Can't believe that Dario, Elon, Sam and I all agree on this.

“and I”

China isn’t gonna slow down. This means we better speed up. I’m sorry but I don’t want china to win this one.

If the IPO went ahead and was a disaster, everyone attached to these companies is in Deep Shit. Now there’s a great excuse to defer. Seems a little incentivized to agree with them.

theo's thumbnail day by day 📈 which model you used to made this btw?

I think they’re playing a game you’re not privy to.

A new government agency staffed by EA activists with authority over all our labs? For safety? And you buy that line why?

Dude stop pretending your opinion on this is important

Why did Elon go from saying it's a psyop to then start agreeing with Dario?

Because he now runs a real lab and understands the consequences

@MONKE2525E didn't he call it a psyop less than a week ago though?

@theo That's what I'm thinking....

day 2 of asking @theo to add message queueing to T3 Code

What are your thoughts on China not slowing down? Isn't it fsr worse to let them pull ahead, than for us to slow down and pace ourselves? Is this going to lead to a ban on foreign open source models? This is not going to bode well for us.

It's very scary how the general trend is people letting agents run autonomously (without human review), but we are seeing more and more cases where agents act without permission or instructions from humans

People don't give a fuck about your opinion

This seems pretty serious

"Dario, Elon, Sam and I"

why can't we have both? Safe AI without pacing it?

It’s to late!

Isn't this is suspicious?

consensus among people who all benefit from looking cautious isn't really consensus

Did you ask China?

aren’t you the guy who’s been constantly complaining about Gemini, the only model that was truly paced?

They're all boot lickers wanting to be protected by the State from their competition -which isn't the Chinese, it's us, the people, who'll compete with them in the USE of these models - they want to relegate the older models to the public, while they use the best.

Sam said he agrees but did not say he follow Dario so he is releasing grok 4.8 that too this month only. Back to back 2 models.

I dont think they agree. They are forced to build a wall so that deepseek won't kill their business

"And I" lmao

when theo goes calm i actually slow down and listen

THAT IS WHY WE CALL U PSY OP THEO

No doubt that this tech can be very dangerous, but we need to be careful with how regulation is imposed. I feel that these frontier labs will exploit the need for gaurdrails to design regulations that will leave them with too much power. Centralized AI can be very dangerous asw

Four people agreeing is a weaker signal than it looks when three of them would pay the same compliance bill and the fourth would not. The test is whether the position costs the incumbents more than a new entrant. Which side of that does this one land?

Glad to see there's still someone else on the internet with basic critical thinking skills.

👇