS Midway studied abroad in Japan, the Persian Gulf, and the Philippines, doing all kinds of things that aircraft carriers do. Her brawling days over, the USS Midway was decommissioned, and in 2004 she fulfilled her lifelong dream to teach history to all who climbed aboard her port in the San Diego Bay.

Salk Institute

Jonas Salk, who discovered the polio vaccine, approached famed architect Louis Kahn with his dream of a research center where scientists would live and work in monastic harmony. The final design in 1965 cut out the community conference center and the village, and the existing buildings house Salk Institute laboratories. In Salk’s original instructions, he invoked the name of Picasso as the desired level of artistic acclaim. The regular irregularity of these concrete shapes are cubist-adjacent, and would make the late Spanish painter proud.

The architectural form here is brutalism, which you will also see in the Geisel library below. Brutalism uses bare concrete to create a hulking sense of presence. The style is much less popular now than it was in the mid-century, as it became aesthetically associated with totalitarian bureaucracy. There are some elegant examples of brutalism though, and this is one of them. The teak wood accents certainly help. Since the Salk Institute and its prestigious founder, San Diego has become a leading center for biotech firms specializing in cancer research, genetic disorders, and infectious diseases.

Geisel Library

One of the fascinating tricks in good architecture is to make something colossal seem to float in the air. This experiment in brutalist futurism makes concrete hover like a UFO at the centerpoint of UCSD. Attempting to prioritize light and expand the building upward from a single entry point, the Geisel architect originally planned to build a sphere. The final plan, built in 1970, resembles the interior scaffolding that would remain if one built a sphere-shaped building and removed its shell.

The Geisel Library has two submerged stories and six above ground. It is named after Theodore Seuss Geisel, aka Dr. Seuss, who donated $20 million to the preservation of this terrifyingly whimsical building. Dr. Seuss lived and worked in San Diego in the final stage of his life. UCSD honors him with not just this building but an on-campus statue.

Torr Kaelan

Emergent urbanist concepts fold neatly into this oragami jewel on the outskirts of a redeveloping neighborhood. In 2015, when Torr Kaelan was built, East Village had a notably tougher reputation, especially to the south and east, right where this building is situated. Perhaps Torr Kaelan accomplished its goal of extending creative development into struggling parts of town.

Two offices and two townhomes make up Torr Kaelan, adding dimensionality through diverse interactions. The building splits down a glass center as if struck by lightning, taking in golden light by day and pouring it back onto the street by night. The entire building is solar powered and collects rainwater for the gardens. Torr Kaelan aptly houses the Quigley Architectural offices, who designed both this building and the equally interesting Central Library just around the corner. Perhaps Torr Kaelan’s gaelic name, meaning “rock outcropping,” is a nod to Frank Lloyd Wright’s Welsh-named Taliesin.

Still, the main complication with East Village is the same as it is for many cities on the West Coast and across America. The homeless population congregates around the library here, making the neighborhood feel unsafe. For all the ingenuity and wealth compounding upon itself in these urban centers, we still haven’t solved that problem. Most of the time, we can’t even afford to see the homeless as human beings. The business of building and maintaining exclusivity is simply too demanding. And still, the waves break on their own accord. Some days I look at the question-mark-curl of the waves and I wonder: what will make sense of this world?

Honorable Mentions

Bali Hai

Built in 1953, “The Hut” was originally associated with a 1935 movie starring Clark Gable. A restaurateur bought it in 1954 and renamed it Bali Hai. Its Polynesian theme is accented by the use of 15 rare woods, restored in 2010. It is a beautiful venue, with 360 degree views that make this destination a phenomenal place to observe the San Diego skyline.

Rady Shell

This 2021 acoustic wonder speaks for itself. Built on manmade land in the bay, the Rady Shell is the summer home of the San Diego Symphony. San Diego would hardly be complete without its own beautiful venue where the sun, the sea, and music meet.

Many friends and family after the 7/17 funeral wanted a copy of my eulogy. I decided to put it up here to memorialize the event. Others who didn’t know him may find it interesting as they experience loss and grief in their own lives. One thing to note is that I wrote this as a speech to be delivered, so it may read differently or awkwardly on the page. Also, in my speech, I didn’t have time to get into who Dale Carnegie and Ernest Shackelton are. If you don’t know them, I’m sure the internet will help you there. If you don’t know who Genghis Khan is, nobody can help you.

The poem at the end began as something I started writing about people I would meet while traveling. When my dad died, I made it about him. As such, it continues to take on different meanings, which I think is appropriate as one particular loss ultimately puts you in conversation with a world full of people who are passing through each other’s lives, just trying to make it through the day.

From as early as I can remember, I could tell that my dad was odd. He always had this peculiar effect on people. People would confide in him. They would smile more than usual. They would make exceptions to their rules. They would drop the façade and start to breathe. It’s like there was this electrical current running through him, sparking people to life.

Truthfully, the legend of my father followed me everywhere. People might stop me in the bathroom at church and tell me, “Your father is a good man.” They might look at how I sewed up my broken backpack with a shoelace and say, “You’re just like your dad.” It took me a while to piece together what drove him and just how he became so embarrassing. I think the easiest way to explain him is through his three idols: Dale Carnegie, Ernest Shackleton, and yes, very predictably, Genghis Khan.

In Dale Carnegie, my dad learned how to live outwardly, always regarding others. What I, when I was younger, regarded as general goofiness, turned out to be in the end an alchemy that withstood the greatest hardships by focusing on others needs. He fought the deadening forces of the world with spectacle and tenderness, blazing from his joy in simply sharing life with others. And people took notice and loved him for it. Every Christmas Eve, we’ve watched this movie called It’s a Wonderful Life. We just sit down as a family and quote the damn thing from beginning to end. It’s atrocious. There are a lot of things about that movie that relate back to my dad, but there’s just this one line… “No man is a failure who has friends.” Well then my father just might be the most successful man in the world.

In Ernest Shackleton, Dad learned to withstand anything by taking every difficulty as an opportunity to show that the light prevails. Life was nothing to my father if not an adventure. And he always said, “An adventure begins when something goes wrong.” But he recognized what mountains were worth climbing and what weren’t. When he won a bid to climb Mount Everest, he declined. You see, his mom died when he was 10, and his best friend died while rock climbing at 18. Although this didn’t stop my dad from climbing the tallest mountain in North America, it might’ve worked in his mind when he declined climbing the tallest mountain in the world, because he had his first baby on the way, and he didn’t want to risk leaving his family behind. Dad made more sacrifices than anyone ever should. And when we as a family say we wouldn’t wish his paralysis on anyone, anyone, we mean it. Never before or since have I been face to face with how cruel the world can be to take almost everything from the one person that appreciated it the most. But in his response to that tragedy, Dad taught us a depth of empathy I couldn’t have possibly otherwise learned.

And he couldn’t have done that without the strength of will exhibited in his estimable colleague Genghis Khan. You wouldn’t know he was constantly in pain, or that it might take him four hours to do a simple task on the computer, because he just took up every burden without mentioning it. If it weren’t for this mindset, he would’ve made the extremely practical decision to stop biking after he became paralyzed. But several years after his accident, he worked up to biking all the way around White Rock Lake with whatever little leg function he had. Dad was unconventional. He was unpredictable in every way, except that he would seek out every challenge with pure grit and determination. Although his business in Group Dynamix was so unique that many weren’t sure what the market was, or what the product was, he charged forward with this idea that he could provide a solution to loneliness by unifying people through adventure. And people leave that building remembering something they forgot. That life shouldn’t be taken too seriously, and that the important stuff comes when you least expect it.

Life is strange since my dad died. You can feel the vacancy where someone so large once was. But in that vacany is a stillness. And in that stillness is a peace. I was going to finish with some song lyrics, but instead I wrote a poem that I’ll now share.

I wish for one more sleepless night, praying you might walk again.

Making deals with God that I know that I can’t keep.

I remember how you ran through the world

like a shatterspark shackleton.

You were a cup of wine to all you met,

so flood my heart,

so flood my heart.

But I’ll see you when I see you,

in the golden hall of the sun.

At the fountain at the heart of the city,

beyond the obsidian veil.

Oh that I were not already dead.

I expect to see you just around the corner.

I’m turning, turning, and you’re never there.

Last I was here was with you.

Oh that we were not already dead.

But there’s a rumor that love never dies,

and a tree of life at the center.

So I’ll see you when I see you,

beyond the wall of the stars.

We will walk up to the pillar together.

And I’ll see you when I see you.

We only have until forever.