João Paulo Pelúzio

@Noahpinion Good thing US has some companies entering production on sodium-ion too

Cool! And I'd say this is a genuine Chinese innovation. Even though the technology was technically invented elsewhere, it wasn't viable before.

MoonshotAI will overtake OpenAI and Anthropic before the end of the year or will they? at least that's what the hype kiddies on X want you to believe So let me make it falsifiable. They are saying: - China / MoonshotAI is catching up - they are catching up generally (not just coding, but almost all domains and including restricted models like Mythos 5) - the gap is currently ~1.4 months based on Artificial Analysis Index and benchmarks provided by MoonshotAI, where Kimi-K3 beats Opus 4.8 in 30 of 35 benchmarks, and GPT-5.6-Sol in 19 of 35 benchmarks (they ignore the existence of all Mythos variants) - China is not catching up due to distillation, so they should overtake US labs Their implicit prediction then is: - a chinese model / MoonshotAI will overtake Anthropic (and OpenAI) on the Artificial Analysis Index by: - Median: 2026-12-24 (80% CI: 2026-09-17, 2027-09-14) Since they claim that chinese models are as general as american models, we should see unsolved mathematics, physics, and more being solved by chinese models at higher rates than american models. Speaking to its generality Kimi-K3 should surpass Opus 4.8 and GPT-5.6-Sol on the majority of these benchmarks: - METR Time Horizons, FrontierCode, MirrorCode, UK AISI cyber ranges, ExploitBench/ExploitGym, CritPT, FrontierMath T4, ARC-AGI-2 / ARC-AGI-3, WeirdML, ALE-Bench, GSO, MRCR2/GraphWalks - vibes --- Some other things that are more speculative and downstream of China overtaking US models: - more involvement by the USG - stricter export controls on semis - potentially a Manhatten-style project, as we will be behind in 2027 and are racing against China - also in the cards: US banning chinese models or US labs distilling from chinese models --- I have already stated my position clearly. Chinese models are generally ~6-8 months behind, with some domains like coding behind slightly less. Kimi-K3 did not significantly shift my estimate on the gap and it currently does not change my outlook on the future, but we will have a MUCH clearer picture once we have all the benchmarks I mentioned earlier. The main reasons for my position: - Kimi-K3 doesn't even beat Mythos Preview, a ~5 month old model - We will likely not see much larger open models than Kimi-K3 for several months, likely not until early-mid 2027 - Meanwhile Anthropic is sitting on a 10T model since ~February, OpenAI likely just finished the training of GPT-6, which should also be around that size, and more 10T param US models are coming from SpaceX AI, Google and Meta. - We are currently not seeing the true frontier of models. Anthropic and OpenAI are currently sandbagging as the legal situation for releasing new frontier models is unclear. - Historically, chinese models have been more benchmaxxed than US models, meaning their benchmark numbers do not translate to real world performance as well as their american counterparts - GPT-5.6-Sol is still 2-3x more token-efficient on the Artificial Analysis Index than Kimi-K3 (while likely being smaller, ~2T) - US labs have more compute --- I'm very happy that Kimi bros released this model. It's a great model and probably the first really useful chinese model.

@BTC2TMOON @howardlutnick Is to them im talking about, it is completely stupid, data centers are money printers

@PeLuzZIo_ @howardlutnick I’m the 1% bud. I’m just telling you how your average person in the US is thinking. It’s so popular that states are halting projects due to pressure from their constituents.

When President Trump said we were bringing semiconductor manufacturing back to America, he meant it. Today, TSMC announced another $100 billion investment in the United States, bringing its total commitment to $265 billion. Promises made. Promises kept. 🇺🇸

A few thoughts on Kimi K3: - The Moonshot team is absolutely cracked. Some of you might remember that K1.5 was released on January 20, 2025 - i.e., on the same day as DeepSeek R1, and it was ~just as good as R1. IIRC, the model wasn't open-source, which is why no one ever talks about it and DeepSeek got all the glory. - Do K3's capabilities generalize outside of coding, or is it coding-benchmaxxed? In my initial testing, it's materially worse than GPT-5.6 Sol and Fable 5 for certain use cases that do not involve coding. - I will not tire of saying this: the publicly available models in July 2026 *do not matter*. It doesn't matter whether China is 2 months behind the frontier, or 4 months, or 8 months. The only thing that matters is the race to RSI. What do we need for RSI? IMO, two things: (i) a model with excellent research taste; and (ii) tons of compute. Does excellent research taste develop on its own from increasingly strong coding abilities? I'm skeptical. I think it's a much more general ability that is more a function of overall model intelligence than just proficiency at some skill (like coding). - All that said, my estimate for how far behind China is has not changed at all with the release of K3 (IMO, much farther behind than the timeline thinks today). - This model's cyber capabilities are currently unknown, and this is an area worth watching. The few weeks following the release of K3's weights might be interesting times for defenders (or, then again, maybe not). - The geopolitical consequences of the release of this model are also worth watching - both in the U.S. and in China.

@BTC2TMOON @howardlutnick Imagine being this fucking stupid

@howardlutnick This makes you popular with 1% of Americans. 99% will hate you for bringing in Datacenter investments.

@nikicaga Sure, let Dems cripple their states so they can have some power, even as polls are showing how bad it actually is for them when they are against psycho Potus 2.0, bunch of imbeciles

The data center stuff is going to be pretty horrible for Dems long term, but winning power short term is pretty important rn, so let them do what they must

@Noahpinion Good thing rethoric has changed a couple of months ago

My take on data center moratoriums is that they don't matter. Data centers will go somewhere else and AI progress will not be affected. But AI companies should realize how their toxic doomer rhetoric has scared people into hating anything associated with their product.

@m_goes_distance All in US?

biotech ideas we're funding rn: - Stripe for gene therapies. - GitHub for wet labs. - Cloudflare for clinical trials. - Plaid for biomarkers. - Bloomberg for biological age. - Costco for GMP manufacturing. - Anduril for biosecurity. - NVIDIA for programmable biology. - Cursor for molecular biology. - A "WHOOP for cells," not steps. If infrastructure precedes the science, biotech accelerates 10x faster if youre backing this wave, i'd love to chat.

@aaronburnett What? From data i gathered US seems to be investing more overall in orbital data centers

Same country investing an order of magnitude more into orbital data centers than the US. Wonder if there is something to that.🤔 Orbital infrastructure Cambrian explosion is coming whether the armchair engineers believe it’s possible or not.

@JamesRaxz It ia really funny how americans keeping regulating themselves into oblivion while China makes everything easier to build and scale, regulation, financial and industrial policy wise, if anything, this is just pathetic

It was a nice strategic advantage while it lasted.

@tomieinlove If 70% of your household wealth is tied to real state, yeah, it is pretty shitty this level of devaluation

It's a pretty damning indictment of our own societies that the price of housing going down is considered bad.

@jack_watson_hfw @Gaurab Not only that, but many new entrants like Vulcan coming to levek the playing field

@Gaurab The good news is that there are still several US companies that design/make separation equipment. We work for many of them. So while the base has diminished, it is not completely gone.

The United States has roughly two dozen people who know how to run rare earth solvent extraction at commercial scale. The Chinese Society of Rare Earths has over 100,000 members. In 2023, American mining engineering programs graduated 162 students. China graduated about 3,000 from 45 programs. Twelve US universities have shut down their mining engineering departments, including UC Berkeley and Ohio State. The Bureau of Mines, the federal agency responsible for mining research and training, was dissolved in 1996. Chinese pricing drove Western rare earth operations out of business through the 1990s and 2000s. The people who knew how to run those plants retired, changed industries, or died. Rare earth separation requires understanding how 17 chemically similar elements behave across hundreds of interconnected mixer-settler stages over months of continuous operation. That knowledge is accumulated through years of hands-on work in facilities the West stopped running. The average US mine worker is 46. 221,000 are expected to retire by 2029. Building an entirely new critical minerals supply chain means rebuilding the workforce that was eliminated a generation ago. The operational knowledge only comes back by running real plants.

@lthlnkso If only US hasn't made China more favourable to manufacturing given so much regulations US has, good thing this it is changing now, but it needs to be faster

An obvious consequence of AI and robotics is that military power will increasingly depend on being able to manufacture and deploy autonomous weapon systems. This is uncomfortable to contemplate given China's vast manufacturing advantages.

@YuppyYellowKing @Noahpinion Actually, 61% of overall independents are at least proud of being americans and more than half of the remaining are simply not sure (which likely changes depending on presidency). This trend is not anywhere close to dems npr.org/2026/07/01/nx-…

@Noahpinion People ignore though that this is also massively true with independents and isn’t a dem problem. This just seems to be something republicans are causing for everyone else.

Anti-Americanism has become the normal, consensus view among partisan Democrats in America, and this is very bad. We need to turn this around, and "just elect Democrats to the White House" is not a permanent solution to the problem.

@cixliv @resolvedmotion Many are in house designs anyways...

@resolvedmotion The ones actually shipping are using Chinese parts

The reason you don't have any of these in your living room yet is because the parts to make them take too long and are too expensive, namely, the actuators.

@aphysicist @paulg More of those companies likely to appear too

@paulg as hardware development cycles accelerate, these types of companies will only be moving faster

Someone asked why I like Boom so much. Partly because they're building supersonic jets! Partly because they're doing it so well. And partly because we've been through so much together; it was so hard for so many years.

@JosephPolitano Nothing new, industry is also digesting construction from IRA and Chips Act and spending more on workforce and tooling overall, it is a N shaped curve, eventually it will grow again

American factory construction continued declining in official data released today as CHIPS Act projects finish, IRA projects get cancelled, and tariffs weigh on nearly all industries Total US factory construction is down 30% from its 2024 highs and 20% over the last year alone

@andrewkornuta What you dudes are talking about? NOX fellows keep posting many vids of metal cutting and updates on their facility, you just seem butthurt somehow

@jakeknudsen5 @mattfreed @boxcardavid UP those numbers by some orders of magnitude, my brother, i believe in you

We are excited to update our site with some of the awesome companies we have had the privilege of supporting in the last 6 months. A very massive thank you to some of our first supporters, @mattfreed and @boxcardavid. We have grown super fast over the last 6 months, and we look forward to expanding our capabilities and offerings to help serve the future of the physical world. @SkydioHQ @tetradynamics @mundanebot @pivotrobotics @AskariDefense @westmagco

@HoosierInfidel @cremieuxrecueil Huh? Just say you don't like Data Centers and don't know what you're talking about at the same time, it is less pathetic

@cremieuxrecueil Not when the state law and regulators let them raise rates instead. Why add supply when you can just increase everyone’s rates?

Or ppl who don’t want our electric bills going up because state govt won’t tell the utilities to make the big customers pay for their usage and impact.