Regular readers will know that for a couple years I’ve been saying that Chinese open source AI would win the AI “war” because it’s cheaper and non proprietary (prices can’t just be raised suddenly, or capacities taken away.)

Over the last few months there’s been a lot of screams coming from regular AI users. OpenAI and Anthropic moved to token based billing, which is to say “you pay based on how much you use.” They still weren’t charging full rate, but they were charging a LOT more and users were not happy. One company spent 500 million by mistake: they forgot to put limits on how much their employees could spend.

Oops.

Nor are ordinary users exempt:

I Went From $3,000/Month on Claude to $5/Week on DeepSeek

And honestly? 80% of my work is identical.

For the past two months, I was burning $3-5K monthly on Claude Code. Every idea from design to development to testing – full end-to-end automation, even simulating users to test my products and provide feedback. Extremely token-intensive.

But Claude’s caching sucked, making it insanely expensive. Then I discovered DeepSeek V4.

The numbers: • Claude: $5 input, $25 output per million tokens •

DeepSeek: $0.28 input, <$1 output (with their current discount) • DeepSeek cached: $0.0002 – literally less than a penny The caching optimization is game-changing.

Once DeepSeek has seen content, it basically stops charging tokens. My result: $5/week vs $1,000/week for the same workload.

So now Microsoft has created their own minor Deepseek fork, and will run it on their servers to power Copilot. You can still use a version run by US labs, but if you can’t afford, or justify that, you can use the Deepseek version.

Driving the news: Microsoft says companies using Copilot Cowork will pay based on how much compute they use.

• The company tells Axios it is exploring a fine-tuned version of DeepSeek V4, or another open-source model, as a lower-cost alternative to the Anthropic and OpenAI models now powering Copilot Cowork.
• Microsoft says it expects to make a lower-cost model available in the coming weeks and will confirm its choice then.

Worse than this, there’s beginning to be serious pushback on whether AI is all that useful. Uber’s COO opened the door back in March:

In perhaps the most high-profile example of this growing concern yet, Uber COO Andrew Macdonald acknowledged during a recent podcast appearance that gains in productivity simply weren’t being reflected in the oodles of cash the company has been shelling out on AI.

“That link is not there yet, right?” he told Rapid Response host Bob Safian. “I think maybe implicitly there is more that is getting shipped, but it’s very hard to draw a line between one of those stats and, ‘Okay, now we’re actually producing 25 percent more useful consumer features.'”

“If you’re not actually able to draw a direct line to how much useful features and functionality you’re shipping to your users that trade becomes harder to justify because it’s not free,” he complained. “AI is not free.”

As far as I can tell there’s little evidence that US priced AI is more cost-effective than the employees who were laid off because it was so great. I rather suspect that in most cases, it’s less cost-effective.

But more importantly we have the “it’s better to be wrong with the crowd” effect moving against AI. In almost all positions, including executive ones, if you’re wrong in the same way that everyone else is wrong, it’s no big deal. If you’re wrong against the crowd (say not getting into AI when the rest of your industry is) and it turns out that AI is the next big thing, well, you’re fired.

So much of the AI mania was driven by this and a relentless hype cycle. Now that important people are beginning to push back on it, it’s no longer required to be all-in on AI. And that’s bad for Anthropic and Claude.

AI is not the next coming. It is not going to make it to general AI (not this generation of large language models anyway) and while it does have some utility the US frontier models cost far more to operate than any conceivable return most of their customers will receive. It isn’t the “get rid of three-quarters of your employees” super app corporate leaders were promised.

And to the extent it is useful, well Chinese open source models are more cost effective. As good? Generally no. But they keep catching up, and paying 70 to 97% less makes up for being somewhat behind.

So to the extent that AI is a real industry, odds are high China’s going to win the race. Since the models that will win will be built off open source models that’s not a crisis for anyone, it’s a good thing, far better than a proprietary future.

BUT it does mean that US AI expenditures are probably going to turn out to be the biggest misallocation of resources in centuries: bigger than the housing bubble and bigger than the dot-com bubble (which at least did have a world changing technology behind it.) Not quite the Dutch tulip bubble, but at least the Dutch got lots of pretty flowers of that, instead of massive ugly data centers.

Business is driven by stupid people engaged in group think, especially in the West, far more than most people will admit. Everything Silicon Valley does these days is someone trying to create a monopoly or oligopoly so they can be insanely profitable, while China actually competes on price, and that’s why China keeps eating the West’s lunch.

I’d cry, except that an open source AI world is a far better one than a proprietary one, and every tear some Silicon Valley tech bro cries over a lost opportunity to make a monopolistic buck an angel gets their wings.

What I write here is for the benefit of everyone, but alas, I live in capitalism and I, and the site, take money to keep running. If you value the writing here and can, please subscribe or donate.

I recently stumbled across a story about the Supreme Court (not exactly bleeding heart liberals) refusing to let Alabama use nitrogen suffocation:

An Alabama man facing the death penalty by nitrogen gas was spared Thursday as the U.S. Supreme Court refused to set aside a lower-court ruling that found the method is unconstitutionally cruel, issuing a brief order that came well after the hour originally planned to initiate Jeffery Lee’s execution…

…During the previous Alabama nitrogen executions, the inmates shook, pulled at the restraints and exhibited labored breathing. During the state’s last execution by nitrogen gas, 30 minutes elapsed between Anthony Boyd exhibiting signs of being impacted by the gas and state officials closing the curtain to the viewing room to signal the execution was complete.

The idea is that you breathe, but the gas you’re breathing is nitrogen, so eventually you die.

Of course this is going to suck, anyone who’s ever suffocated or had serious breathing issues knows that one of the worst feelings in the world is not being able to breathe.

There’s been a lot of this sort of thing going on: the company that used to sell drugs for execution stopped doing so, and various US states have been looking for alternatives. The prisoner in this case wants a firing squad, figuring it’s quicker and less painful.

Meanwhile up here in Canada we have legal assisted suicide. It’s a controversial program, because it seems like the government or various relatives are a little too eager about it. (After all, dead people don’t take up hospital beds and dead relatives don’t cause problems.) I think assisted suicide is often a good thing, but easily abused, however we’ll leave a moral deep dive for another article.

The thing is there’s never any criticism that it is a painful death. I looked into it. They give the patient:

1. An anxiolytic and sedative drug: Midazoloam.
2. They give them a drug to put them into a coma-like state with a rapid acting drug: Profofol. Then,
3. They give them a drug which causes paralysis, including of the lungs. Rocuronium. The patient dies of suffocation, same as with helium (or Hemlock, which Socrates was executed with.)

But the patient doesn’t suffer, because they’re deeply unconscious.

This protocol works, it’s well known, so why not use it?

Because Alabama and other US states want the prisoner to suffer. Moaning about expense is ridiculous, however expensive it is it’s cheaper than keeping a prisoner on death row, same as it’s cheaper than keeping a patient in hospital.

Executing prisoners without causing undue suffering is a solved problem. Alabama and other states like it just want the prisoner to suffer, so they keep searching for a method that will be painful and courts will allow.

What I write here is for the benefit of everyone, but alas, I live in capitalism and I, and the site, take money to keep running. If you value the writing here and can, please subscribe or donate.

There always seem to be a handful of revolutionary technologies perpetually out of reach: fusion energy, quantum computers, and full self-driving cars are always in this list, and it seems like there’s also some battery technology which will finally let us fully decouple from fossil fuels in there as well. Although lithium batteries have allowed some ground-based electric transportation, the energy density is still not enough to enable full electrification, especially for things like aircraft. Solid state batteries may be on the verge of changing some of this, though, and a team has recently put them to work in a test aircraft to help make some headway with this novel battery chemistry.

The main contributing factor of these batteries’ improved energy densities is the ability to use a solid lithium anode, which has much higher energy density than the graphite-based anodes in modern liquid electrolyte batteries. Solid state batteries also have improved safety, since the solid electrolyte is generally not flammable and the battery itself is less prone to thermal runaway. The tests in this aircraft, a modified motorized glider, bear this out as well. With a standard lithium ion pack the team was able to harness 250 Wh/kg and with their new solid state battery they managed 410 Wh/kg, which let them fly the craft up to 24,000 feet (7,315 m) with the help of some wing-mounted solar panels.

Of course, a motorized glider is a long way away from battery-powered commercial flights, but tests like this are an important step on the way to de-carbonizing one of the more impactful industries on the planet, as well as hopefully making it less expensive to operate aircraft in the way EVs are generally much cheaper to operate than their internal combustion equivalents. But the limiting factor to adopting solid state batteries isn’t going to be implementation but rather the discovery of a cost effective way to manufacture them at scale. It’s the same reason we haven’t seen mass adoption of things like algae-based biodiesel or economic carbon capture yet.

Trains and big cities are more efficient on a typical cost basis than cars. The operating costs of trains are such that even the unsubsidized costs of big-city metros and bus networks are a fraction of those of cars. For example, New York City Transit, despite its high operating costs, manages to serve a linked trip for around $6, which works out to an annual cost per user of around $3,500, and if it had the cost structure of London or Berlin this would be $2,000 or a bit less than that; American cars average $7,000/car per year in private spending.

And yet, cars have one singular advantage: they commoditize location. Public transit ideally works in large cities at specific locations, based on historic contingencies like national capitals, religious significance, or river crossings and harbors that may no longer be relevant with modern technology. It’s decommoditized, in that there is only one New York, one Philadelphia, one Chicago, etc., and the cost of moving is high. Public transit itself doesn’t lend itself to competition, because it requires extensive scale to ensure connectivity and high frequency. This is why public provision is almost universal, and the exceptions either involve a high degree of public coordination such as the Verkehrsverbünde in the German-speaking world even if elements are contracted out or are Japanese cities with such large systems that competition between a JR and a private operator still leaves each competitor with much scale; even generally privatization-happy states like Singapore keep the systems broadly public in planning.

What this means is that cities and public transit require a public sector that can keep up without the discipline of market competition. This means public-sector innovation, with competition taking place in the political sphere as in European cities or in the technocratic one as in Singapore. If this doesn’t happen, then the system suffers. If, for example, the New York MTA folds to a strike by the LIRR train drivers in which the union demands are so unreasonable that even the left-wing city mayor Zohran Mamdani doesn’t side with the union, and gives the drivers large increases in pay while still allowing them to collect double pay for driving both a diesel and an electric train, then there’s no easy way to move to a competitor.

Cars and suburbs instead commoditize location. If the city can’t provide adequate public services, people can just leave. It’s particularly easy if the municipality that falters in providing services is not a large city but a small suburb of one, as in the boroughitis of New Jersey. Cars facilitate that, in that they scale down better. There’s no way to squeeze anything the size of Midtown Manhattan or even the center of Paris into one auto-oriented place (Los Angeles has a weak central business district), but that’s fine, a region can take the hit on income and still function with worse scale; Dallas is not a poor region. There are real problems in this setup with higher transportation costs and with job centers with worse scale, but sometimes it’s worth it to take the hit if it means not having to deal with unaccountable government that one can’t leave. If there’s no mechanism to improve governance – say, if there is such democratic deficit at the local level that it’s not possible for voters to coerce the city into improving education or public transit or housing or any other devolved issue – then that usually equally affects city and suburbs, but one can move from one suburb to another at relatively low economic and social cost, and this has a disciplining effect to some extent.