Actually, he probably wants cash.

They make things; they're not investors. Often these kinds of "investments" from companies like NVIDIA are not cash injections into companies, they're in-kind. So if their chips and GPUs are priced at let's say $10,000, and they make a $1B "investment", they're really giving their investment 10,000 GPUs, not $1B in cash. But note, they're priced at $10,000; it costs NVIDIA likely $2,000 to make and the rest is overhead and marketing. So an "investment" in OpenAI comes back to them in the form of sales, that are non-cash sales but still accounting sales.

Now they go public, and let's say their investment nets $3B on the $1B. That's essentially like making $30,000 on a $10,000 product that cost them $2,000 to make; a real windfall.

To be fair, there is a long tradition that predates Trump in the DoD if using the work performing system for defense contracting.

Yep. Just what we need: an AI capable of making kill decisions for autonomous weapons that is deep down deeply racist, fascist, and used primarily for NSFW images.

America!

And that's not a great thing. I worked 6 years for one of the primes, and it's clear how they operate. Anthropic's problems would go away if the leadership starts acting like Lockheed; kowtow to the DoD officials by calling them DoW, and say "yes sir, whatever you want sir". Hire a bunch of retiring O4s to O6s who aren't going to get their stars and give them cushy mid-management jobs, and then stack on top margin after margin while delaying projects left and right.

It's funny because having worked at a Prime before and seeing this very thing happen, I also don't believe very heavily in the military industrial complex. The concept was that industry would push the military to war because sales were driven by weapons usage, but that never really materialized. Rather, it was a welfare state. The DoD is a terrible customer, buying things in fits and starts, changing requirements in the middle of a program, and squandering R&D budgets on pet projects and nonsense. Meanwhile the contracting officers are too lazy to go direct to a Tier-2 or Tier-3 supplier for an interesting idea, and instead farm it out to a Prime who just subcontracts to the Tier 2 and puts an overhead fee on it. Meanwhile the warrant officers for a given technology are reluctant to change anything without 10X the proof of capability and safety studies than would be normal, meaning half of our military's subsystems are so legacy compared to what's available commercially that our military is eminently hack-proof because no modern hacker knows how to hack an abacus and a hamster wheel in code written in ancient Egyptian that is the backbone of many sub-systems.

The Primes on the other hand have regular bills to pay and workforces to maintain and see this insane way of doing business that the Pentagon does, and adapts to it milking as much overhead as possible so they can level out their monthly payroll expenses without too much labor disruption. All the while Congress has no idea what to do to fix the issue, so they impose restrictions on government employees where they have to report even a lunch meeting over $25 or get investigated while we squander billions with bad bureaucrat managers in DoD.

The one thing the Trump Admin is doing somewhat right is targeting this exact issue; somewhat right in that it's an issue that needs solving so they got that right, but wrong in that I don't think they know how to fix it.

Sorry for the rant. I loved my time working at a Prime and I still cherish it, with some great colleagues that I still keep in touch with. But once I started seeing how it all worked, I was just stunned with the ridiculousness of it all. And now we see effectively mid-tier at best contracting bureaucrats trying to manage something as fast moving as AI with all the subtlety of the Titanic, and with likely similar outcomes. And what's sad is that the DoD should cave to Anthropic. Claude is good, and the military does lots of things that don't involve weapons; it has the world's most complex logistics chain, a huge healthcare system, major R&D programs, huge humanitarian programs, it led to the development of game theory, it has (or had until Hegseth) one of the world's best leadership training programs; all of those aspects of the military could benefit, and if they really want killer-AI weapons, as bad as that is, I'm sure Musk will sell it to them with Grok. It's painful to watch what could be amazing utilization of AI become a giant s-show.

the power side can't be solved. Power capacity growth is a linear growth model; it simply cannot meet the exponential growth that these companies are projecting. On-site generators have a 90 month lead time right now. The grids and current generation can't keep up. Nuclear power, which would be ideal, is a decade away at best to meet their needs. They'll hit a physical barrier in power generation and then the growth will stop and the bubble will burst.

Not to mention inflating stock market indices and laying the groundwork for a bubble burst followed by a recession.

I agree with the assertions in this article. Even if AI builds a vastly better system than SAP or Oracle, companies won't buy it, and it comes down to an old saying. It's slightly different in every industry, but it comes down to this:

No one ever got fired by buying from GE/ThermoFisher/Oracle.

For this particular case, given how critical business systems are, a wrong move can screw up the business royally. That means the executives and the CEO even, but all the way down to the buyers are on the hook of it goes wrong. If it does though, the outcomes are simple. If they bought from Oracle or SAP, and it screwed up, well Oracle and SAP know better so that's their fault. However if you bought from a no-name company, particularly one built by AI that is known to hallucinate, well that's your fault for not buying from Oracle or SAP who should know better.

Even if the product is better, it doesn't matter when people's reputations and careers could be impacted by it going wrong. It's far less risky to blame Oracle or SAP (and I've delt with both, both offerings are complete bloatware and a load of garbage) then it is to take the blame yourself, and no amount of excellent code can overcome that simple sociological fact.

It's a byproduct of what happens with all technologies, the Gartner Hype Cycle.

A new technology comes out. People immediately think it is the be-all, end-all solution to many things. People apply it to numerous things because they can.

Eventually, many of the things they claim the technology will do don't pan out. Then you head into the Trough of disillusionment. People begin to ask questions like "We can do this, but should we?" Many potential applications die off completely. The few that remain survive, and gradually grow into becoming a useful solution.

So yes, of course people used an LLM to generate passwords. Because they can. Doesn't mean they should, but we're not quite at the "should" stage of LLM applications yet.

And for the bubble talk, yes AI is a bubble. The reason it's a debate is because we talk about bubbles in binary terms, but when you look at it with the Gartner Hype Cycle, you'll see that it's not yet ready to fall, but applications like password generation and quite a few others show that it will. It won't collapse completely, because there are good applications of LLM AI; it will climb out of the trough and find valuable, useful tools. The debate about whether it's a bubble or not is irrelevant, what we should be discussing is when a contraction will happen (and it will), what it will look like when done.

One of the big differences between Amazon and WalMart, and fundamentally is a problem with the assertion in this story, is that Amazon isn't really bigger than WalMart, but Amazon also is a very different company. 90% of WalMart's revenue is from online sales and retail sales, and this dwarfs Amazon's retail business by a large margin.

However Amazon does two things that WalMart will never replicate, and in some ways gives them an enormous edge.

AWS was a byproduct of Amazon building their own logistics empire, and they rented out the excess server space. When this took off, it became it's own thing. There is no world where WalMart will go into cloud computing infrastructure; it is a retail giant. The fact that Amazon has a core business that is fundamentally different than WalMart's and is very successful and profitable in it's own right gives Amazon a diversification edge in it's retail business.

The other thing Amazon does is it rents out it's logistics empire. There are now many businesses that piggyback off of the massive distribution chain that Amazon has built. I know many small businesses that are successful and profitable that sell products entirely through ecommerce, and they use Amazon extensively. The business carries no inventory physically, it does no order fulfillment, and does not need to worry about getting shipping labels correct nor does it need to set up a global distribution chain; they pay Amazon to do that and as a small business they can buy a large business' entire logistics system. They use Amazon's store front for orders, Amazon's marketing to get their products up, but they are their own brand and P&L. 60% of sales on Amazon come from third party sellers, and 70% of those use Amazon's logistics services, which has globally over 1,300 fulfillment and distribution centers.

This is something that WalMart will not do; they provide shelf space for other manufacturers to sell their products, and they use their own logistics chain to move things around their stores and retail outlets and some delivery, but it's the old-school retail model that they control; Amazon's is fundamentally different in that they operate the logistics but the sellers control what they want and where. Amazon's model allowed for multiple small businesses to compete with larger companies who went through the retail model, and in return Amazon got enormous data on consumer trends and spending in both their core categories but also ones they're not in to. There were bad stories here; Amazon went into their own products and competed with their own partners when it was profitable, true, but there are still hundreds of thousands of companies that are successful today because of Amazon's system.

But those are why WalMart doesn't dominate here. WalMart still does dominate; it's retail business still dwarfs Amazon, and it serves a demographic who doesn't do a lot of online shopping and is quite large. But Amazon does many things that Walmart won't do which gives them a competitive edge.

With all due respect, why? Cancer is cells that grow out of control, but they're still just the cells of the animal. In your gut, your stomach acid destroys the genomic material; you can't catch cancer from eating it. If you could, then by extension you would also absorb the natural DNA of animals you eat; you'd be genetically part chicken, cow, pig and fish if it were possible to absorb the genetic material of what you eat. So what's the issue?

Cancer may suck for the creature that has it, but in a factory farm cancer is essentially free extra meat (if it's a muscle tissue tumor).

I disagree with your points. It is flesh of a living being. They're taking the muscle cells of living beings, using modern techniques to encourage the cells to replicate which cells do, and in the end they produce a constant stream of muscle tissue. The only difference is the fatty side of things; a grown organism will have fat mixed in depending on it's diet and how it's raised and the cut, whereas this will likely be pure muscle tissue with very little fatty tissue.

But by your definition, it's not protein masquerading as meat. It's literally the cells of cows grown in the same way that cells grow to form muscle fibers, that just happen to be in a tank instead of a cow, so how does that not meet your definition?

Actually I don't think it does threaten ranchers; rather it just adds another source.

The US produced around 25 billions pounds of beef per year.

Current lab-grown meat facilities can operate at a several kilogram scale. Optimistically people are suggesting we can get between 22,000-44,000 lbs per cultivated meat per year in commercial grade bioreactors, but no one has proved that yet and bioreactors are notoriously finnicky at scale. For reference, a bioreactor is basically a fermenter, like what's used to brew beer, except much more complex and different recipes. The open secret about brewing is that it's quite hard for a brewer to make a consistent product and controlling the fermentation process, so often it comes out tasting different than the primary product. If it's not toxic and tastes reasonably well, they slap a new label on it and call it a "seasonal". That would be much more difficult to do with lab grown meat though, people will expect a certain flavor quality. The point of this anecdote is that controlling a bioreactor fermentation process is quite challenging even today particularly at scale, so while some suggest we could get to 22,000-44,000 lbs per year per bioreactor, it's still an optimistic goal.

But let's be generous and assume 44,000 lbs/year/bioreactor. At 25,000,000,000 lbs per year of beef produced / 44,000 lbs per year per bioreactor (and these are large scale ones available today), you end up needing approximately 590,900 bioreactors to entirely replace the cattle ranching industry, in just the US, not globally. Or around 60,000 bioreactors to even make a 10% inroad on just US beef production. For comparison, the entire pharmaceutical market operates today around 5 to 8,000 bioreactors total; the number of bioreactors sold per year at that large scale is in the hundreds per year, two orders of magnitude less than waht you need just to make a small dent in just the US beef market. ALso the current production cost is around $17/lb and selling at around $50/lb; compare that to the average price of beef which is around $8 per pound (all cuts of all types across the nation).

So no, cattle ranchers are not worried.

In unrelated news, Meta stock is down 2% today. It should be down even more.

My child likes to use that term a lot too, and I keep telling them "Life is unfair. All sorts of people are going to be unfair to you. If you don't protect yourself, then you only have yourself to blame." It may be unfair, but it's not unlawful, so there's no reason to not do it and if they can do it so can many other companies.

It's interesting, because the Chinese have a habit of this kind of reverse engineering. They're very good at it, and you often hear the West (Note, I am American so definitely a Westerner) decry Chinese practices are unfair. From their perspective, what's unfair is how China was more than 50% of global GDP for nearly 1,000 years, and their economy literally defined major events; the Silk Road was the path to wealth, but the Middle Eastern nations were in the way and getting a cut, leading to Bartolomeu Dias being the first to round the Cape of Good Hope (so Europe could bypass Middle Eastern merchants to deal directly with China), and Columbus sailing West; it was always about reaching China's riches. Then in the 1800s Europeans usurped China's position primarily through growing opium in India and selling it to China, destabilizing and corrupting the government. As far as they're concerned, that was 150 years of unfairness to them, so turnabout is fair play.

History is obviously much more complex than that, but "unfair" is a subjective term and a crybaby term. More importantly, it shows the weakness of AI spending if you have to spend billions of dollars building something and then someone can copy you for millions.

The good news is that most of these deals have new lockup periods being . It's typical for lock up periods to last 180 days; these new ones are being staggered out for over 2 years or more. Even if if some of these companies are legitimately building a solid, investable business, with the sheer size of these IPOs the number of newly liquid shares could tank the stock price just from an oversupply of shares. So most of these big rounds they have to show decent performance for at least 2 years or they lose big. that goes for employees and founders too.

It's something, and I hope it makes a difference, although I don't know if it will. NVIDIA is still sky high in my opinion and they're already public, as is some other "AI" companies like Meta and Alphabet. But given the sheer amount of spend going on by these companies in the market and how they're affecting the overall economy, it'd be better for all of us to have some more insight into these companies' operations that is required for a public company, even if it ends up being a painful insight.