Oh, To Be Rich Poor Alive Dead In America

After some inordinately long periods of sickness over the holidays, followed immediately by the need to evacuate my home for ten days with my family during the Eaton Fire, I have not felt much like writing. My head has been a blender filled with scattered thoughts and ideas mixed with familial obligations, life necessities, work, frequent brain farts, and bananas. Which is to say, the usual state of mind for most attention deficient adults in 2025, but instead mixed on the highest setting until my brain has been more liquid than solid for some weeks.

As I feel my brain has begun to congeal into something resembling a neuroplastic thinking device again, I have a few other fun pieces in mid-flight and ready to take off to your inboxes. But today I am loathe to tell you that after the events of the past 72 hours, I feel compelled to write about AI yet again, even though at this point it pains me to even see the letter “A” and the letter “I” adjacent in print or even in the general vicinity of each other. I can’t even pnt my house or expln why I gned weight.

The infernal machine of the market continues onward and today the market took what is called - and I believe this is a term of art - “a fatty dump.”

Why it took this truly epic and stinking fatty dump has been predictable even if the timing and precise catalyst was unexpected as market catalysts tend to be.

The TL;DR version is this: A Chinese AI start-up released an AI product (DeepSeek) that is just as good if not superior to the AI products of the American techno-behemoths (OpenAI, Anthropic, Meta, Google, etc.), and they released it on a vastly more cost effective basis, with far fewer and less powerful computing resources, and in less time than their American counterparts. Basically American capital brought a knife to a gunfight but the knife was a trillion dollars and the gun brought by the Chinese was a few million dollars that also evaded U.S. sanctions ostensibly employed to prevent this very scenario.

From here the math is simple.

For the past two years and counting, the barons of Silicon Valley technocapital and the VCs who back them have proclaimed like heralds before the coming of a great war that AI has come to solve all of our problems. Your knee hurts? Put some AI on it! Your ice-maker is broken? Stick some AI inside it! Your girlfriend dumped you? Why not put yourself inside some AI - hubba hubba! But of course in order for AI to solve all of our problems, AI must be American because America is America. And because the AI must be American they need more money and more energy to make it within America. And to make it in America they need more money and energy than has been required for essentially any industrialized capital project in history, except maybe those six days when our most famous founder-entrepreneur, God, made the Earth before he took a well deserved nap, Shabbat shalom.

The belief was that in order to make the Captain American AI supercop investment banker doctor lawyer filmmaker who will elevate us all to the next higher plane of existence, the technology needed to scale, which is to say they needed to spend all the money to buy as many computer chips as possible, mostly from Nvidia which exploded their valuation to over $3 trillion, and they needed all that energy to power dem chips so you can have a deep relationship and virtual sex with the preggers Sonic the Hedgehog of your dreams. Essentially it has been argued that “bigger is better because bigger must be better” (AMIRITE LADIES?!) even though bigger has thus far resulted in plateaus of performance growth trajectories for nearly all AI products, which is academically known as the “Scaling Problem.”

However it’s been known for quite some time that the Scaling Problem may not actually be a problem that needs to be solved, or if does need to be solved, whether or not it can be solved at all within the bounds of our existing understanding of computation and physics. Which is to say bigger may not be better just because it’s bigger. (AMIRITE FELLAS?!). Many academic papers, longstanding AI researchers with decades of experience, and general critics of the technology have been arguing for at least a year now that the Scaling Problem cannot be solved just by pouring more money into endless data centers cooled by a given municipalities entire supply of drinking water and powered by nuclear reactors that won’t be built for at least a decade.

So this past weekend, despite spending hundreds of billions of dollars, nigh trillions of dollars, our great heralds of the Valley realized that they were potentially not heralds at all, and were possibly more akin Humpty Dumpty upon the wall, teetering with ambition and ignoring all evidence to the contrary until they had a great fall (in market capitalization). And no I am not only referring to the egg shaped man himself Marc Andreessen.

Someone else did it faster, cheaper, and more effectively than they could and they didn’t see it coming, which made Nvidia lose more money in valuation in a single day than any company in history, although admittedly it is also the largest company by valuation than any other company in history. Because if it can be done faster, cheaper and more effectively, then why would anyone be spending billions of dollars to buy the largest quantities of most expensive hardware possible?

And therein the fatty dump was taken as every investor looked at every executive at every publicly traded company whose line item budgets for “BUY MOAR CHIPZ TO FEED BOTZ” in 2025, and went “whaaaaaaaaa” at the same time and realized that perhaps the current market conditions are based not only on bad ideas but potentially on bad ideas which are also unprofitable, which is a fate far worse for an idea than just being bad. Plenty of bad ideas are profitable, just watch literally anything on TikTok.

This whole experience would be less offensive, even with the losses to my own personal stock portfolio and your 401(k), if AI was not being promoted by those who stand to benefit the most, Cui Buono, as a cure-all to everything ever on the face of the earth past present and future.

It should be said once again that I personally have uses for Claude / GPT / Perplexity and use all three products each week for some task or another. I do actually find them quite useful for ongoing education and research in particular.

But have these AI technologies meaningfully changed my life in the way their biggest proponents and creators have said they would? No, they have not, at least not in a way that is substantially different than any other new technology or tool which has made something previously more difficult into something less difficult. Although I do think think Apple Intelligence may have actually made my life marginally worse. Fuck you, Siri.

The lately more obvious potential (literal) control of our society by the - Bernie Sanders voice - “billionaire class” of technology scions aside, it is important for us all to be even more skeptical of those who equivocate the most about the virtues of our new automated technology era, and who have been pushing as hard as possible on a market-moving scaling model of AI to solve the Scaling Problem, unarguably for no other reason than “nUmBeRs BiG!” when alternative methods were known to be available and which have now been shown to be yield better results.

The head-spinning discourse which has been articulated to support the scaling of AI is meant to leave you in awe.

“AI is going to kill us all!”

“But it’s going to make your life into fully automated luxury communism!”

“AI is going to take your job!”

“But it’s going to make your job easier!“

“AI is going to create superbug biological weapons!”

“But it’s going to cure cancer and be your hospice nurse as you die from the superbug biological weapons instead of cancer!”

Some, all, or none of these things may be true in the near or far off future, but regardless, they are all directly contradictory claims which are being made constantly and consistently inconsistently by the people who want you to fear their technology which will simultaneously make you poor rich alive dead where you apparently exist in some approximation of a Schrödinger Cat-like state, which consequently means you fear them because they make and own the technology that does this, and therefore you will stand in awe of this thing that will supposedly kill you but that you absolutely must buy as much of as you can before it does.

In sum, if not in short, the entire market is (was) potentially being propped up because of a very small group of companies, their leaders, their overlapping boards of directors, and mutual venture capitalists wanted to use a lot of money to impress you with that money and subsequently increase valuations in order to build an “incredibly easy” way for you to order for groceries (see below) when they could have just as easily not done that, but for a dick measuring contest of datacenter construction that we will all inevitably bear the cost of.