I'm reminded of export restrictions on 40-bit encryption 30 years ago. It will pass when chips get cheaper and things become less one-sided.
Hormuz moment for Mythos
Isn't this exactly what Dario wanted?
it took a long time for this to actually go out.
for what id expect to be an in memory switch, 2-3h is a while
Well, it was great while it lasted - I had fable build me a bunch of stuff this week that opus was just screwing up too much and could never finish. Good thing there are plenty of choices now even if US gov fucks up US AI.
I feel like this is part of the “negotiation” between the US Gov (Trump) and Anthropic and other labs for an equity stake. “If we have a seat at the table, and a confederate engineering team embedded, then we can ensure you won’t let nefarious actors use the models.” Either that or a temporary gate on Anthropic to benefit other labs (think maybe SpaceX started trading today).
.
Is it just me or is this really bad for their ipo, insane valuations and data centers that are being built r n? This might end the hype…
Interestingly, I am yet to lose access.
Probably Anthropic just needs to commit to handing over some shares to the Trump presidential family after their IPO and this will be solved.
This is just cost of doing business in corrupt Soviet vessel states like the USA.
HN is mostly bitter cynics that think they understand AI safety better than the people that actually work on it.
As a non American, I am simply mad!
American companies have, and continue to, gather data for free from across the globe and, despite our willingness to subscribe, we non Americans will be restricted from latest tech.
This is a big middle finger to me, and my gut reaction is to take my subscription to Mistral and not believe a dime of statements from Americans-- people, companies and government.
Biggest "Meh!" moment of my life so far...
Stupid government run by warmongers
Well, they also reset the quota
Qwen is all you need
My main read on this is that European governments will more aggressively invest in regional AI labs.
For Nvidia chips you could've deluded yourself that the US is just anti-china, that position is harder to argue for right now.
Finally, they face consequences for their IPO pump fear mongering rhetoric
This is the best marketing Anthropic could have hoped for. People crave what they can't have.
$5-10T of the US economy already operates under ITAR or EAR export restrictions, there is nothing novel about this order.
Aerospace, defense, semiconductors, telecom, advanced manufacturing equipment, etc
SpaceX entire operation is under ITAR, because even though their rockets are not weapons, rockets are treated as weapons for export purposes
Anthropic is big hall monitor energy. Clownish behavior and exaggerations constantly. A holier than thou attitude derived from rationalists at LessWrong.
It's really sad. I've been using Fable for the days that it was out and it was, and is probably the greatest model I've ever used. It's doing sane choices on its own, it's fully autonomous work, has done enormous, has brought enormous additional value to my work. I could hand off the development of an Android app fully to the model. It worked for hours, came back, and the app was just perfect. It's sad to see a government of incompetent people, of war criminals, and liars to stop a commercial model while they fully rely and embrace the capitalist idea within their own selfishness. I really hope Anthropic can win this war, and I really hope that this government will soon be history and that all of these guys that are responsible for bringing madness over the world will have a fair end.
US Government does bizarre, erratic thing which will likely be walked back shortly. Spectators nonplussed. Film at 11.
Europe really needs to get some useful sovereign capability and right quick.
That's one way to pop the bubble. Where are all of those billions going to go now, if not to make better models?
So eventually, you will have a massive string of data centers working to full capacity and whose only client will the US government?
So the US government was able to shut down that upgraded version of that slot machine in Anthropic's casino because of how powerful it is?
There is something called the Streisand effect and they are about to unintentionally get a bunch of more token gamblers into their casino.
We'll see if this backfires hard, but then again constant doomsaying will get yourself under scrutiny and self exclusion (due to the 30+ day retention clause) and this is exactly what Anthropic wants for free marketing.
If I read that right, the "jailbreak" is to ask the model to fix the codebase and then it exposes the flaws? That sounds like a gap that is nearly impossible to fix while retaining high capability. Like you want it to be able to fix your codebase...
I feel like this is more of a marketing campaign for Anthropic than anything.
USA_RESIDENT=true is back lol
You can't do that. A lot of engineers got fired because of Fable 5
I'm not going to downplay it. I've been coding since the 80s and using these models since 2023. 10 minutes after using fable I told my colleague this is a new era. It is the difference between sonnet and opus. I didn't think this was possible.
It was over before this announcement. After a couple of days, even though the model was set to fable, it felt like opus. We are back to sticks and stones.
trump hates dario. simple as that.
Their entire marketing strategy has been unwarranted fearmongering. This is completely unsurprising.
Did Opus 4.8 just get a lot dumber because of this? My sessions are making so many dumb mistakes it wouldn't make before...
Wow this is wild...but I guess it makes sense now why they had such an overly sensitive on Fable usage before. Perhaps they were already in a back-and-forth with the Trump admin about the Fable/Mythos release and what safeguards are needed.
I mean it should be clear to everyone that this is just the consequence of Sam Altman's lobbying activities.
it wasn't even that good...
Just in case you need evidence for the need of AI/LLM sovereignty.
This might be the biggest favor to anthropics valuation that Trump could have done
WH really don't like Anthropic. That is what this is about. The people who warned about the risks of using American cloudservices have become right. We should at least in EU see a ban on Azure, GCP, Cloudflare and so on.
Backfired...
And the chickens come home to roost. That’s what you get for your theatrics around Mythos!
In my head: The conversation politicians are having with other AI CEOs!
"How dare you release this model to poor people? This belongs only with the ultra-rich!"
They can say whatever they want... but I just have this gut feeling that this is part of it.
This is another historical moment in AI.
a fable for the ages:
pigs get fat, hogs get slaughtered
Pff. China has a massive opportunity to reign supreme on frontier models, and I feel like they are (commercial) Mythos/Fable capable within months, maybe even OSS as well.
Anthropic seem f’ed, not f’ed all at the same time. All this will come to light, but the hunch from the model reviews is that it’s all PR. Their model isn’t even close to all that for the govt to shit their pants over.
welcome back 90s crypto wars
> If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.
But what about the pelicans ?