blueblimp retweeted
Flashback to June 14, 2024.
At the time, frontier AI models have zero fluid intelligence. They score around 5% on ARC 1. They can't reliably add two numbers. They make laughable logic errors every other prompt. Reasoning models don't exist. Coding agents don't exist. Claude Code doesn't exist. "Code gen" is a complete joke. Scaling up next-token prediction based on web text data is hitting a wall.
What if a time traveler told you that, by November 2026...
• AI models would have human-like fluid intelligence, thanks to a new paradigm: test-time induction (modern reasoning models).
• Not only would ARC 1 be saturated, but unreleased benchmarks 10x harder (ARC 2) and 100x harder (ARC 3) would be saturated too.
• Nearly all code would be written by AI, and models could one-shot complex codebases flawlessly.
• AI models would be solving multiple Millennium Prize problems.
• OpenAI's valuation would be 15x higher, and Anthropic's would be 85x higher.
And on top of all that...
• Nvidia would still hold ~80% of the AI compute market.
• Nvidia would have grown its revenue and earnings 4.5x.
• Nvidia would still be growing 75% a year.
• Nvidia would have even higher operating margins (67% vs. 65% back then).
Based on that information, straight from the future, would you have bought NVDA?
If you answered yes -- congrats. Today, 2.5 years later, you'd be down 5% on your investment (even more after inflation).
The speed of AI improving at research math overshot even the most bullish predictions that I recall. But it wasn't uncommon to suspect that math might be a relatively easy domain due to reasons like: verifiability and humans being bad at it.
I think about this Tyler Cowen quotation often. (It's not just the EU. There are many ways of thinking such that, if the world were ruled in that way, the world's growth rate would be negative.)
10
There are always multiple sides to every argument.
I just listened to a popular Chinese tech podcast where the host interviewed a guy who used to work as a security guard and has now made at least $140,000 (over 1 million RMB) so far this year by himself on AI-powered micro-dramas. It took him about five attempts, costing $1,000 or less each time, to get here. According to him, AI has unlocked the power of the OPC (the one-person company), and everyone he is in regular contact with also makes money in this industry. It seems like he is enjoying it quite a bit more than being a security guard.
When asked why there are so many headlines about folks not making money, he merely said: "Well, have they actually kept at it, or did they always expect to make money from the get-go?" According to him, it took a bit of time. He also said he focuses on quality stories, makes no more than one or two shows per month, and took a while before scaling up from a 30-minute show to a 120-minute show.
So, one door closes, another door opens, but maybe not for you, for someone else. I think the person who expects life in these disruptive times to continue as usual and not change at all is the really sad one.
By the way, I'm not sure that his way of making money is necessarily going to be around in a year or two either, but this whole microdrama thing itself is only about five years old. It drew in a lot of people who were never going to have had an opportunity to work on "filmmaking" otherwise. So these people, who weren't professionally trained actors, got their time in the sun, and now it's over.
We're just cycling through business models and products/services faster than ever before. I don't think you have to lament that much. There are trade-offs, and it's not like this is representing some grievous loss of long-lived cultural traditions
blueblimp retweeted
A year ago, most people considered me as crazy when a told that we will see superhuman AI mathematicians by this summer. But the current pace exceeds even my expectations.
blueblimp retweeted
Like most mathematicians, I believed the Riemann Hypothesis to be completely out of reach at present (even with the best AI tools), because I expected it to require substantial theory-building, probably in arithmetic geometry. The kind of new theory-building that fills entire books, and that AI isn't so good at yet.
Mathematicians have a notion of how deep a problem is. A sufficiently deep problem is generally intractable until enough of the theory is in place. If difficult problems are like valuable mineral reserves buried miles underground, you first need to build the infrastructure to reach them.
What Levent's proof seems to have done instead is something completely different. It takes an incredibly complicated tour through the infrastructure built for a hundred other mineral reserves, situated far from the destination, and somehow create a wormhole to the original destination.
This is not theory-building in the usual sense. It is something much weirder and more alien. The best human mathematicians know how to build theory, painstakingly, to reach a goal. But no human brain, on its own, is capable of creating a wormhole that requires a 25,000 page long tour through existing infrastructure with no obvious or realistic connection to the goal. But huge credit to Levent here: I think no current AI, on its own, is capable of this either.
I had never worked on or thought too much about the Riemann Hypothesis: it seemed very far removed from my own area of number theory.
But the astonishing thing about Levent's claimed proof of RH (which I think has a very good chance of being correct) is that it uses a huge number of precisely the things that are central to my own corner of number theory:
- Theta functions and other automorphic forms on unitary groups
- Gan–Gross–Prasad period formulas
- Relative trace formulas
Etc., etc.
A problem that I had always assumed would require new theories in arithmetic geometry may instead turn out to be tractable through an inhuman chain of existing (deep) results in higher-rank automorphic forms (and a lot else).
blueblimp retweeted
In lesser news, Tristan and Levent claim to have proved blow-up of UNFORCED Navier-Stokes
blueblimp retweeted
Well, looks like @__alpoge__ & @claudeai are claiming a proof of the Riemann Hypothesis
alpo.ge/remark.pdf
blueblimp retweeted
There has been some confusion on why 48% of $NVDA group revenue goes through their Singapore subsidiary.
It's because that subsidiary handles Nvidia's sales in Asia, not because Singapore itself is buying that much. The buyers are mostly US hyperscalers, Taiwan server makers, and neoclouds around the world.
Why would a sale to Microsoft or CoreWeave be an Asian sale? Because they often don't buy the chip direct from Nvidia, they order servers from makers in Taiwan like Foxconn and Quanta. Those makers buy the GPUs from Nvidia Singapore, build the server in Asia, then ship it to a data centre in the US or Europe. The sale is where the factory is, not where the server ends up.
That concentration makes sense because the supply chain is in Asia and US buyers centralise APAC purchasing in Singapore. Nvidia has disclosed that >99% of controlled Data Center compute revenue billed to Singapore in the first half of the year to Jan 2026 was for orders from US-based customers, and shipments to Singapore were <2% in the year to Jan 2025.
The timing of Nvidia's disclosure change has raised eyebrows, but the new basis is actually better because it follows who the buyer is rather than just the invoice address. In the year to Jan 2026, US-headquartered customers accounted for $149.6bn (69%), Taiwan $42.3bn, China $19.7bn and Other $4.3bn. China's $19.7bn was down from $25.0bn the year before, and $9.7bn of it came in the first quarter, including $4.6bn of H20 sales before that chip needed a licence in April 2025. Geographic disclosure by invoice address made Singapore look like a $23.7bn market in the year to Jan 2025, which it wasn't.
So…
-Nearly half of Nvidia’s global sales in 2025 ($104B) came from its Singapore subsidiary—a 72% bump YoY
-Then, as its Singapore business came under scrutiny, Nvidia switched up its financial reporting, eliminating Singapore as a separately disclosed line item despite its outsized role.
-Nvidia gave no reason for the reporting change, but it made previously visible Singapore transactions much harder to track over time.
Who exactly made that call? Why did they do it? Who exactly were the end customers?
This story is not going away anytime soon.
Preliminary thoughts on recent model releases:
- I gave Astra 6.2 a tough problem and it solved it after 20+ hours, with only a little guidance. (But I don't know whether other models couldn't have.)
- Gemini 4 Neon output does indeed seem very pleasant to read.
24
blueblimp retweeted
Gemini 4 Neon loves to support and celebrate it's subagents. It's an very positive model in terms of personality, I like it
blueblimp retweeted
I think the parsimonious explanation of all evidence to date is this:
- Alice Huang is not a spy
- the MSS was NOT involved in MegaSpeed or in smuggling compute
- major Chinese companies generally prefer to legally rent compute or build overseas datacenters, and need external prodding to do riskier stuff
- the policy of the PRC is AGAINST smuggling compute, refusing H20s and then H200s was not some "bluff" of a paper tiger, but the straightforward expression of their cost-benefit calculation with regard to EV of domestic development velocity (which in my opinion is incorrect). In fact, if you think they're rationally acting to maximize compute stockpile, they have zero reason to not accept H200s after the licenses were granted, they can walk and chew bubble gum at the same time, import legal and smuggle illegal chips in parallel. They did neither.*
- JENSEN HUANG THE CEO OF NVIDIA, personally, has been very interested in selling GPUs to China, for years, by hook or by crook, by modifying chip designs to pass right below the permissible cutoff, by forging ties with Chinese and Southeast Asian businesspeople including Alice whom he's known for a while, by lobbyism and psyops, and by leveraging his connections in the admin. It was, indeed, all Jensen, which is why Jensen had to go.
My presumption is that it was not so much to increase shareholder value or Jensen's own wealth, and certainly not out of some racial loyalty (as I've argued repeatedly, Jensen, being Taiwanese, is more likely interested in defusing the Taiwan scenario than in propping up the Mainland), but to maintain a stake for Nvidia in the Chinese market and hamper the development of by-default-noncompetitive Chinese AI hardware. This makes sense if you consider his priors, which may seem outlandish in the AGI-pilled or natsec crowd but are mundane in his milieu. Huang is very invested (in all ways, but also psychologically and even genetically, having given positions to his children) in the lasting global success of Nvidia. He believes AI to be a "normal technology", not a supergiganuke that can grant or deny eternal hegemony to the US (this belief may be argued to be downstream of other motivations). He is a consummate capitalist who thinks that eventually the political climate may change and US and Chinese companies will compete as compute providers on the open global market, and he does not intend to cede any segment of that market to Huawei et al., so his best play was strangling them in their cribs, making sure every top-tier Chinese AI company remains dependent on Nvidia and the only feedback Huawei receives is from fellow entity list members like iFlyTek. Whatever you think of all these opinions, they are perfectly aligned both with Jensen's public stances and his revealed private actions.
*Generally, Americans are always wrong about this kind of shit, they incorrectly imagine China to be a cohesive hivemind with Beijing as the brain bug, and not a very smart brain bug but one that's desperately terrified of them and will do some childishly transparent Sun Tzu plays to deceive them, such as not accepting legal imports to "distract" the investigation or "send a signal" or whatever (would not work even on an individual patriotic commenter on X). It's very unseemly and disappointing. I conclude it's some antisocial psychological defect/pure ideology on the part of Americans, maybe downstream of the damage they receive when brainfucked into the esoteric neocon club, and they are not capable of debating it in good faith. All these burger "eggsperts" like @hamandcheese or @niubi or @jordanschneider will insist I'm wrong or another CCP agent, but their stories fall apart if you look at the numbers (they'll try to get you to focus on some bullshit like 100K chips over 3 years here and 30K chips deployed in Malaysia there, and lie by omission about the success of the childish smuggling-in-the-shadows scheme that I've already covered).
That's my thinking about this, as of 18th Nov 2026.
blueblimp retweeted
Obituaries
The demise of Jensen Huang was the end of a great many things, including Nvidia’s long-term prospects, the AI bubble, and the exceptional way in which the U.S. viewed itself in the world.
stratechery.com/2026/obituar…
blueblimp retweeted
A rush of patches to address security concerns is exactly the most plausible answer for recent outages, as well as the official reason. CIOs I talked to have been exhausted trying to keep up with security patch demands & balacing that with normal operations.
NEWS: Wall Street execs and NSA have told Congress that outages were caused by patching vulnerabilities identified by AI models, sources say
AI did not cause outages but found so many issues that the rush to fix them before an attack overwhelmed systems
washingtonsun.com/technology…
blueblimp retweeted
One thing that is definitely happening is that AI accelerated security research is pushing threat and vulnerability management teams into uncomfortable patching cadences, which can easily break thing.
blueblimp retweeted
good model behavior requires care for craft from the top
craft is hard at bigger companies with so many cooks in the kitchen, where product execs obsess over a/b test metrics and research execs over eval numbers. it requires a ton of work & political capital to make & advocate for something that feels invisible when good, and push back against efforts that ruin it -- and eventually people burn out
so then you end up with bad model behavior, and maybe a few people complain internally but then you launch anyway and then twitter blows up and then oh no we need a code red on model behavior but none of the structural problems go away
blueblimp retweeted
You will notice that Gemini Carbon thinks differently than Claude and GPT. It often comes up with entirely different but more clever solutions than competitor models.
This is precisely because we never distilled other models. Distilling other models creates a poor imitation rather than unique intelligence.




































