So Chinese labs are driving essentially towards commodotized intelligence. Even if its a few months behind the US.
Is this a classic 'commoditize my compliment' situation? They want to sell the hardware and infrastructure behind AI and make the software part not the value driver / moat?
I can see it. But also even two Chinese labs sinking 100s of millions USD into training isn't exactly commoditization. It's still a ton of effort with dubious payoff.
If there was some grand strategy for all Chinese labs, surely it'd have leaked by now. I think its more likely that:
- Companies can still make money from commodities
- Chinese labs only have 5-10% the valuation of OpenAI/Anthropic, so massive monopoly profits aren't necessary. Profit expectations for tech companies in China are really low in general, complete opposite of the US.
- Open weighting is a great way to get talent/attention/reputation
Apparently open source models is an officially stated national strategy now, which is somewhat reassuring that we'll probably keep getting these for a while.
Until it's no longer officially stated strategy, and something else benefits the CCP. See rare earths policy. Changes constantly, to whatever maximizes their leverage.
The aspect that you can count on is that China sees AI as a long term strategic necessity, so they will support domestic players for an extremely long time. But unless you're a Chinese lab that shouldn't necessarily be a comfort to anyone else.
That's how nearly everything works. If people don't think action is in their best interest they don't act. And you can substitute any country for China - although if anythign the Chinese have been unusually competent compared to everyone else. the US has been adopting a foreign policy with a deliberateness of action that could be likened to someone having a seizure.
I expect the Chinese will do something wildly stupid at some point, but they've been behaving quite responsibly so far.
In his speech https://www.xinhuanet.com/politics/leaders/20260717/72728b6f... Xi Jinping mentioned "open source" (开源) exactly once in a list of buzzwords. Deriving from that an "officially stated national strategy" that ensures the free beer won't stop flowing is motivated reasoning. The most popular LLM chatbot in China is Bytedance's Doubao. Do you think this national strategy will involve forcing them to release their weights?
Another buzzword mentioned by Xi is "safety and controllability" (安全可控), which is generally in tension with open source, since that involves giving up control. Historically, when there's a conflict between openness and controllability, the Chinese government tends to choose control, see the Great Firewall.
I'm not really sure that China's labour costs can be described as "low" these days, going by GDP per capita [0] they're above the global average. They're a little lower than the Europe median I think but the gap probably isn't strategically important.
The Chinese edge tends to come from their environmental laws; they allow their industry to do things that the West would ban.
I have no source for this but I feel like the environment thing is too frequently repeated (also without any evidence). They're pumping out solar like mad and places like the Xinjiang deserts actually have nowhere to put their excess electricity from what I've read previously.
They also have a "cancel" culture pointed at the right place, in that whenever there is some exposé of some xyz it usually does get some reaction from the authorities. Even if in the end no real punishment is meted out due to corruption or politics or cover ups the very least is that the perps gave to change a hiding spot. Which is very much unlike the datacenter gas turbines running in people's backyards in the US right now.
Furthermore if you take a look at the Politburo meetings and political documents etc it is clear that Xi is actively trying to solve the environmental problem. In the latest people's meeting it was also a talking point. It is a matter of "face" for them at this point. Contrast with "drill baby drill"?
It is easy to talk about the "environmental" problems in China but I think it is very last decade.
As a disclaimer I live in Asia and I frequently visit China but I'm not Chinese nor do I get any so called "social credits" which actually don't exist in China. The information gap between what the west thinks they know about China and what it actually is is sometimes kind of just a huge joke.
I think the environment thing is really overblown as well.
What’s not overblown is the ease with which the government can get big projects done. And there are some downsides to that, specifically when it comes to property rights.
The Beijing public transit trains are a great example. The routes are wildly straight. So straight they allow extremely long trains.
Thats possible, in part, because the government can just reappropriate the land in a way that governments in the west can’t without engaging with a contentious public.
And that’s just one example. That sort of thing permeates many different kinds of projects and leads to a particular kind of efficiency.
I've been saying the same thing for the last 8 years to my european and american friends. Maybe in another 8 years they'll finally realise, or they will quietly adjust to the new normal where China is no longer viewed as this cheating, dirty and backwards competitor.
> I have no source for this but I feel like the environment thing is too frequently repeated (also without any evidence).
> They're pumping out solar like mad and places like the Xinjiang deserts actually have nowhere to put their excess electricity from what I've read previously.
You've found a pretty compelling piece of evidence right there. There is a reason they're the ones pumping out solar like mad - because they have relatively lax environmental laws. If they had western-style environmental laws the panels would have to be manufactured in India or Africa or something.
Obviously it isn't the only criteria because I'm sure there are places with lower environmental requirements than China; but the big selling point for the last 20-odd years seems to have been more that you're allowed to manufacture there and they have a competent workforce rather than that the wages were amazingly low.
Reminds me of my first trip to Japan, a lifetime of robotech and ghost in the shell had me shocked and disappointed at how run down and stuck in the 80s Tokyo and Hiroshima was.
If you study the last 5k years, then the real question how on earth did the Europeans (primarily) and then the Americans managed to stay ahead of China for as long as they did (the "Needham Question").
Confucius came before Plato and most important discoveries came from China. The two wrote a the "good citizen" ought to be and how a city/country/empire ought to be run. Even meritocracy has been traditionally stronger in China, over centuries.
In the west we're talking about it since Homer, but in practice has been applied sporadically, while China has been much more consistent.
China's labor costs are low for most labor-intensive manufactured goods that higher-income countries import from them. They're high for most labor-intensive manufactured goods that China imports from lower-income countries. It's all relative. Not all manufacturing is super polluting (nor is all manufacturing labor-intensive), but if you tried to replicate a non-polluting labor-intensive Chinese factory's cost structure in a high-income country, you'd still run into the problem that the workers you'd need to hire have better options available.
I'm not sure how true is this claim about weak environmental laws. I don't have any data to back this up, but I was in China in 2013 and two years ago and the difference in air quality in major cities is staggering. So, maybe they destroy the environment but for sure they no longer do that near the big cities.
To be honest I do not thihk you are wrong. Not in a way AI talent is underpaid, but they unlikely need to offer $1,000,000 sign up bonuses to just create AI lab. Its still pays better than average SWE job, but not x5-x10 more like in US.
Though in manufacturing China is dominating because of immense optimized supply chain, economy of scale and now workforce with know-how.
I honestly cant compare with US since I'm not from US, but to produce something in EU you will have weeks or months lags on logistics of source materials and components. And some industries only exist in some countries, etc.
In China in one single place you have anything you need to produce from a toaster to a car in hands reach or might be few days at most.
And I guess China did the same with education. Education is a lot cheaper there so they just have more talent.
Are you implying air pollution in China is not bad anymore? I wish that was the case because I would visit a lot more but you need to educate yourself a little bit just go look here: https://aqicn.org/map/china/
I'm sure it's better than it was before and it's not India bad but it's almost always considered unhealthy in most parts throughout the whole year
Great, so you plonk data centers into the desert. Now you need to cool them, and bestow them with excellent network connectivity. Those are things that isolated, arid locations cannot do very well.
So, plonking data centers into the desert (or on orbit) isn't really a silver bullet, is it?
China has an explicit national policy to build high bandwidth network connections across the arid Western half of the country called the digital silk road. And on the scale of China's other projects in the west (e.g. mining, large scale agriculture, chemical plants), the water usage of a data center or ten is a rounding error.
I don't know if they're actually building AI DCs out there, but it doesn't seem completely unreasonable like you're suggesting.
Of course it doesn't seem completely unreasonable to build out power distribution, cooling, and high-bandwidth network interconnects throughout a region, but once you've done that, you are no longer facing the empty, barren desert scenario.
So if China builds out infrastructure that can be maintained in perpetuity, then more power to them, and the NIMBYs won't have a leg to stand on.
Conversely, in these United States, I've seen literal tons of data centers being plonked into urban centers: literally at intersections where lots of people drive by in daily commutes, there are data centers now where there weren't before, and they are also found on the outskirts of civilization, because they require more security than accessibility. So data center siting can take multiple factors into account here, depending on their purpose.
And can we please stop hammering on the "AI Data Center" meme, because in my experience, a DC is a DC, whether they choose to make it AI or not AI, they're probably all AI from now on, but it just adds fuel to the Luddite fires to say so...
Oh, they're trying, and the locals are not happy, due to both the rise in electricity usage (mostly NOT solar) and environmental damage. Look up the datacenter planned in Box Elder County, Utah.
- AI investment is basically the only thing keeping the American economy treading water at the moment, and kicking the chair out from under that industry benefits China tremendously.
I'm definitely not saying that's the only factor, but I think it's naive to assume it isn't at all a factor.
I don't really believe the idea that there's some grand plan by the Chinese government to use their AI labs to undermine the US companies. Maybe now it's more of a reality, but I don't believe it was there from the start.
That said, if were going down the rabbit hole of saying that the Chinese labs are now part of a larger geo political strategy by the Chinese government, then I think Taiwan is part of the equation. If the frontier models require the best chips and those are mainly coming out of Taiwan then it's hard to imagine a world where the US allows China to make a move on Taiwan without a fight. If frontier level models can be run on chips being made elsewhere, then Taiwan becomes less important geopolitically. I don't think we'll ever get to a world where the US is just like, "Fine China. You do what you want.", but China has to assume that there will be a lot less resistance from the US is Taiwan isn't such a key component in the AI race.
In the same way that the US going to war with an important Chinese oil supplier: 85% of Iranian oil going to China and comprising ~14% of China's oil imports.
There's always geopolitical reason behind the reason.
I think it’s pretty wishful thinking to believe that the Iran war was some complex N-dimensional chess to make Chinese oil costs higher.
Even if the current President was known for his strategic global thinking and ability to keep quiet about the “real goals,” other explanations seem more likely.
It's not, but it's not coincidence either. The reason for the US entanglements in the Persian Gulf intensifying are because it's a choke point for oil, and the reason the US feels freer to mess around is because the US became an oil exporter. So the fact it's hurting China more than the US (in some ways) is related to the factors that led the war to start, even though it's not 27D hyperchess.
You think too much. Trump got duped into a gamble and didn’t win as he thought. If the hope is to hurt China it’s unlikely to be working very well. Their energy sources and oil sources are very diversified.
Trump wasn't 'duped'. He did underestimate the level of involvement in Iran that would be required.
It is hurting China. Their economy has slowed to a crawl as a result of stalled oil and has delayed any real gains they had by at least a couple of years.
Bibi tried many times. Every president before Trump understood the risks and refused the wild plan. Only Trump fell for it.
China’s factory outputs are in fact increasing. So is their export. You wouldn’t see this if they are energy constrained. China’s problems are weak consumption and deflation. It started way before this year and has very little to do with Iran.
He was definitely duped by Benny. Him underestimating was part of the dupe. He did not just come up with this shit on his own. Someone fed it to him, and it's most likely Benny. And he slurped it all up. And yes, he made the decision, this is well documented. You can't just push the blame onto advisors. In that case Hitler would be innocent, he could just shove everything on his mad advisors.
It's certainly more complex than all but a handful of people know.
For example, part of the dynamic was, according to the NYT (though they may have been fed bad intel), that Ahmadinejad was apparently working with Mossad and could've been a partner for a new leadership team in Iran. Who saw that coming?
It also brings to mind Kennedy. He campaigned hard on, in part, the "missile gap" with the USSR. Gets into office, gets read into the real intel, realizes the "missile gap" was essentially a MIC psyop of strategic 'leaks' to the media and that the US was in fact far ahead of the USSR in strategic arms. Kennedy was furious -- but due to it all being classified, could say nothing in public. But it did start a pattern of mistrust that lasted his presidency between him and the MIC. That all didn't come to light later.
Later, like the Ukraine/Starlink thing. The media jumped all over Elon for allegedly holding back Ukraine and shutting off their terminals close to the front lines. More recently it turns out that was the US government, they were afraid Ukraines rapid progress might tempt Russia to deploy tactical nukes. Elon was innocent.
Theres layers upon layers of things we don't see and won't know for decades after the fact. And as Hillary showed with her email server, digital records are more ephemeral these days, so there will be a lot of things we simply never know.
> And as Hillary showed with her email server, digital records are more ephemeral these days, so there will be a lot of things we simply never know.
Ironically, part of the story behind the email server was that State Dept IT was so bad that provisioning a new email server was a months-long request.
There is definitely strategy and direction coming from the top. CEOs have been summoned for conversations. There are new rules like no Ai lovers that certainly show they are thinking about things.
The AI CTO from The Linux Foundation, Matt White, spoke about his recent trip to talk to the Chinese Ai labs ~month ago.
> Chinese labs only have 5-10% the valuation of OpenAI/Anthropic, so massive monopoly profits aren't necessary. Profit expectations for tech companies in China are really low in general, complete opposite of the US.
It’s really amazing to see that the competition is creating better quality models for everyone - and am really happy that some of these are open source (or partially os).
Regarding the valuation, that maybe points a finger to the over valuation of the US companies ?
> Either Chinese labs are worth more or OpenAI/Anthropic are worth less. One of those is true
The reason that's not inherently accurate is because it assumes the economic models are identical. That couldn't be further from the truth. It'd be like pretending a globally dominate Google search engine should be worth as much as eg Baidu or Yandex. No doubt Baidu is an advanced search engine, and that's not what defines its economic characteristics: the markets it operates in does, the market ownership does, the ad engine does, the ability to sell does (advertiser relationships), etc.
OpenAI is very likely to bolt on a massive, global advertising platform to GPT over the coming years (and yes, it'll take time to build up). The Chinese labs will have a very hard time replicating that to a global audience the way the US tech companies have repeatedly proven they can. OpenAI will poach proven ad system builders from Meta and Google.
Also at scale Kimi (or DeepSeek and others) don't have to worry about local deployment. The number of param keeps growing, I highly doubt consumer hardware will catch up. I am just glad Qwen team is still putting up great bangers.
This is strategy by Chinese government, so much of US economy is invested in AI. Releasing free or cheap versions of the models undermines US economic growth. It’s asymmetric strategy that makes sense if you are close second in AI race. If situation is reversed, US would do the same.
Google did this by creating Android to undercut Apple. Many US tech firms did this strategy in the 2000s and 2010s of supporting open source alternatives to their opponent's closed source money maker, to undercut the competition. Back then, it let open source have a big boost and we all benefited from that. Hopefully open weights models will do the same so that AI can be more democratized. I wouldn't want to live in a world where, for example only Anthropic or MSFT have top AI and the rest of us have nothing.
Android wasn't created to undercut Apple. It was started long before the iPhone project was public. Android was created because Google felt their apps could be huge on mobile (correct) but that mobile operating systems of the time were painful and frustrating to develop for.
Android was even an acquisition. However, it was open sourced after the iPhone. The iPhone was announced in January 2007 and released in June 2007. The first AOSP announcement/source drop was in November 2007.
I think it is very likely that Google open sourced Android to undercut iPhone. Android was initially developed for phones with a keyboard (similar to Blackberry). The introduction of the iPhone made it clear that touch was going to be the future, so Android was quite far behind before it was even released. Besides that, my recollection is that the first Android releases were pretty bad compared to iPhone OS. I know a lot of people (not necessarily Apple fans) who looked down on Android.
Open sourcing Android was a great move to rally manufacturers around Android and gather a large group of early enthusiasts.
No, open sourcing it had been the plan from the start. I was there, I was at the internal meetings where these plans were announced to employees.
The main impact iPhone had on Android was delaying it to redo the UI to make it nicer. Not trivial changes but not fundamental business strategy changes either.
First Android releases weren't bad at all. I'd say the UI was more 'classic' feeling and not as nicely animated than the first iPhone, but the first Android actually had far more features. And some critical ones too, like you could actually write apps for it.
> First Android releases weren't bad at all. I'd say the UI was more 'classic'
Maybe in your test labs, but on real HW it was sluggish, laggy and barely usable. Source: I owned the original Samsung Galaxy and tried some friend's HTC Magic and they were horrible. The first real good Android phone, perf wise, was IME the Galaxy S II.
Android the technology wasn't, but Android the commercial product along with its business model was.
The first iPhone came out in June 2007. In November of that year Google and partners announced the Open Handset Alliance and the open sourcing of Android.
> Releasing free or cheap versions of the models undermines US economic growth.
This would be like saying that the creation of the PC undermined the 1980s economy because it hurt IBM's profits. Instead the consequences of cheap, ubiquitous personal computers grew the economy 10-fold.
There is an entire class of wholesale model providers that stand to gain from open source models. Then there are companies building platforms on top of LLMs that would otherwise be impossible with closed models due to cost. And there are entire enterprise use cases that would simply be non-viable at $50/million tokens, like OpenClaw, Hermes, etc.
This is a talking point that plays into the frontier labs' desire to be seen as "too big to fail". While yes, several hundred billion dollars have been invested in AI, (a) much of this is in the form of circular Monopoly-money deals, and (b) the US GDP is over $30 trillion annually. The real economy - the one that makes food, builds homes, provides medical care, etc. - is so much bigger and more important than the AI industry.
That's not to say that an (inevitable?) AI crash won't be the spark that ignites a big recession. We are well overdue for one.
> The real economy - the one that makes food, builds homes, provides medical care, etc. - is so much bigger and more important than the AI industry.
The real economy has seen pretty poor growth under Trump tariffs chaos though and I'm not sure the US economy could survive a crash of the tech companies
It's exactly the same strategy followed by the Silicon Valley for the past 3 decades to beat the IBM-era companies. Google gave everything free. Social media stuff was free. Video calls are free. Bloggers gave content for free. A lot ot open-source product developers did managed offerings of the free products.
Free + Open is simply how you earn credibility and user trust when you don’t already have it.
That doesn’t mean the trust is unearned once gained, or a bait and switch, or purely Machiavellian either btw.
Consumers and businesses need credible branding to feel like they can trust vendors who provide them with the products they value or deem mission-critical, because it creates accountability and makes it less risky to depend on.
Open source addresses the credibility/accountability/branding/counterparty problems simultaneously, and adds to a permanent intellectual commons we all benefit from. It’s legitimately just Good
Chinese companies also have invested money and people to grow AI model，though China's energy costs are far lower than everybody else's. Combined with lower labor costs and currency arbitrage, it's completely natural that Chinese models are priced so aggressively. I use them myself, and the price-to-performance ratio is honestly unbeatable.
It's just too much 5D chess to be possible. Government spending billions into AI companies and pressuring them to open source at a hope at denting openAI and co? Why not just pour the cash into electric cars or steel where there's a guaranteed return? It's just absurd.
At a state level it's a pretty good defensive strategy against the American AI industry, as a whole, gaining a monopolistic advantage. This prevents the US from using this as a coercive leverage through things such as tarrifs, export bans etc.
In a world, that's increasingly dependant on the AI "opium" that the US is dealing, it's conceivable that the current administration could try to sabotage something like, say Chinese-European relationships, by threatening to cut access to Anthropic or OpenAI products for Europeans, if China cosies up to the EU.
The disadvantage of trying to use these leverages, is that once the genie's out of the bottle, the other party will divert their focus quickly, so it only works if the US truly has a choke-hold on frontier AI. Otherwise, you just scared the other party into never trusting American frontier AI ever, and you didn't even truly hurt them, because they already have a quick fix from China.
Unfortunately behind a paywall, but there's an article in the previous issue of Foreign Affairs about how the Trump administration has fumbled this coercion strategy repeatedly because they overestimated their advantage in different markets (tariffs on Canada, Iran, etc) and what an actually effective strategy can look like https://www.foreignaffairs.com/united-states/how-fight-econo...
tl; dr: You need to have a monopoly, the enemy should not bounce back quickly, you shouldn't cripple your own economy doing it.
AI may just be the next economic offensive the Trump administration fumbles, because China planned ahead by incentivising development of close-second alternatives to their frontier models.
1. There is a “space race” mentality happening at the national level with respect to AI. So, China is committing to the race.
2. Even if the race turns out to be a dud, China is hoovering up massive amounts of data as customers throw everything into their prompts. This is useful for all sorts of national objectives. Why hack when you can just put up a shingle that says “Artificial Intelligence” and customers hand over their data willingly?
National strategy. Thanks to smart investments, China's energy costs are far lower than everybody else's. If intelligence becomes a commodity, then China will be providing it for the world. Incredible leverage.
Whatever you want to attribute it to, the Chinese labs would get zero attention if it weren't for their models being downloadable, so it's a marketing play. You can't run their largest models on consumer hardware, it needs a hosted service/professionagrade equipment. Which means using their service becomes an option, especially at their pricing. So that's the play.
Open weights are indeed a commercial risk because there's no shortage of companies that'll make a business out of running MaaS (model as a service), close & resell a fine-tuned or distilled variant (DeepSeeking of Llama / Cursorification of Kimi, for example).
China sanctioned and voted for UN resolutions to sanction Iran in the past. Then Trump baselessly abandoned the JCPOA. Even the US own DNI said it is the consensus of all 18 intelligence agencies that Iran was in compliance and not seeking nuclear weapons. China, correctly, called the renewed sanctions illegal and refused to follow them
Energy is not the main reason. China would still stand with Russia without energy.
China could never support a pro-Western Ukraine in the first place. The same logic applies to Iran, North Korea, etc.
Although China cannot directly express support because of foreign trade, things like the Communist Youth League playing "Katyusha" on the day Russia launched the invasion, and the information direction on social media, have already made their position very clear.
> Energy is not the main reason. China would still stand with Russia without energy.
Of course it will because Russia also has cheap raw materials not just exclusively oil and gas. At the same time China does not invest into Russian economy because they don't want competition.
Weak Russia is more beneficial for China to buy commodities for cheap and then sell Chinese products back.
> China could never support a pro-Western Ukraine in the first place.
China is one of largest Ukraine trade partners before and during the war. Ukraine even lobbied for ability to buy Chinese drone parts using EU funds when EU unable to fulfill.
So nope - China dont care. They are trading with both Russia and Ukraine.
Having trade relations does not mean China supports a country. US has always been China’s largest trading partner, but China <-> US relations can hardly be called "friendly".
The CCP government has absolute control over domestic public opinion. The pro-Russia direction on Chinese social media is very clear, although as the war became unfavorable for Putin, voices supporting Ukraine have been allowed a little more in recent years.
What the US supports is what China opposes. From Iran to Venezuela, from Gaza to Ukraine, except for a very small amount of "human consensus", it is always and will be like this.
First things first: the CCP must dominate China itself. For this job one is the global discrediting of democracy in favor of cults of personality. This is the single objective of the Putin-Xi-Khamenei information system. China global hegemony would merely be an instrument of CCP internal control. Similarly Khamenei supported Hezbollah, Hamas, Houthi, Iraq militias ... to control his own population.
Destroy any hope of profitability, prevent further capitalization, ultimately bankrupt them.
Hasten the popping of the AI bubble.
Drag down the stock market.
Put a dent in the GDP (alleged growth).
Cause investors to pull back, further depressing economy.
Devalue the dollar. Kicking off an interest rate doom loop.
Challenge USA's hegemonic role in the emergent multi-polar (post neoliberal) world order. (Gleefully supported by the ruling coalition's anti-globalist America First faction.)
US "AI tech giants" are foaming at the mouth about "dangers" ans "safety", about "how US customers can not be trusted with AI", how much everybody should pay them, and how many $Trillions should they make, at US customers' expense.
For each "AI tech giant", there are millions of other US companies that will benefit from broader AI availability. US economy (and US tech economy) != Anthropic + OpenAI.
Not US-centric, top dog-centric. The US is just the current top dog.
The US is a barrier to a number of key Chinese goals, like reunification with Taiwan. So of course most of what they do will take the US into consideration.
Not everything is an 8D chess conspiracy mate. If they wanted to do this, they’d wait until OpenAI and Anthropic were at their absolute peaks of investment and then release an open source model that was as good. That’s not what they’ve done at all.
A hundred million is table stakes. They get better results by stoking the fires (bluffing) and keeping their opponents locked in a war of raising and re-raising the stakes. By amping up the pressure, they drive more and more investment into the bubble.
China also benefits massively through all of the commodity hardware they make. Sure, they aren't competitive in the chip arms race, but they dominate all the component and power electronics markets. Think about all the power supplies and other commodity components needed to build a gigawatt-scale data centre.
Granted I've only used it for a few hours, but to me K3 still appears below Sonnet, even below 4.5.
I have their highest subscription, so it's not that I can't find uses for it and it has sides to it I really like, and it performs really well in some situations, but 2.7 also gets totally lost on tasks Sonnet and Opus has no problems with, and it looks like that is still the case with K3.
That said, I'm doing things with these models that are a lot more complex than the average app people will throw these models at, so I'm sure there are lots of use cases where it will perform better than what I'm seeing.
I don't know about Kimi, but in my experience other providers like z.ai and minimax don't serve the same models via their subscription plans as they do with api pricing.
They're clearly quantised - spelling mistakes, wonky thinking section dividers, missing whitespace are the immediately obvious tells but along with that comes degraded quality, and it seems to vary based on time of day.
I wouldn't be surprised if Kimi did similar things with their subscription plans - try using it through openrouter and see if you notice different behaviour.
Frontier labs release frontier models to the public only if there is market pressure to do so. Anthropic is not even hiding that they have been using Mythos internally for months now.
I wouldn’t be surprised if OpenAI (so much for “open”) is using GPT-6 internally already.
It appears that peasants like us are not going to get access to frontier AI anymore at any price.
Anthropic had Mythos-Preview for many months internally, but from available sources was an active work in progress, and it seems they started releasing it via Project Glasswing to partners before the final checkpoint was available.
very unlikely that they are holding models back. models are very quickly depreciating in value and have internal cutoff dates. that would really not make sense from a business perspective. and doing an entire training run without commercializing it also seems like a huge waste.
That assumes progress is linear, which it certainly is not. We’re also assuming Chinese companies are releasing their best stuff when our own government is dictating what American companies can and cannot release.
I don’t know for sure that they are weeks or months behind. I doubt anyone outside of 3 letter agencies knows that. The pace of AI is crazy fast and China is notoriously secretive. We could be comfortably ahead or China could have the top model by the end of the year.
I don’t think any of us have enough information to know what’s really going on, but I suspect it’s a very tight race.
Mythos/Fable release was significantly delayed because Dario is pushing for regulatory capture.
It also plays into the hype because there was no one else releasing a model better than Opus. The same thing was with the scarcity crap. They said Fable will not run on subscription, then they will run just for a little while and then forever... Once you get the game they are playing it starts to become quite laughable. If you check his history he was doing the safety/terminator stuff since the days at OpenAI when the models couldnt even calculate 1+1 reliably
Is there any actual proof of this? Everything suggests it's a sincerely held belief and there's been no whistleblowers claiming otherwise. "Since before they could add 1+1" works against your argument because the models had no commercial value, no guarantee they would ever have commercial value, and plotting regulatory capture was about as serious an idea as regulatory capture of antimatter engines.
All three models were quite likely undergoing training at the same time, with Ant a couple of months ahead. Glasswing was announced 3 months ago, Sol and Kimi were already being trained at that point. They are taking snapshots and doing experiments the whole way through anyway and all models will continue to be trained after release. Kimi 3.1 could be better than Fable 5.1
Nah that’s not how that works. There are literally an endless stream of examples of the model providers tweaking backend stuff and improving prompt responses without using new or retrained data. And even if they added new training data to an existing model, it isn’t true to say that model ‘just came out’. But even then, when providers release better models based on new training data they release it as a new dot version, which boosts their sales from the announcement.
Fable has been out for more than a month - I didn’t have the preview version and I was using it around June 10th, when my Claude subscription expired. Saying “Fable only really came out 2 weeks ago” is just factually incorrect all around.
It’s a nationally-sponsored arms race with Western proprietary AI, so as long as the capital sponsors of Western AI lose their shirts and Chinese LLMs become ubiquitous, I think that qualifies as a win condition. Individual labs are no different than running a local restaurant — yeah, it would be great if you land a massive success story, but just making a bunch of money before moving on to the next thing is cool too, even if it’s just via state-sponsored venture capital.
It's the same reason Meta open sourced Llama and AMD open sourced FSR. When you're behind it is a prudent strategy because it undermines investment in the private frontier. Once you're on top you pull the rug and go closed source. There are no morals in this anywhere to be found.
> 100s of Millions
That is utter peanuts given the stakes. This is competition between two super powers for the most important technology in human history.
A more charitable reading of this comment is that you can make a game theoretical argument for these behaviors that doesn’t require altruism. And that argument therefore doesn’t preclude going closed at a later date.
“He’s only being good because he likes how it feels, or values goodness, or exists in a social context where doing the right thing is socially rewarded! That has no bearing on whether he, intrinsically, is good!”
Not to get all philosophical but this makes no sense outside of the context of a very specific post-Protestant, engagement/outrage-driven social media context.
If people are “only being good because it’s in their best interest” the last thing you should be doing is arguing against valuing good things, or making it impossibly difficult for someone capable of doing good to be trusted. Also literally the basis for Western (Plato, good as attractor state, res publica) and Eastern (kongzi, filial piety, social harmony) civilization btw.
I think what they're saying is that any morality related arguments during the free phase are irrelevant because their actions are not based in morality. Put differently, they aren't giving you something for free out of the goodness of their hearts.
I think it’s one thing to give you something free forever, and another to deliberately foster dependency / suck all the oxygen out of the room just to abuse it.
It’s also not a binary thing. You could truly start off with the noblest intentions but succumb to lesser-evil thinking or unforeseen political/personal complexities, lose influence or control to those with less pure intentions, be bought or become a political pawn in a more extractive endeavor without realizing it, etc.
Reasonable adults generally understand that “free stuff” costs real time and money to provide, and that businesses can only sustain it when it helps them sell their products. Unfortunately that means “free” appeals most to people with lots of time, no money, or lacking in the reasoning/adulthood departments.
In an extreme sense, yes. If the free stuff was given for a prolonged period of time with the expectation that it would continue indefinitely.
In modern terms, it's enabling others' pathological dependency on your free stuff.
The same way a normal parent would teach their child to become self sufficient instead of providing for them for their first 30 years of life and then tell them to figure out how to live after being used to not working.
I mean, it's pretty sad but there are examples of these things happening, and I definitely wouldn't say the parent is blameless for allowing this to happen.
Google on it's own is spending 1000x that in a single year $280B in '26 alone iirc
I could also argue with GP's comment that some US companies are driving towards commoditization as well. One released a model this week and another announced their series d today.
Not the person you are asking, but... In my view, the two most important technology clusters in human history are writing and industrialization. The former made culture (cognitive work) cumulative. The latter made physical work approximately free, with the result that we now have 5% instead of 95% working on producing food. Now we have a technology that could potentially be a culmination of these two: The industrialization of cognitive work/culture.
It isn't absurd to think this is one of the most consequential technologies ever. It probably doesn't make sense to rank technologies (which build on top of each other after all) by importance, but it's fairly clear this is a historic breakthrough.
Agriculture is the most important set of techniques and technologies in human history, it gave us the possibility of staying in a place.
There's also fire.
Writing was very important, I'm not so sure about industrialization, maybe as modern men we might think that way. But human life was perfectly liveable and conceivable without it. Even writing. Big civilizations like the Inca empire didn't really have writing (although they had the quipu system for record keeping).
If there's a civilization collapse, we will still be doing fire and agriculture. We probably won't be doing industrialization and computers.
Fire predates modern humans. I will concede agriculture as up there with the two I mentioned, but I think you underestimate how radical the shift of industrialisation was/is, and I think this is because history teaches us about highly atypical lives from the past. City dwellers, kings, etc...
In fact you could reasonably argue that life for most people didn't change that much from establishing agricultural settlements all the way up to industrialization. Most people spent most of their time procuring food and clothing/shelter using roughly similar methods.
So far it rates "quite important" and certainly not "paradigm shift".
I've just spent most of the day wading through vibed blogs and what not to glean info on ... setting up LLMs! The content on the web is already in a parlous state and rapidly getting nicely but rather the same presentation and mostly right but somewhat wrong, often in crucial parts.
On the bright side I did find an absolute belter of a vibed site and it was only mildly wrong but I was able to sort that bit out and to be fair I think it was written up by an expert with help from a LLM and had a genuine human mistak in it.
Anyway, I suggest you might like to look at the internet as a whole as a paradigm shift and reserve judgement on LLMs and this decade's version of AI.
They want to bankrupt the US AI industry and lock down the future development (and the substantial financial gains from more capable models). Essentially the China car manufacturer playbook, but for LLMs which are ~software, so much faster to execute.
To your last point, it can’t possibly be sustainable, so it reads to me more as a short term FUD attack on American dominance in this space. It might have cost them half a billion dollars to train this model, and they’re going to make nothing off it. How many more times can they afford to do that? It’s going to get more expensive to train AI going forward, not less.
I also have a suspicion that the benchmark numbers are not real.
Is this a classic 'commoditize my compliment' situation? They want to sell the hardware and infrastructure behind AI and make the software part not the value driver / moat?
I can see it. But also even two Chinese labs sinking 100s of millions USD into training isn't exactly commoditization. It's still a ton of effort with dubious payoff.