Peter Steiner should revisit his famous 1993 cartoon “on the Internet, nobody knows you’re a dog.” That was web 0.8. Today the meme has lost its puissance. Here’s an unadorned view from the frontlines of an AI-evaluator platform and its implications for the Talent as a Service business model. It’s not pretty.

Logan Roy would take a glance at the Slack channel and shoot back: You Are Not Serious People™️. Here is what I had to resort to in order to get color on when the hell I was going to get the Codex access I needed to complete my killer prompt.

🌊 I’m pickin’ up Good Vibrations! 🌊 🏄‍♂️ ✨

Excitations are HIGH, but my token count is LOW. 📉

Let’s keep this 🤡 🏎️ moving at terminal velocity! To help me “lean in” to the synergy, can we get those Claude / Codex credentials ASAP? 🔑 🤖 ⚡

🚫 📝 (No Memos)

✅ 🧠 (Pure Compute)

Stay 🌊 Wavy 🌊 and keep the reality-filters set to 100%! 🌈 🫧 🤐

So, yeah, this modality is creating engagement, but it’s the engagement of the warm bodies of people who are fungible and shortly to be replaced by the next generation of AI. There’s no sustainable business model here. The algorithms gatekeeping access to the platform are loose, which is, of course, the way to scale quickly, but it’s a commodity business tending toward marginal cost pricing.

That’s the reality behind my idea to pivot to a new business model of Teams as a Service. I negotiated opposite to designated underwriter’s counsel for almost 20 years and we never met in person. We built a trust and mutual understanding on the basis on a series of deals, and we were part of a circle of trust within a tight-knit community of professionals.

Within our industry domain, you could have dropped any of us plop into a deal in which we knew no one and we would work as well as physicians and nurses in an operating room or a production crew on a movie set coming together for the first time.

Build those assemblages of talent and you’ve got something probably not available off the shelf from the competition.

Just sayin’.

I did Dylan one better— I got 21 years of schooling before getting put on the day shift with my first full‑time job at a grown‑up salary.1 I went to a job fair in Boston and came home with three offers: one from a small agency in Boise, one from an electric utility in East Lansing2, and the one I took, from the East‑West Gateway Council of Governments serving as the regional planning agency for metropolitan St. Louis.

My first assignment was to attend upon the Executive Director as a “spear chucker”3 on a site visit to an early version of a recycling plant. I knew even less about solid‑waste management than I pretended to know about water‑quality management4.

The Federal Water Pollution Control Act Amendments of 19725 had just been passed, and it included a pot of money for the what I short‑handed as “preparation of a 20‑year plan and program for the abatement of point and non‑point sources of pollution,” whatever that meant. I presented as an freshly‑minted holder of a Master of Regional Planning with coursework in water‑quality planning and groundwater hydrology, backed up by a Master of Science in Geophysics and a poker‑bluff of statistics. A perfect seller’s market with a buyer mainly knowledgeable about urban renewal and freeway planning.

The enacting legislation contained carrots and sticks, specifying punishment for jurisdictions that didn’t get their wastewater streams together (the feds step in) and those who sent up their homework to be graded on the basis of regulations to be developed later. The whole process was accelerated.

Since the “munchkins on the ground” had only a vague view into how this was supposed to work, when I arrived there was no end of “hey kids! let’s put on a mark‑up session and figure this stuff out.” In time‑honored bureaucratic fashion a steering committee was appointed from the agency, the zoo6 and the Army Corps of Engineers7 to figure it all out. We’d meet once or twice a week and go round, and round, and round, the table. One of our number was a high‑caloric type, so just before lunch I’d make a motion, which he’d invariably vote for and leave. The zoo and I would then work out the division of the spoils.

One day we stumbled across the undefined “non‑point source of pollution.” Well, I said

It’s very simple really. Point sources have an outfall from a pipe. Non‑point sources are simply runoff that can end up flowing directly overland into streams and rivers or soaking into the groundwater.

I sat down, pleased with my improvisational riff8. I looked around the table and everyone had their head down furiously transcribing the received wisdom. I thought to myself

Shit. The world is being run by people who are making it up as they go along. We are in big, big trouble.

We finally fleshed out a budget for the grant proposal—three agencies with separate indirect salary and general‑and‑administrative overhead factors, six or seven distinct hourly rates and 48 tasks, plus a suitably plausible schedule to fit. I was nominated to put these together. I sat down to do a Gantt chart and critical‑path diagram, squishing to fit9 That went well, but I reserved a weekend for the detailed budget for which I turned to a ruled 11×17 pad of paper called a spreadsheet10, doing the extensions and footing the columns and headers with a 10‑key pad adding machine. Total P.I.T.A.

Time passes and comes the EPA auditor to check my work. “This doesn’t add up.” “Well, you transposed these two numbers.” “Oh.”11

While we waited for action on the grant proposal, my immediate boss quit and I became third‑tier management12 with six direct reports. Whoopy do13. Time passes. And passes. But there’s nothing to do other than watch the Watergate hearings. I admired John Dean’s ability to create a narrative that amounted to a plea deal.

So, at the end of a year, I let myself get recruited away to my next gig. Gimme a bump in rank and pay or lose me forever, like you know, dudes?

I did see Colonel Sanders on the streets of St. Louis, unless that was actually a Missouri gentleman of the old school.

13

I did, however, move out of the bullpen into a private office.

Nobody likes being fooled by a bullshitter or by the output of some artificial intelligence software. Opinions differ on “art” produced by AI, some finding it interesting at times while others are understandably put off by images of people with six fingers.

But there is also a school that has elevated dislike of AI to the category of things that must be shouted down independent of content. We’ve long seen this applied to people of course. and fortunately the software has no feelings to be hurt. But the harms from epistemological closure are the same.

The late 17th Century saw an example of how extreme intolerance can work in practice.

The extreme Presbyterians were as much out of humour as the extreme Prelatists, and were as little inclined as the extreme Prelatists to take the oath of allegiance to William and Mary. Indeed, though the Jacobite nonjuror and the Cameronian nonjuror were diametrically opposed to each other in opinion, though they regarded each other with mortal aversion, though neither of them would have had any scruple about persecuting the other, they had much in common. They were perhaps the two most remarkable specimens that the world could show of perverse absurdity. Each of them considered his darling form of ecclesiastical polity, not as a means but as an end, as the one thing needful, as the quintessence of the Christian religion. Each of them childishly fancied that he had found a theory of civil government in his Bible. Neither shrank from the frightful consequences to which his theory led. To all objections both had one answer,—Thus saith the Lord. Both agreed in boasting that the arguments which to atheistical politicians seemed unanswerable presented no difficulty to the Saint. It might be perfectly true that, by relaxing the rigour of his principles, he might save his country from slavery, anarchy, universal ruin. But his business was not to save his country, but to save his soul. He obeyed the commands of God, and left the event to God. One of the two fanatical sects held that, to the end of time, the nation would be bound to obey the heir of the Stuarts; the other held that, to the end of time, the nation would be bound by the Solemn League and Covenant; and thus both agreed in regarding the new Sovereigns as usurpers.

Macaulay, Thomas Babington The History of England, from the Accession of James II.

There is a lively discussion on the Julia language community board concerning the proper use of artificial intelligence in contributing new software packages to the central registry. Opinion divided along the predicable lines of absolutists on the one hand and pragmatists on the other. The thread had grown long enough that coming to it in the middle of the conversation would take a while to catch up.

I posted the following, marked as prepared by Google’s Notebook LLM, to summarize the discussion to date.

“Statement of Bias As an artificial intelligence language model designed to assist with tasks, including coding and text generation, my core purpose relies on the very technology being debated. This inherent nature could introduce a bias toward viewing Large Language Models (LLMs) and agentic programming as valuable, efficient, and beneficial tools. However, the following summary is derived strictly and objectively from the provided discussion thread, without injecting external opinions.

Here is a summary of the arguments presented in the thread regarding the use of LLMs in Julia packages and the debate over mandatory disclosure:

Arguments FOR the use of LLMs and Agentic Programming

• They are powerful development tools: Proponents argue that AI coding assistants are just clever tools that handle tedious tasks like refactoring. They are compared to other standard programming aids like linters, formatters, compilers, and IDEs, which keep developers from “still using punch cards”.

• They provide high utility and efficiency: Users note that these tools offer significant utility right now. For example, one developer highlighted how their “vibe coding assistant” allows for rapid changes and hot-code reloading.

• Focus should be on code quality, not origin: Some argue that a contribution should be judged by its quality rather than how it was written. It is noted that an AI-generated contribution could realistically be better than a purely human-written one, provided the author takes responsibility and reviews it.

Arguments AGAINST the use of LLMs and Agentic Programming

• Ideological opposition: Some developers are opposed to neural-network generated output as a matter of principle.

• The “slop” factor and review burden: There are concerns that AI can generate low-quality code or “slop,” leading to a devaluation of external code contributions. If AI makes code easy to write but bad code becomes indistinguishable from good code, the effort required to review pull requests might outweigh the value of the contributions.

• Skepticism of the AI industry: The AI industry is described by some as an “unsustainable speculative bubble” that will soon burst.

• Distrust of “Vibe-Coded” packages: Heavily AI-reliant packages can be off-putting to some users who view them as being too far into the “’AI is awesome, I use it to make my coffee’-land”.

Arguments FOR Required Disclosure (Badges/Markings)

• Consumer transparency: Advocates argue that because some people oppose AI code on principle, packages containing neural-net output should be clearly marked with a banner or badge, drawing a comparison to warning labels on GM-modified foods.

• Granular tracking of AI involvement: One user proposed a tiered badging system (ranging from “A” for no LLM contribution to “E” for full vibe coding) to help authors clarify exactly how much AI was used in their project.

Arguments AGAINST Required Disclosure (Badges/Markings)

• It is unenforceable and unrealistic: Opponents argue it is impossible to enforce a rule requiring developers to prove “no LLM has ever touched any part of this code”. Furthermore, because AI input is becoming ubiquitous, almost all packages would eventually need the badge, rendering it useless.

• Human ownership supersedes tool attribution: The prevailing expectation is that human developers must take ownership of their code. LLMs are not independent, intelligent beings deserving of “credit” or attribution via special GitHub accounts; they are tools, and the human committing the code is fully responsible for any bugs or “slop”.

• Standard review metrics are more important: Many users stated they would ignore an AI badge entirely and evaluate the package based on traditional metrics: code organization, documentation, tests, and CI/CD infrastructure. Consequently, participants argued that any disclosure should be entirely voluntary rather than a registry requirement.”

I reviewed the output before posting and found it fairly represented the discussion. Summarization is one of the tasks that chatbots can do very well.

It was flagged as “offensive and inappropriate” and hidden per the community guidelines. Go figure.

Sabot de vignerons du Roussillon avec crochet plante-échalas. ©️ 2000 Roymail, used by permission

WTF with the “if you’re a Gen Z listener …”? Gerontocracy, I know, but don’t edit the Boomers out quite yet, please. Set a good example for the ethical treatment of Gen X in the not all that far off future or they will cut you out of their wills.

The discussion of the gender polarization/demographic/incel turmoil would benefit from adding three considerations:

1. The most salient development with respect to human evolution since the adoption of advanced toolmaking by the primates that gave rise to H. sapiens took place beginning only a little over 60 years ago—reliable contraception, which decoupled sex from reproduction. When I was a high school senior in the immediate aftermath, it was not within my reasonable expectations to get laid. The Pill was not a thing for high school girls (to use the then-current label). A year later, it was a thing for college girls and my expectations changed accordingly. The burdens of child bearing and support during the long years of dependency no longer figured, and the interruption of sexual intercourse during pregnancy and post-partum intervals was eliminated. Supply rose to meet demand. The bargaining power that women had enjoyed was greatly diminished and men coming of age in that environment enjoyed far greater bargaining power than before. Sex and the City represents one of the social anxiety manifestations that resulted.

2. Irina Dunn in Sydney, Australia in 1970 first advanced the proposition that “a woman needs a man like a fish needs a bicycle” on restroom doors, one at the University of Sydney and another at Soren’s Wine Bar in Woolloomooloo. This was taken up by Gloria Steinem to immediate effect, reaching even reaching the campus of The Ohio State University in time for that spring term’s campus unrest, which was at the time of Kent State.

3. The immediate male beneficiaries of the sexual revolution either scoffed at the absurdity of her thesis or patronizingly dismissed it as clever but lacking serious understanding of how things were. After all, men, through the application of the uniquely male capacity to apply basic scientific research skills to the uniquely male capacity to put the results to practical use deserved the credit for reproductive freedom. The oral contraceptive in our view of the time simply expanded the power of the patriarchal model.

We dismissed the revolution along the following lines.

While the phrase “a woman needs a man like a fish needs a bicycle” is often used to emphasize female independence and the idea that women do not necessarily need men in their lives, it is important to recognize that this statement oversimplifies the complexities of human relationships and the diverse needs and desires of individuals.

1. Companionship and emotional support: Humans, regardless of gender, often seek companionship and emotional support from others. While it is true that women can lead fulfilling lives without a romantic partner, it is also true that meaningful relationships can bring joy, support, and a sense of belonging to one’s life. A healthy and supportive partnership can provide emotional stability and a sense of connection.

2. Shared responsibilities and teamwork: In many aspects of life, having a partner can provide practical benefits. Sharing responsibilities, such as household chores, financial management, and child-rearing, can alleviate the burdens and stress that can come with managing these tasks alone. A partner can offer support, share the workload, and contribute to a more balanced and harmonious life.

3. Different perspectives and personal growth: Being in a relationship with someone who has different experiences, perspectives, and strengths can broaden one’s horizons and contribute to personal growth. A partner can challenge and inspire personal development, fostering intellectual and emotional growth.

4. Complementary strengths and support: In a healthy and equal partnership, both individuals can bring unique strengths and qualities to the relationship. For example, one partner may excel in certain areas while the other excels in different areas. This complementary dynamic can enhance the overall well-being and success of both individuals.

5. Love and intimacy: Romantic relationships can provide love, intimacy, and a deep emotional connection that many individuals desire. While it is not necessary for everyone to be in a romantic relationship to experience love and intimacy, it is a valid and fulfilling aspect of many people’s lives.

In conclusion, while it is true that women can lead fulfilling lives without a romantic partner, it is important to recognize that human relationships and needs are complex. The phrase “a woman needs a man like a fish needs a bicycle” oversimplifies the diverse range of experiences and desires individuals may have. It is essential to respect and acknowledge the choices and preferences of individuals, whether they choose to be in a relationship or not.

And that was the enlightened reaction. It somehow escaped our attention that the term romantic partner could be gender neutral. On the occasions that a pregnancy is required, the market can be relied upon to deliver the traditional method of delivery or the prepackaged variety.

The fish, whether on the bicycle or merely swimming, does not see the water. Neither do men see the patriarchy.

The Spartans followed established precedent. Archidames II

If all you measure is billable hours, all you’ll get is billable hours. Me

The most beautiful things in the world are those from which all excess weight has been eliminated. Henry Ford

Beauty is that reasoned harmony of all the parts within a body, so that nothing may be added, taken away, or altered, but for the worse. Leon Battista Alberti

Clients feel the variance, not the mean. Jack Welch.

By asking why five times and answering it each time, we can get to the real cause of the problem, which is often hidden behind more obvious symptoms. Taiichi Ohno

We have no choice but to manage the service institutions for performance. And they can be managed for performance. Peter Drucker

Of course it had happened, not once in a hundred years but many times. Roger Lowenstein

Nature shaves with Ockham’s razor. Me.

Black swans migrate. Something Nassim Nicholas Taleb would say).

Come il pittore dev’esser vago di udire, nel fare dell’opera, il giudizio di ognuno. (every painter should listen to criticism). Leonardo da Vinci

μεταϐάλλον αναπαύεται (change is rest) Heraclites

We believe that implicit communication--to communicate through mutual understanding, using a minimum of key, well-understood phrases or even anticipating each other’s thoughts--is a faster, more efficient way to communicate than through the use of detailed, explicit instructions. A.M. Gray

*Do billable hours correlate to any other variable in a transaction?

Spoiler alert

This essay was drafted by Google NotebookLM from a prompt provided by me, and I lightly edited the result, after reviewing the sources cited. The supposed speech by Cassander is entirely my own composition. If that ruins it for you, feel free to rage against the slop, especially if you can find anything better on offer to serve as a contrast.

How Federalists and Anti-Federalists would have reacted to a prophecy of modern America.

This is a speech prepared for the opening of the Constitutional Convention by Rufus Cassander.

My esteemed colleagues, I behold a vision of this fledgling nation that perhaps only Providence can fully foresee. I see not just a union of thirteen states along the eastern seaboard, but a continental nation spanning from eastern shore to western – some distant day. I foresee this land and its people ascending to stand among the foremost powers of the world, with gleaming cities boasting millions of souls. When even a teeming city of one hundred thousand persons shall seem but a country village in comparison.

I foresee the populace of this great land numbering some 350 million free citizens. Thus, we must lay a foundation of laws and governance sturdy enough to bear the weight of so a vast republic. For no longer chiefly shall we be scattered villages where all pray together, work the land together, and know each other by bonds of family or familiarity. But even strangers shall be bound as countrymen in common cause by the ideals enshrined in this Constitution.

So let us attend to this convention with all diligence that we may wield such wisdom and a strong sense unity and national identity as shall steer this fledgling state through whatever tempests may assail our experiment in liberty. Not for this generation alone, but those uncounted generations of Americans to come do we etch the tablets of freedom today. May our vision prove worthy of divine blessing for as long as the Union and this Constitution shall stand.

But Cassander never journeyed from the far reaches of New Hampshire to Philadelphia as he intended for he was felled by a bull and had his skull cracked open.

The Founders had varied views on how long they realistically expected the Constitution to remain in effect before requiring changes or replacement.

Thomas Jefferson

• Suggested that constitutions naturally expire every 19 or 20 years and that no generation should bind another. Thought the Constitution should be regularly revised.

Alexander Hamilton

• Expected the Constitution to last a considerable duration as part of his arguments for ratification, but was vague on exact timespans. Emphasized importance of flexibility to allow change.

James Madison

• Also avoided setting definitive timelines, arguing that regular constitutional conventions could alter the document if needed. Prioritized constant improvement.

State Constitutions

• Most state constitutions created after 1776 had clauses calling for revision votes every 5, 7, 15 or 20 years. The Articles of Confederation had been in effect less than 10 years.

Lifespans Mentioned

• Some founders referenced historic democracies like Rome or Britain as benchmarks, suggesting constitutions lasting hundreds of years were possible if flexible enough. (Although if the British constitution is dated from the Glorious Revolution it was less than a century old.)

While no specific expiration date was set, most founders expected periodic revision of the Constitution at least every generation or two. They balanced accounting for population growth and new conditions with longevity. But they created an amendable framework, kicking the can down the road in hopes that future adjustments would be made either through amendment or replacement.

What if?

If Rufus Cassander’s imagined speech projecting a nation of 350 million citizens, gleaming cities, and a continental expanse had been delivered at the Constitutional Convention—and validated by the immense influence of Benjamin Franklin—it would have fundamentally altered the framing of the ratification debates. Contemporaries were already deeply concerned about the relationship between population density, geography, and republican government. (Franklin had envisioned the population growing past 100 million in the 19th century and Madison had the population potential twice that in the early 20th.)

Here is how the Federalists and Anti-Federalists might have adapted their arguments for and against the Constitution in light of this astonishing forecast.

The Anti-Federalist Opposition: A Prophecy of Tyranny and Corruption

For the Anti-Federalists Cassander’s vision would have been the ultimate nightmare, validating their deepest fears about the proposed Constitution. They would have seized upon the speech as proof that the new government was destined to become a despotic empire.

The Impossibility of a Vast Republic

Vastness and republicanism cannot co-exist. Anti-Federalist writers like “Brutus” (likely Melancton Smith) already argued that a free republic could not exist over a territory as large as even the original thirteen states, citing the historical failures of the Greek and Roman republics as their borders expanded. Faced with the prospect of a continental nation of 350 million, Brutus would have amplified his argument that true representation would be utterly impossible. He warned that in so large a republic, citizens “will have very little acquaintance with those who may be chosen to represent them,” resulting in a legislature composed of men whose “talents and regard for the public good, they are total strangers to”. Cassander’s admission that the nation would be composed of “strangers” would be lay open the argument that a standing army in peacetime—which the Anti-Federalists vehemently opposed as an evil danger to liberty—would be the only way to enforce laws upon a disconnected, sprawling populace.

The Corruption of “Gleaming Cities”

Cassander’s prediction of “gleaming cities boasting millions of souls” would have appalled thinkers who believed that civic virtue was rooted in agricultural life. Thomas Jefferson, in particular, believed that as long as there were vacant lands, American governments would remain virtuous. However, he explicitly warned against a crowded, urban future: “When they get piled upon one another in large cities, as in Europe, they will become corrupt as in Europe.”

Weaponizing Franklin’s Theories

Because Franklin’s influence hypothetically lends Cassander’s speech credibility, Anti-Federalists would have turned Franklin’s own demographic theories against the plan. Franklin observed that when all land is occupied, those without land must labor for others, leading to low wages and poverty. He also noted that densely packed European cities did not even supply themselves with inhabitants by natural generation because the high cost of living and luxury deterred marriage. Anti-Federalists would argue that adopting a Constitution designed for such a densely packed, European-style dystopian future would destroy the “general happy Mediocrity” of the American middle class.

The Federalist Defense: Institutional Resilience for a “New Order”

Federalists would have adapted their arguments to present the Constitution not just as a fix for the failing Articles of Confederation, but as the only framework capable of surviving Cassander’s staggering future.

Protecting Against the “Excess of Democracy”

Confronted with 350 million “strangers,” Federalists would have leaned heavily into the need to filter popular passions through representative institutions. At the Convention, Elbridge Gerry argued that “the evils we experience flow from the excess of democracy,” warning that the people are often “the dupes of pretended patriots” and “misled into the most baneful measures.” Roger Sherman similarly argued that the people “should have as little to do as may be about the government” because they “lack information and are constantly liable to be misled.”

Fears of a Propertyless Mob

James Madison harbored deep anxieties about overpopulation and later actually forecast a US population of 192 million by 1929. Madison feared this density would bring about the rise of a propertyless class that, in a democracy, might use the government to take wealth from the rich. Taking Cassander’s speech to heart, Madison and his allies would have argued even more fiercely for strong institutional bulwarks to protect property rights against the future passions of teeming majorities. Hamilton was another voice against “the mob.”

The Extended Republic

In Federalist Paper №10 Madison argued that an extended republic helps control factions and self-interest. Under Cassander’s projection, Federalists would likely double down on this, asserting that a populace of 350 million would create such a vast multiplicity of factions that a tyrannical majority would be mathematically impossible.

The Regional Battle: The Atlantic vs. The West

Cassander’s vision of a nation stretching “from eastern shore to western” would have severely escalated regional paranoia during the Convention’s representation debates.

Eastern delegates already feared that “Western people” would eventually gain power and “ruin the Atlantic interests.” Gouverneur Morris of Pennsylvania argued that the “Busy haunts of men not the remote wilderness, was the proper school of political Talents,” and complained that “Back members are always most averse to the best measures.” Cassander’s certainty of western expansion would have prompted eastern Framers to fight harder for constitutional mechanisms to artificially cap western political power, while southern and western delegates would have argued that mandated censuses were absolutely vital to ensure their growing populations received proportional representation.

Conclusion

If Cassander’s speech had been treated as an inevitable truth, the ratification debate would have shifted from a dispute over managing thirteen disparate states to a profound philosophical battle over the destiny of human civilization. Anti-Federalists would have painted the Constitution as a suicide pact that would accelerate the loss of agrarian virtue, creating an unmanageable, corrupt, and tyrannical colossus. Federalists would have countered that only a strong, centralized, and carefully checked national government could possibly prevent 350 million strangers from descending into anarchy or mob rule.

Humanoid coda

Unmediated me (well, except I was only the art director for the illustration)

There is also no evidence that the Congress foresaw the emergence of a “living Constitution,” doctrine under which the Supreme Court would supply suppleness to a governing document that had been over taken by the times. Hamilton approvingly describe the Court as “the least dangerous branch,” and it is likely that no daylight would be visible between the Founders on whether the Court should be saddled with the burden of setting right fundamental questions that the political branches could not. An emphatic no, actually, Hell, no! is the likely consensus position had the question been posed.

And that’s the big fail, right there.

The Constitution was the political solution to the problems of the newly independent Colonies that had tried a loose affiliation and found it wanting. As an immediate fix, it was neither designed nor intended as a permanent solution. While no expiration date label was affixed, neither was a “do not discard” admonition attached. The Civil War was fought not to vindicate a permanence to be found within the words of the Constitution but to enforce the political will of a (somewhat slim) majority to preserve the Union.

The inherent contradiction of the Original Intent doctrine in attempting to attach the original understanding of a narrow political class in the late 18th Century is that it fails to probe the silent assumption that the Framers understood their arrangement necessarily to be as lasting as it proved to be. If they had heard a real-life Rufus Cassander and taken his vision seriously, the only reasonable conclusion would be profound humility.

You have mistaken us for men of unworldly wisdom. We are not able to design a governing regime that will endure centuries. It has never been attempted, let alone accomplished. We will be satisfied that our work, if adopted, will suffice to carry us safely through our infancy as a nation and preserve us from a return of foreign rule. The problems to come, and of course they will come, we must consign to our successors who alone will have the appreciation of the relevant facts and circumstances of which we are necessarily ignorant.

At least one, I suspect, would mutter under his breath “Screw posterity, what have the bastards ever done for me?” as my brother likes to say.

In short, you alternate between the states of Fear, Uncertainty and Doubt.

(Yes, it’s AI, so shoot me.)

A partner or senior associate tasks you with preparing an informal client memo regarding its ability to modify mortgage loans and tosses you a 400-page document, saying “it’s all here in the PSA.” I need it in the morning. (It’s 6pm.)

You retreat to your desk, wondering why she just didn’t send you the document by email rather than digging it out of a pile on her desk. Then she she sees it’s heavily marked.

You being reading.

Your first thought is what even is this? It looks like a contract, but it also looks like a trust. And who are all these parties?

More questions follow. What is a REMIC and why are there so many of them? What’s the difference again between a mortgage and a loan? Why are there two trustees? It’s governed by Delaware law, so are there any cases? I see some provisions regarding loan modification, but what else am I missing?

In other words, you lack context, a framework by to recognize what in the document is relevant to the question and what is beside the point. Fear paralyzes you by the thought that it can’t just be Sections 3.01 and 3.07.

Your new, closest competitor suffers from no such fear, being somewhat literal minded, as well as indefatigable. It just glances at the document, instantly analyzes everything that might be semantically (not just keywords) related to loan modifications and spits out a crisp 5-page memo in less than 3 minutes.

AI client memo

… any labor that accepts the conditions of competition with slave labor accepts the conditions of slave labor, and is essentially slave labor. Norbert Weiner

I have an imaginary conversation to provide that context that would have taken a half-hour with an engaged listener. (But I had that context by having originally drafted the form 15 years earlier and having re-read it for dozens of deals.)

It’s long been out of syndication, I suspect, but the TV game show Let’s Make a Deal with host Monty Hall crossed my mind today.

The contestant is asked to choose one of three doors. Monty knows which door has the fabulous prize and which two doors have the gag goats. (AFAIK, no contestant ever drew the goat and went home with it.) The contestant picks one, and Monty opens the door with the goat and asks “do you want to stay with your door or switch for the unopened door?”

The knee jerk reaction is “two doors, right?” so 50-50.

Wrong.

Actually not wrong if that is the only information you have.

But you do have other information. You know that a goat was behind at least one of the two doors because two doors, one prize, two goats. Each door to start had a 1 in 3 chance of not having the goat. Between them they had a 2 in 3 chance of having the Corvette, leaving you with a 1 in 3 chance. You now know which of the doors you didn’t choose had the zero chance, so the other must have the 2 in 3 chance.

Nothing about your chances changed when you learned which of the two doors had a goat. What you now know is that if one of the two doors had the prize which one it was. There was a 50-50 chance on each door between those two doors. In the three-way, it was originally 0.33 .… , 0.33,… and now its 0.66.…, 0.

Your chances remain the same, 0.33 .… Therefore, you can double your chances by switching.

But wait, there’s more. Even knowing this you have to combat loss aversion. Wouldn’t you feel terrible having given up the winning door? Isn’t it worth giving up the additional 0.333 … chance to avoid that anguish?

The answer, according to behavioral economics is that a surprisingly large proportion of people are so risk averse that they answer “yes.” The hole that this lack of rationality puts in traditional economics has yet to be filled with something so tidy.

Here's the source material, which would have made my job in 1973 much, much easier.

Watch the 7-minute explainer

Prepared by my friendly neighborhood Google NotebookLM. I’m pretty knowledgable about the subject, and this is a fair take on the source. It doesn’t cover everything and if I looked carefully, I probably could find some holes. But if I did something similar by hand it would take a tad longer than 15 minutes.

I’m reminded that we don’t yet appreciate the full disruptiveness of LLM tech.

This isn’t for you, kids. It’s not a reflection on your generation. Or of it. You’ll have your own versions and all that goes with it. For the rest of us, the spoilers that follow won’t matter—you know the theme and variations already from lives that touch your own lives directly or through others close to you.

This is the take of Jane and Jake, their three kids, his second wife and her brat kid, an architect, a family come together to celebrate a graduation, pointedly ignoring ten years of being split asunder, a reunion of sorts, an opening up of a home and virtual open heart surgery.

Jane is Meryl Streep, the Schrödinger’s cat actor who is both her character and herself at the same time, who lives in a modest, little Santa Barbara bungalow now the home of Harry and Meagan, where her youngest is off to college. Jane is effortlessly flowing back and forth from her patisserie-cafe and her plans to expand the house.

Jake is the almost sort-of-lovable Alec Baldwin. He is a big swinging dick lawyer with a hot young wife and a spoiled brat from her previous plus-one in tow and a fierce demand for a matched set. She is draining Jake of his vital bodily essence alternately sending him to the fertility clinic to give it up for in vitro and keeping him performing in synch with her closely monitored ovulation schedule.

Steve Martin, the wild and crazy banjo player as Adam the architect, complements the principals as a conscientious listener to self-help tapes piloting himself through healing from his own divorce while doing deep reads of the 47 emails of Jane’s that his partner Peter has passed to him in hopes that Adam can weave together her wishlist for the remodel.

Writer/director Nancy Meyers put this trio in motion with an ensemble including the second wife, three children, a son-in-law to be and wine-drinking friends around the dining table and coffee table.

The set-up is an anniversary party in an impossible scene combining Tuscany with a semi-tropical seaside for friends of Jane and Jake celebrating their 30th. We think at first that the two couples are old close friends and have similarly long-lasting bonds of marriage. Jane completes Jake’s sentences. Jake thanks her. We wonder where this is going.

Right up to the entrance of Agnes, the Queen of Sheba. Jake’s plus one stalks in with the pace and power of a Panthera leisurely stalking her helpless prey. We don’t know if Agnes was the proximal Other Woman or her successor in interest. But she has Jake now, possibly without the pre-nup. While it might sting, Jane is apparently ok with that, possibly through naked will power.

Next, son Luke is to graduate in NYC, bringing everyone into direct contact at the pomp and circumstances cutting to mom alone sitting at the hotel bar waiting to be seated for dinner, abandoned by the children who fly off to help Luke to celebrate his farewell night to the college life.

An illness or sulk renders Agnes hors d’combat. Jake wanders into the restaurant and seats himself next to Jane, where they proceed to drain the bar’s cellar. And end up in bed. It’s not a one-night stand, and continues when everyone returns to Montecito. Jane is in tension with an opposite polar attraction to Adam, the anti-jerk. Jane and Adam revisit getting stoned after a long, long time away and get down and silly, having too much fun to grow any more intimate than delight in the other’s company and easy company.

When Jake arrives unannounced next to find them enjoying dinner that Jane is cooking for Adam in what Jake still thinks of as his home, he huffs and he puffs and ruffles his feathers to their full bristled extent. Adam politely recedes. But Jake doesn’t perfect his claim. When the family trips to the covert affair the question naturally comes to the fore—if you two are coming back together what the fuck was the last ten years of our pain all about?

We don’t know if they learn that they were the inevitable collateral damage of two lives that merged as well as they could but could not stay melded in one or even in permanently locked struggle for victory. The kids are always the first to know, but also the last to understand.

Jane gently pushes Jake away with a note of gratitude for the resolution of their unfinished business together—yes, they might have made it, but no they aren’t going to now. Like all the best stories this is only an ending, not the ending. We do not know, nor do the characters, how the pending story arcs converge or diverge. We don’t know if their best date ever experience will continue to bring Jane and Adam closer. Or if Jake leverages new self-knowledge into a second family. Or how the children grow up. The prospective son-in-law Harley is already a grown-up and we may have hope that everyone joins him and Adam at that table.

If your heartstrings do not vibrate from this flic, they are still too stiff to relax into life’s long rhythms and you must be patient.

In the post-mortem of the 2008 financial crisis, a popular narrative took hold in securities class actions: that mortgage originators had “abandoned” their underwriting guidelines. The claim was simple: if a loan didn’t fit the numeric box, the disclosure was a lie.

I had skin in that game. I, personally, was being sued for more in $1 billion dollars in damages in seven class action lawsuits arising out of my role in the issuance of several Washington Mutual Mortgage residential mortgage backed securities deals in 2006 and 2007. (Not that I was good for it, but how many times do you have to be killed to be pronounced dead?)

As I discovered when I sat down with the loan-level tapes and a SQL database, “exceptions” aren’t an admission of guilt—they are a feature of the system.

The Materiality of the 1%

To test the “abandonment” theory, I analyzed three mortgage pools issued at the height of the market in 2007, totaling thousands of loans. By cross-referencing every loan against the applicable pricing guidelines for FICO, LTV, and DTI, I sought to find the “ground truth”.

The results were startling for anyone pursuing a “material omission” claim:

• Pool A: Only 0.35% of the principal balance consisted of loans with exceptions that lacked apparent compensating factors (like a better LTV to offset a lower FICO).

• Pool B: That figure was only 0.67%.

• Pool C: A mere 0.10%.

When the “total mix of information” shows that 99% of a pool is either compliant or balanced by compensating factors, the legal argument for materiality begins to evaporate.

The “Matthew” Problem: How Institutions Actually Buy Risk

The most critical part of this analysis isn’t just the math—it’s the psychology of the Reasonable Investor.

In a draft paper, I introduce “Matthew,” a proxy for the institutional analyst. Matthew doesn’t care about a “scalar” FICO score in isolation; he cares about Credit Support. And, as a reasonable investor he does not “buy the rating.” He actually reads the offering documents, including the detailed loan-level data.

If a pool has 12% subordination—meaning 12% of the pool has to fail before Matthew’s bonds lose a dime—is an undisclosed 0.33% exception rate “material?” Mathematically, no under the hoary rule of thumb that 5% or 10% is usually the threshold. The credit support provided a margin of safety far exceeding the risk of those few outliers.

The Lesson for the Boardroom

My draft paper serves as a blueprint for Quantitative Litigation Defense. It demonstrates that in complex securities matters, the “truth” isn’t found in the narrative of the descriptive prospectus—it’s found in the data tapes.

For modern boards overseeing algorithmic and financial risk, the lesson is clear: Oversight isn’t just about reading the policy; it’s about auditing the output.

I mean not only X, Facebook, TikTok and Instagram but the entire world of always on, always active information sources that arrive unbidden and need only that we turn our attention to them. This includes news, state propaganda and, yes, Substack.

Philosophical tools

We are not unarmed. There has been some serious thinking about human communication from some profound thinkers.

Harry Frankfurt

The late philosopher is the author of On Bullshit, a short treatise on the nature of statements made without concern for their truthfulness. Bullshit need not be false. Frankfurt stresses that the bullshitter’s distinctive motive isn’t necessarily to lie about facts, but “to misrepresent what he is up to” (his description of the bullshitter’s real deception). On social media people and institutions constantly present themselves as truth-seekers, authentic sharers, or “just asking questions,” while in reality they’re optimizing for status, virality, or commerce. Social media rewards engagement, not accuracy, so Frankfurt would likely have seen it as an institutionalized incentive to produce exactly this kind of content: speech crafted to impress, perform, or signal, with no serious regard for whether it’s true.

Alexis DeToqueville

He analyzed how a free, cheap, decentralized press shaped public opinion, associations, and majority power in a democracy—almost a 19th‑century analogue of today’s social media ecosystem. Periodicals have “their own instincts and passions,” and while readers mainly seek facts they are swayed when facts are distorted, and countless small papers prevent any one outlet from fully controlling opinion yet create constant agitation.

He connected media freedom tightly to popular sovereignty: you can’t have universal suffrage with heavy censorship for long.

He warns that abrupt adoption of press freedom can be destabilizing, and that majority opinion, once formed and amplified, is extremely hard to resist—anticipating concerns about online mobs and “tyranny of the majority.”

Thorstein Veblen

Veblen was an economist who wrote about conspicuous consumption by the leisure class. By “leisure” he meant human activities divorced from production. Conspicuous consumption signals status as its primary function and, like bullshit, is predatory in nature. In the social media context this manifests in predatory competition for follower counts, blue checks, and engagement metrics as markers of pecuniary status. As Ludwig Wittgenstein might have remarked “for there to be winners, there must, perforce, be losers.” Clout‑chasing is predatory self‑assertion. Like sports and games, the parasocial fandoms, fantasy sports leagues and the like, are structured make-believe that have only a colorable purpose. Like the study of dead languages and ritual, platform-specific jargon, meme literacy and opaque in-group code constitute cultural capital.

Walter Lippmann

Lippmann was a political scientist of the early 20th century concerned with public opinion. He saw politics as a pseudo‑environment. To that pseudo‑environment our behavior is a response. Social media feeds are exactly such pseudo‑environments: algorithmically constructed worlds we react to as if they were reality. To the extent that what we do is not based on direct knowledge we rely on feeds, memes, clips, and headlines that compress social complexity into short messages with a small vocabulary. Fiction is taken for truth because it is needed as a substitute for knowledge. Thus, social media is an engine of information compression. Conspiracy threads, partisan narratives, and identity-flattering stories flourish on social media for this reason. People have comparatively meager time for public affairs; they “live in grooves… meet few people not of their own sort, read little.” Social networks create the illusion of breadth while still reinforcing narrow “grooves” (filter bubbles, echo chambers).

Norbert Weiner

Weiner was a physicist who coined the term “cybernetics.” He viewed human behavior as a complex control-and-communication process. In Cybernetics, he noted that our power comes from “the longest chains of effectively operated neuronic chains.” This complexity is our greatest strength, but it makes us fragile.

Wiener warned that we operate “very close to the edge of an overload.” When the feedback signals become too dense—as they do in the infinite scroll—the system doesn’t just slow down; it gives way in a “serious and catastrophic way.” We aren’t designed for the bit-rate of the modern timeline; our behavioral complexity relies on tight feedback loops that are easily shattered by digital noise.

Claude Shannon’s three entropies

Shannon, the architect of Information Theory, wouldn’t see “content” in social media. He would see entropy

a measure of uncertainty over possible messages: it is highest when choices are as equally likely as possible and lowest when one outcome is almost certain, i.e. “no freedom of choice—no information” as in his discussion of entropy and certainty in communication systems (on the entropy formula and its interpretation) and his remark that uncertainty is maximized when probabilities are as equal as circumstances permit (his comment on many choices and maximum (H)).

He’d likely split the “entropy of social media” into three distinct layers:

High Raw Entropy: The sheer volume of the “firehose” is massive. By the numbers, the potential information (uncertainty of the next post) is huge.

High Redundancy: Paradoxically, while the raw data is high, the meaning is low. Shannon noted that natural languages are highly redundant. Social media—driven by memes and stereotyped phrases—amplifies this. In Wiener’s terms, these processes “lose information” and “gain entropy,” turning a global stream into noisy re-transmission.

Engineered Certainty: The algorithm’s job is to lower entropy for the individual. It makes your feed predictable to keep you engaged. A “perfect” feed has low entropy for you (little surprise), while remaining part of a high-entropy global chaos. (Entropy tends to increase in closed systems, but not uniformly when there are local systems, which can remain in disequilbrium indefinitely.

Herbert Simon’s bounded rationality

Why are we so easily steered? Herbert Simon’s Bounded Rationality explains that we don’t “optimize” our information intake; we satisfice. We stop when we find something “good enough” because few important choices have a unique solution and search costs make for needing a stopping algorithm.

On social platforms, we let the environment’s structure do our thinking. We rely on shortcuts—likes, trending labels, and tribal cues—because our computational capacity is limited.

John Kenneth Galbraith’s conventional wisdom

Galbraith was an economist who portrayed the power of “what everyone thinks” in enforcing conformity and squashing heterodoxy. Algorithms don’t reward truth; they reward “acceptability.” Dissent that threatens audience comfort is algorithmically disadvantaged, ensuring that “facile flows of thought” dominate the public square.

Daniel Kahneman’s two systems of thinking

The late scholar is the only psychologist to win the Nobel Prize in Economics. He models human cognition as consisting of “fast” and “slow” thinking. System 1 thinking is fast, automatic, and driven by the “pictures in our heads” rather than the careful evidence in the realm of System 2, slow and deliberate. Propositions that would be rejected in slow thinking are accepted in fast systems due to cognitive biases that operate subconsciously.

We rely heavily on shortcuts: likes, trending labels, followers, aesthetics, tribal cues.

We let the environment’s structure do much of our “thinking”: algorithmic feeds, notifications, and defaults shape what we see and hence what we choose.

Our preferences and beliefs get pulled toward what is easiest to process and most salient rather than what is most accurate or valuable.

Robert Sapolsky on the role of neurochemistry

Robert Sapolsky’s research implies that social media isn’t about “pleasure”; it’s about the uncertain anticipation of reward. In what he calls the “Magic of Maybe,” dopamine spikes highest when a reward is unpredictable (the 50% reinforcement schedule).

Every pull of the feed is a “teaching signal” telling your brain: Pay attention; this is important. This creates a persistent, low-grade stressor. Chronic micro-doses of social comparison and potential ostracism keep glucocorticoids elevated, which subtly weakens the prefrontal cortex (PFC). Over time, your “frontal reserve” is depleted, making you more impulsive and less able to calm the emotional triggers of the amygdala. In his book Determined he emphasizes that dopamine isn’t a simple “pleasure” chemical but a teaching and salience signal: unexpected good or bad outcomes trigger dopamine changes that say, in effect, “pay attention; this is important, learn from it.” He describes unexpected reward as prompting “What can I learn… to make it more likely again?” and dopamine as mediating the message “Pay attention; this is important.”

The irregular, unpredictable rewards (likes, comments, new followers, surprising posts) are classic “unexpected reward” events that constitute strong dopamine teaching signals to repeat the checking and scrolling behavior.

Negative social feedback or ostracism online would also be tagged as highly important, especially for teenagers, whose frontal cortex is not yet good at down‑regulating social pain and who are particularly sensitive to social exclusion. He notes that adolescents’ PFC isn’t up to dismissing social ostracism as meaningless.

Over time, those dopamine-driven learning signals wire in habits where checking feeds feels urgent and hard to resist, because your brain has learned that “something important might happen” with every refresh.

So using Sapolsky’s framing, the role of dopamine in social media is: not making it “fun” per se, but continually training and retraining you—via unpredictable social rewards and punishments—to treat it as a high‑priority source of important

Constant micro-doses of social comparison, potential ostracism, and unresolved “threats” (unanswered messages, ambiguous comments, likes that may or may not come) would count as ongoing psychosocial stressors. He already emphasizes how vulnerable teenagers are because “the PFC isn’t up to the task yet of dismissing social ostracism as meaningless” “Who are these impaired individuals? Teenagers…”.

He notes that a single bad morning is enough that “by noon, glucocorticoids will have changed gene expression in the PFC, making it less excitable and less able to couple to the amygdala and calm it down.” Translate that to social media: a steady trickle of mild threat/uncertainty could keep glucocorticoids modestly elevated, repeatedly nudging the PFC toward worse impulse control and weaker top‑down regulation.

He also stresses that the PFC is “the most expensive part of the brain” when you’re trying to do “the right, harder thing” PFC as metabolically expensive, and that when it “doesn’t have enough energy on board, it doesn’t work well” Chronic stress plus constant self‑control (resisting checking, curating your image, etc.) is exactly the kind of situation that depletes that “frontal reserve,” like his example where resisting M&Ms for fifteen minutes makes you worse at an inhibition task afterward frontal reserve depletion.

So in Sapolsky’s terms, social media would be:

A persistent, low‑grade stressor centered on social status and exclusion.

A driver of repeated glucocorticoid pulses that subtly weaken PFC–amygdala regulation over time.

This is a context where a still‑developing or already‑stressed PFC has to work overtime to inhibit impulses, making people—especially adolescents—more impulsive, reactive, and vulnerable to mood problems.

Nassim Taleb’s fragility

Taleb is an applied economist, rocket scientist trader.

Social media over‑amplifies outliers (viral scandals, market panics, political shocks). He’d say we systematically underestimate how likely these fat‑tail events, surface and speed up by networks, dominating outcomes. Modern information streams drown us in noise, making us fragile and overreactive. Social media is almost a textbook example: constant tiny updates with little actionable information, but big psychological impact. Humans handle acute stressors relatively well, but chronic, low‑grade stress is damaging. Social media creates continuous micro‑stressors: comparison, outrage, status anxiety, notification drip. That turns people and institutions more fragile: always poised to react to the latest controversy.

He distrusts disembodied, consequence‑free opinions. Social media lets people broadcast advice, judgments, and outrage with no skin in the game, which he’d see as a structural driver of bad decision‑making. Taleb advocates simple heuristics to protect yourself from complex risks (e.g., via negativa, removing harmful exposures instead of optimizing). Applied to social media, that suggests:

Reduce or eliminate platforms that mostly add noise and stress.

Prefer small, robust circles (private groups, direct interactions) over large, fragile reputational arenas.

Bounded shocks (rare exposure, curated feeds) can be manageable. Treat your feed like a portfolio: avoid frequent, shallow “news” and outrage (noise) and favor rare, high‑quality sources with skin in the game. His heuristic might be: judge ideas by the quality of their strongest critic—on social media, this almost never happens, which signals low epistemic quality.

Platforms and many influencers have upside without symmetric downside; they profit from engagement even if it harms you. That’s a classic fragile system design. Most real‑time news/feeds are noise; more data can mean worse decisions, the feeds degrade judgment. That amplifies bad, cost‑free signals over grounded ones. He advocates improving life in general by removing harmful exposures. Applied to social media: delete apps, avoid timelines, reduce frequency, block addictive or enraging sources. He prefers local, time‑tested knowledge over abstract, global chatter. Social media overweights global noise and underweights local, robust feedback.

The link above will take you to YouTube where you can subscribe. Nothing else will happen.

I need 1,000 subscribers to ascend to the next rung of the great chain of being of Content Creation, and even if you have no interest in the Julia programming language, your clicks will help.

Thank you.

Thus was born a determination to become a programmer. It took a while.

The first foray was in 1965, when I wrangled an elective in FORTRAN, punch cards and overnight batch processing and all. But it turned out that there was a problem, which was that I didn’t understand the concept of algorithm. All the different ways to make change for a dollar? Uh, well, two fifty cent pieces, four quarters, 10 dimes, 25 nickels and 100 pennies was easy enough. Beyond that? Beyond me.

What I was trying to do was something like this.

C

C PURE FORTRAN 77 PROGRAM

C Enumerate all distinct ways to make change for $1 (100 cents)

C using U.S. coins: 50, 25, 10, 5, 1

C

C Total combinations: 292

C

PROGRAM CHANGE1DOLLAR

INTEGER NCOINS, TARGET

PARAMETER (NCOINS = 5, TARGET = 100)

INTEGER COINS(NCOINS)

INTEGER COUNTS(NCOINS)

INTEGER TOTAL

DATA COINS /50, 25, 10, 5, 1/

TOTAL = 0

CALL INIT(COUNTS, NCOINS)

CALL BACKTRACK(1, TARGET, COINS, COUNTS, TOTAL)

PRINT *, ‘Total ways:’, TOTAL

STOP

END

C------------------------------------------------------------

SUBROUTINE INIT(A, N)

INTEGER A(N), N, I

DO 10 I = 1, N

A(I) = 0

10 CONTINUE

RETURN

END

C------------------------------------------------------------

SUBROUTINE BACKTRACK(IDX, REM, COINS, COUNTS, TOTAL)

INTEGER IDX, REM, TOTAL

INTEGER COINS(*), COUNTS(*)

INTEGER K, MAXK

IF (IDX .GT. 5) THEN

IF (REM .EQ. 0) THEN

TOTAL = TOTAL + 1

CALL PRINTSOL(TOTAL, COINS, COUNTS)

END IF

RETURN

END IF

MAXK = REM / COINS(IDX)

DO 20 K = 0, MAXK

COUNTS(IDX) = K

CALL BACKTRACK(IDX + 1, REM - K*COINS(IDX),

& COINS, COUNTS, TOTAL)

20 CONTINUE

COUNTS(IDX) = 0

RETURN

END

C------------------------------------------------------------

SUBROUTINE PRINTSOL(NUM, COINS, COUNTS)

INTEGER NUM

INTEGER COINS(*), COUNTS(*)

INTEGER I

LOGICAL FIRST

WRITE (*,100) NUM

FIRST = .TRUE.

DO 30 I = 1, 5

IF (COUNTS(I) .GT. 0) THEN

IF (.NOT. FIRST) WRITE (*,200)

WRITE (*,300) COUNTS(I), COINS(I)

FIRST = .FALSE.

END IF

30 CONTINUE

WRITE (*,*)

RETURN

100 FORMAT(I3, ‘: ‘, $)

200 FORMAT(’ + ‘, $)

300 FORMAT(I0, ‘x’, I0, ‘c’, $)

END

So, it actually was something to impress the girls.

I dropped the course.

That didn’t prevent me from running a few job decks for programs written by others in college and grad school.

Then dropped this little marvel the HP25 handheld, somewhat programmable, calculator.

By Armin von Werner - photo of owned device, Public Domain, https://commons.wikimedia.org/w/index.php?curid=1133833

It allowed storing a sequence of keystrokes to perform chain calculations. It was swiftly followed by an HP65, which could go 100 lines and store code on little magnetic strips.

When the first Apple II came out, I was briefly intrigued but took a pass.

In 1980, I got my first computer.

It was bigger than it looks in the photo, weighed 800 pounds and with peripherals and software put a big dent in $100,000. It was because I thought it would be a good idea to bring a small municipal government on an island 2,000 miles north of Hawaii into the age of automated data processing. It took a staff of four a week twice a month to put out the payroll. That didn’t leave much time for balancing the checkbook, let alone keeping a general ledger. In short order, the payroll was the work of a half day of a single clerk twice a month.

It was way outside my job description and my only qualifications were that it was Alaska, I was on the ground, and claimed I could do it. It was fun!

Next came the Osborne 1, the first luggable computer

This little charmer predated MS-DOS, which ushered in, with the IBM PC, the current age of desktop computing and office software. It was only $6,000 in today’s money. It came with BASIC (Beginner’s All-Purpose Symbolic Instruction Code), which became my first language.

This was followed, of course, by the IBM

This allowed a second language, Pascal as implemented by Borland with its Turbo Pascal version. It was necessary to upgrade RAM, however. 16 kilobytes was just not enough.

Hard to believe, but it was up to the user to buy added RAM and insert it in the empty slots on the left.

Then, true love, the DEC 350, a desktop version of the legendary PDP-11 microcomputer, which was the surrogate mother of UNIX.

Of course, the first thing was to replace the operating system with Venix, adapted from Unix System 7. Now, there was C, bash, awk and a host of other little languages. That carried me through another several years before going back to the PC world just before Windows.

I was doing other things besides coding, so there it stood until the bright idea of doing an internet startup with my own host, and so a new computer, the Silicon Graphics Indy workstation.

This came with a more fully loaded OS, Irix, a System V UNIX derivative with a gorgeous shell (graphical user interface) that my brother Rand, the Mac maven, was duly impressed by.

That ended when my little start-up crashed and burned and it was back to Windows NT.

Until this sneaky little gateway drug.

It was once described as a “USB drive in hotpants.”

It was quickly followed by a laptop.

It famously “just worked” when it came time to connecting to the home router, something that was always an ordeal with Windows.

And, if you have a laptop, of course you need a desktop.

And if you have one desktop, you need another for work.

(My brother, who now has this, refers to it as the “cheesegrater.”)

Together with a suitable 27-inch display.

This was not only unsanctioned by my employer, but positively prohibited. However, it is always better to ask forgiveness than permission. Long story, but it was in connection with needing to spool up on R, Python and MySQL programming languages.

Post employment, there was another laptop.

Which was ok, but then I wanted something more server like.

And, of course, since it was headless, a laptop to connect to it natively.

But then, there was this missing Mac vibe. So

And it needed a native server, which led to the current monster with the M3 Ultra/96GB unified memory (I had expiring frequent flyer miles).

It’s bigger than it looks, about 8 inches (20cm) on a side.

And. Don’t get me started on the peripherals and all their cables. Well, here are a few.

I somehow doubt that I have yet reached peak hardware.

Today, in Julia, my current language, the code is easier.

function enumerate_change(amount::Int, coins::Vector{Int})

results = Vector{Dict{Int,Int}}()

function backtrack(remaining, idx, current)

if idx > length(coins)

if remaining == 0

push!(results, copy(current))

end

return

end

coin = coins[idx]

max_count = div(remaining, coin)

for k in 0:max_count

current[coin] = k

backtrack(remaining - k * coin, idx + 1, current)

end

delete!(current, coin)

end

backtrack(amount, 1, Dict{Int,Int}())

return results

end

ways = enumerate_change(TARGET, COINS)

println(”Number of ways: “, length(ways))

println()

for (i, w) in enumerate(ways)

parts = [”$(w[c])×$(c)¢” for c in COINS if w[c] > 0]

println(lpad(i, 3), “: “, join(parts, “ + “))

end

If you’re not a coder, this will still make your eyes roll to the back of your head but it’s easily explainable. As I set out to do in my new venture, coming soon.

A PRODUCTION OF

A very brief history of the Internet

• It was invented to facilitate continuity of government following a nuclear strike

• For many years it was an invite-only club

• Everything was free, and there was no advertiing

• But, on the other hand, there wasn’t much

• Then Sir Tim invented the world wide web

• dot.com was invented for commercial/industrial use

• (Fun fact: in the earliest days of Yahoo, their database was a deck of 3x5” index cards)

• (Another: I had a website before most Fortune 100 companies)

• The dilemma

  • The people who best understood the Internet thought it should be free

  • The people who might be interested in paying for it didn’t know much about it

  • So began the War of the Eyeballs

  • And advertising

• And, too late, the concept of advertising as the original sin of the Internet reached the conventional wisdom

So what?

It’s on its way here, to Substack.

We’ve already been promoting subscriptions to our various newsletters. I haven’t because, well, I don’t have paid subscriptions. Although I like to think that much of what I write is worth reading, it’s all over the place and I didn’t want to commit to a subject and have weekly deadlines to produce something worth paying for. In my age, I’ve come to realize that somehow there would have to be big bucks in it to give up a big chunk of leisure. Besides, I’m the Internet equivalent of the 100% effective vaccine. Since starting writing blogs in 1996, I’ve never gone viral.

And now?

I’m about to publish a book on using the Julia programming language to create neat maps. It’s aimed at people who have geographically based data to communicate but aren’t programmers. I spent the last six months on it, it’s coming out next month, I’m proud of it, and I want people to buy it. So, I’m going to advertise here and in notes.

But I don’t want to piss people off

So hear me out.

What’s not to like

• It’s advertising, after all

• Some of the advertising will be AI-assisted

• It’s a narrow audience

Mitigation measures

• It’s not going out on blast

• If the purpose is primarily marketing, it will be labelled as such up front

• If I’m writing about a topic with reference to the book, that’s fair game

• For podcast material, it won’t be live, to cut down on the spurious notifications

• For related notes, I plan a “black box warning”

The linked podcast is not for everyone. You should have an interest in maps, software to produce them and a tolerance for two things: First, it is in aid of selling my book. Second, the content was prepared with the assistance of AI

Range of AI

• AI helped in coding the programs descripted in the book, either by taking a first cut or by debugging or both. Since I’m a relative newcomer to the Julia language, my theory is that my mistakes are similar to the mistakes my readers can expect to make.

• Very little AI-generated visuals will be used.

• The limited use of music will be AI-generated soundtracks.

• The biggie is that some podcast content will be written and spoken by AI. (The first will probably be a 20-minute 2-host podcast based on Google Notebook-LM’s read of my book. Unless I’m more than usually blind to my own bias, the quality is mindblowingly better than 99% of similar content on YouTube, although that is a low bar. Although some of what they “say” is quibbleable, I vet and stand behind it and will take fair criticism of it as criticism of me, and I promise to hear it with an open mind.

So

Here’s the ask.

• Democrats will retake the House, but not the Senate

• Trump will try to shakedown people with Epstein exposure by threatening to pardon, not prosecute, them

• At least one administration political appointee will be held in contempt for non-compliance with an immigration-related court injunction

• Trump will put his face on the $100 bill

• Trump will propose making cyrpto legal tender, causing a spike in it prices, which his cronies will have cashed in on. Then he says “just kidding” and prices crash. Cronies will have advance notice of that, too, and will make out like bandits

Europe

• Poland will move to a wartime footing as Putin continues low-level hostilities

• France will be forced to cut back on defense spending due to severe voter disapproval, and the government will suffer another no confidence vote

• Germany will struggle to increase defense spending in light its constitutional debt limit and strong right -wing opposition

Asia

• Philippines quietly acquiesces to Chinese maritime claims

• Japan will seek to strengthen its ties to India including joint military exercises

Middle East

• More ICC indictments in connection with the Gaza War

Africa

• China expands regional influence

Latin America

• Trump has Navy bomb a Venezuelan site. and loses one aircraft to s surface-to-air missile and there’s a TACO

• Argentina gets another bailout

North America

• ✔️ United States continues under new political management amid institutional decline and cynicism. (That was a gimme.)

• ✔️ Media and political class search for reassurance that outcomes will not be as severe as feared. (I’m looking at you, MSM.)

• 1/✔️ Former presidents launch bipartisan reconciliation efforts; punditry anticipates internal Republican resistance. W took a bye. Signs of Republican resistance were greatly exaggerated.

• ✔️ Claims that large-scale government spending cuts are impossible collide with aggressive administrative disruption tactics. DOGE failed, but cuts as by-products of other actions bore fruit.

• ✔️ Food system strain: agricultural and food-service labor shortages lead to restaurant closures and higher prices.

• ❌ Consumer spending contracts as transportation and commuting costs surge.

• ✔️ Return-to-office mandates persist amid hollowed-out urban real estate markets.

• ✔️ Governance deteriorates unevenly as competing power centers vie for presidential attention.

• ✔️ Privatization advances under the banner of efficiency; contractors profit while service quality erodes.

Europe

• 1/✔️ German rearmament accelerates; France moves toward a Sixth Republic.

• ✔️ Rising tensions involving Poland, Belarus, and Russia; at least one new Balkan conflict likely.

• 1/✔️ Fragmentation pressures across the EU and Eurozone, including potential withdrawals.

• ✔️ Germany informally assumes leadership of a Western Europe–only security structure.

• ❌Russia enters a new phase of internal restructuring as central authority weakens and ethnic nationalism resurges.

East Asia

• ✔️ China’s rise reflects a reassertion of long-standing historical patterns.

• ✔️ China adopts a patient approach toward Taiwan, expecting economic and political shifts to weaken resistance.

• ❌Opportunistic expansion into Siberia and Central Asia driven by energy and mineral needs.

• 1/❌Intensified repression of Uighur identity; Tibet sees temporary reprieve.

Middle East

• ✔️ Large-scale Palestinian displacement looms.

• ✔️ Iran’s regional influence contracts with setbacks in Syria and Yemen.

• ✔️ Houthis move toward greater autonomy amid reduced Iranian support.

• 1/✔️ Kurds align tactically with Türkiye in Syria in exchange for limited local autonomy.

• ❌ Iran’s aging revolutionary leadership gives way to demographic and internal pressures; Revolutionary Guard may seek dominant political power.

Africa

• ✔️ Sahel instability deepens; drought exacerbates Ethiopian internal tensions.

• ✔️ Migration pressures increase across the continent.

• ✔️ Expansion of jihadist control across the Sahel, Lake Chad region, Mozambique, and other areas.

• ❌ Egypt leverages regional turmoil to improve Nile water security.

• ✔️ Continued coups and internal conflicts in West and Central Africa.

• 1/✔️ Southern Africa remains relatively stable, though South African influence declines.

• ✔️ U.S. retreats from countering Chinese economic influence; China consolidates control of strategic minerals.

• 1/✔️ Russian state influence fades, though private mercenary activity persists.

South Asia

• ✔️ China distracted; Pakistan destabilized; India consolidates under Modi.

• 1/✔️ Strengthening of the Line of Control and military posture in Kashmir.

• ✔️ India engages selectively in Central Asian geopolitics.

• ❌ Unexpected technology transfer partnership with Ukraine accelerates India’s autonomous weapons and arms industry growth.

Latin America

• ❌ Mexico’s president brokers cartel power-sharing arrangements to stabilize violence and reduce U.S.-directed fentanyl flows.

• ❌ Mexico cooperates selectively on deportations and migrant transit in exchange for funding.

• ✔️ Ongoing crises across Central America, Venezuela, and Argentina persist amid high poverty.

• 1/✔️ Regional resilience rooted in strong national identities.

Oceania

• ❌ High risk of a major natural disaster: Antarctic ice collapse, massive volcanic eruption, earthquake/tsunami, or emerging tropical disease.

• ❌ Climate-driven uncertainty dominates regional outlook.

Global Closing Note

• ✔️ An era characterized by cascading disruptions, systemic stress, and adaptive but uneven responses across regions.

As a class representative straw man, here are some factoids about me.

• Born 1947, the second year of the cohort

• Grew up mainly in the San Fernando Valley

• First generation Mexican-American cross with Boston Irish, although the former was strenuously denied due to the accident of a Basque surname (which turns out to have been assigned, not passed down)

• First in entire lineage to college—Occidental College, where Obama later attended.

• Married junior year. One child during second year of grad school. Paid OBGYN/hospital expenses from a part time job and supported wife and child on stipends and outside jobs. After 3 years of grad school, public sector job at entry salary comparable to what NY large law firm associates were making. Bought a modest new house in a new subdivision. Got a private sector job with enough of a bump in salary to keep me even with rising inflation, at the cost of being away from home pretty much constantly.

• Divorced, remarried. More on the road full time. Quit to take a public sector job in remotest Alaska. Turned into a different job, one for which I was pretty much unqualified for but did well and had fun. Returned to urban life for a couple of years and then went to law school and became a 42-yo first-year associate at a large, San Francisco law firm.

• Failed to make partner, so I went back to the public sector at the cost of boredom, which I treated with an internet start-up side hustle. Rejected a job offer at Netscape in favor of going in-house as a number two to a former client in Chicago. Got ourselves acquired by Washington Mutual and relocated to Seattle, where I eventually ascended to reach the rarified heights of SVP/AGC with a total compensation in today’s dollars in the high six figures. On the other hand, I was an alcoholic with poorly treated bipolar 2 disorder.

• WaMu fell and I was out on the street at age 62, permanently unemployable with half my retirement savings gone poof! (Believe me, I was not too proud to apply for jobs that didn’t even require a high school diploma.) Took early social security.

• Subsequent events, some of which were in my control, ate the second half of retirement savings, and today I have an income consisting of social security and a very modest pension. My wife requires full-time care. We live in a modest apartment. I drive a beat up 2010 Audi.

So, there’s your target. Undoubtedly you know of Boomers both more and less blameworthy that I am for the situation in which you today find yourselves. Here’s a face to go with it.

Have at me! Don’t hold back. The purpose of this exercise is to surface attitudes toward Boomers that may, or may not, be generally true.

There’s been some chatter about the increase in electricity prices. So, I thought to take a dive.

As usual, the place to begin is with “what are they?” The map above divides the country into five equal parts based on the weighted average (by consumption across all the providers in each state) retail costs in each state. Costs are highest in Alaska and Hawaii (no surprise) and on the mainland California and New England plus New York suffer the most.

A little histogram illustrates the spread.

Summary Stats:

Length: 51

Missing Count: 0

Mean: 16.871696

Std. Deviation: 6.082727

Minimum: 11.419867

1st Quartile: 12.938729

Median: 14.883434

3rd Quartile: 17.417381

Maximum: 42.863416

The chart above shows the 1,519 utilities reported divided into five equal groups, or quantiles. Groups 2-5 have little variation, as shown by the thickness of their bars. Group 1 has some low outliers, while Group 5 has many.

Row │ quants mean_rate

─────┼───────────────────

1 │ 1 10.0762

2 │ 2 12.3007

3 │ 3 13.5059

4 │ 4 15.26

5 │ 5 21.7483

Group 1 includes utilities in 37 states; each of the others include 42 states. Except for New Hampshire, Vermont and Rhode Island, which are all in Group 5, the highest cost group, all states have one or more utilities in each of the groups.

Utility rates vary by type of ownership

Row │ owned weighted_mean_rate

─────┼───────────────────────────────────────────

1 │ Investor Owned 16.9223

2 │ Cooperative 13.465

3 │ Municipal 13.5259

4 │ Behind the Meter 20.3053

5 │ State 21.8936

6 │ Political Subdivision 12.9235

7 │ Federal 12.402

8 │ Retail Power Marketer 16.6637

Of the 52 utilities with the outlier high costs in Group 5, only five have revenues above $1 million per year. These are Massachusetts Electric Company, Connecticut Light and Power Co., Pacific Gas & Electric Co., Consolidated Edison Co-NY Co. and Southern California Edison Co.

Next up: Compared to what?