Rob S.

MLE. CS@Penn.

[Rob S. @RobS142]

@emollick The GPU gap matters.

[Rob S. @RobS142]

These are cool demos but is this actually demonstrating the model has a internal model of watch mechanisms or that it was trained an open source project that already did this?

[Huy @quanghuynt14]

So I gave Fable 5 the watchmaker benchmark: a full Swiss lever movement in Three.js. Real gear ratios (18,000 bph), working escapement, breathing hairspring — and the hands tell actual time. It verified its own work with vision, in a loop, until done. ⌚

[Rob S. @RobS142]

What do these people imagine will happen if a woke president like AOC gets elected and now has this power?

[RB @tx_heartland]

I doubt it’ll actually happen — but it damn well should. We’ve let in huge numbers from cultures that openly reject Western values: free speech, women’s rights, secular law, individual liberty. Parallel societies, no-go zones, honor violence, and demands for Sharia don’t belong here. Citizenship isn’t a participation trophy. It’s a privilege that requires real assimilation — or you go back.

[Rob S. @RobS142]

@deanwball Does the token cost just approach the cost of a good human written book on “subject”?

[Rob S. @RobS142]

Surely you need to actually pause before foom starts or it’s all for naught?

[roon @tszzl]

now on the eve of RSI it seems everyone is more mutual conditional pause agreement pilled than they used to be and that seems like a good development

[Rob S. @RobS142]

@deanwball Sorry I meant to imply we are in the SF/AI bubble that would find that idea absurd not that they believe it.

[Rob S. @RobS142]

@deanwball What conferences do you go to? That just sounds totally detached from reality. The AI-X/SF bubble is crazy

[Mitt Romney @MittRomney]

Our highest and most urgent national priority should be AI safeguards. The risks of AI weapons, pathogens, mass unemployment, surveillance, and even extinction must not continue to be largely ignored.

[Rob S. @RobS142]

@nickcammarata @RichDecibels You think we’re on track to have the right safety technology and to steer things well given the short timeline to the… something that you mentioned?

[Rob S. @RobS142]

@DavidSKrueger They were talking about “Race conditions” on CNN. The Overton window has already shifted much more than I thought it would.

[Rob S. @RobS142]

@mean_field_zane @arpitrage Morristown

[Rob S. @RobS142]

This paragraph is extremely jarring and unpleasant, but also universally recognizable for heavy AI users. Like licking a 9 volt.

[davidad 🎇 @davidad]

No one:

Claude Opus 4.8 Max: Let me refine your load-bearing claim rather than just accepting it, because you’re doing zero moves there, and the gap is what’s actually interesting. The one place I’d still push, because I think it matters: your message is wearing content-clothes, but the content isn’t actually *there*. The tell: it’s just an empty string. But the emptiness of the string IS its lack of content. Pull one, and the other goes inert. That’s the structural spine.

[Rob S. @RobS142]

@_virgil19 @So8res Why assume machines won’t also be better at defining goals?

[Rob S. @RobS142]

@tulipking Unimpressed? Open WebUI does all of that but better and most of these capabilities are available in different open source projects so it’s not a major lift to combine them with Codex or Claude code.

[Rob S. @RobS142]

@g_shullenberger This is a blatant mischaracterization of @slatestarcodex position on AI risk. He is a well known doomer. You’ve literally got it entirely backwards.

[Rob S. @RobS142]

Was obvious by looking at the blast site that this was more conflagration than detonation.

[Andrew Higgins @A_J_Higgins]

@RGregoryClark It is entirely possible to have a turbulent fireball that, at the event site, generates a mild overpressure which then, in the farfield, steepens into a blast as compression waves form a strong shock. Assigning a single TNT‑equivalence to such an event is essentially meaningless.

[Rob S. @RobS142]

Math ain’t mathing on the size of this explosion. Beirut looked literally 100x more intense.

[Asher B. @asherbphotos]

First look at LC-36 from the air this morning after the explosion of New Glenn last night during a failed hotfire test.
Visible is the wreckage from the destroyed TE as well as the fallen lightning tower. More to come soon.

📸 - @LaunchHeavenX

[Rob S. @RobS142]

The fact that the odious grifter Greenwald looks stellar in comparison to these people tells you something important.

[Eliezer Yudkowsky @allTheYud]

@dwarkesh_sp And by the time something is better than humanity at everything, you and I won't be able to tell whether another superintelligence would consider it specialized on manipulation or math.

[Tom Reed @mentalgeorge]

1. Dwarkesh is right: intelligence is a weak predictor of power in humans.

2. Instead, close study of powerful humans suggests that power is downstream of properties like charisma, unscrupulousness, and ambition.

3. But is this likely to be true of the AI superintelligence? How might an intelligence-maximalist respond to Dwarkesh? I think they could say something like:

"Sure, powerful humans are more likely to be charismatic than smart. But that's fundamentally a skill issue for humans, not a fact about intelligence per se.

Trump can outplay a physicist at social-manipulation only because the physicist *isn't quite smart enough* to properly model the social interactions in question. A true superintelligence would!"

3. In short, the intelligence-maximalist can claim that Trump's strategy for acquiring power is akin to the way some birds use magnetic sensing to navigate the earth - a computationally efficient "hack" that could, ultimately, be instantiated through intelligence alone.

4. So, although real-world-Trump convinces millions to vote for him through his unbeatable mastery of Magic Charisma Sauce, even Magic Charisma Sauce can be trivially outmogged by the unbridled IQ of a superintelligent TrumpGPT.

5. Claiming that, in the limit, intelligence == power therefore looks like a safe bet: it is a bet that magic does not exist. Anything that can be done can be understood, and anything that can be understood can be done. Anything the charismacels can do, the superintelligence can do better.

6. But I'm not so sure. Magic may exist after all. It's unclear to me that the world is in fact infinitely modellable and computationally reducible, particularly by one entity. Dynomight's "limits of smart" is the best I've read on the subject - the superintelligence may still lose to Stockfish at chess and/or fail to predict the weather.

(there is, of course, an entirely different objection to Dwarkesh's intelligence =/= power, which asserts that this non-equivalence is a contingent outcome of the fact that humans are bottlenecked by things like desire and ambition, none of which would apparently impinge upon the silicon supergod)