Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Everyone should slow down AI development except for me

xeiaso.net

Post reply on HN
Print view View original

421–430 of 459 posts

Re: Everyone should slow down AI development except for me

#421

Post by ipaddr »

BoiledCabbage wrote: post #213
ipaddr wrote: post #190

Earlier quoted context omitted.

"Opus delivered mediocre, so I did not let it finish" Mediocre means average / middle of the pack. It sounds like its doing exactly what you would expect nothing more. Why would you stop it? Why would you need exceptional?

> mediocre - of moderate or low quality, value, ability, or performance https://www.merriam-webster.com/dictionary/mediocre Clearly they were using the word to mean low quality. Why would you ask this odd question?

Core meaning: Barely adequate, average, or just acceptable.

Mediocre means of only ordinary or moderate quality—neither very good nor very bad, and often slightly disappointing

The quality is average but expectations of high quality are not met. He expected more but got what he asked for. We overuse top models because of this.

Re: Everyone should slow down AI development except for me

#422

Post by iamnothere »

metalcrow wrote: post #409
iamnothere wrote: post #357

Earlier quoted context omitted.

Nukes don’t get “worse” because of physics, not because of laws. You can’t make a pocket Tsar Bomba even if there were no rules against proliferation. Dial-a-yield nukes already provide a range of yields from tiny to massive, and this provides all the operational flexibility that is needed considering that nobody’s willing to actually launch them in anger. (And this reluctance again isn’t about laws, it’s about the m…

> Nukes don’t get “worse” because of physics, not because of laws. Yes, that's my point, that's why the metaphor is wrong

Sorry, I was tired and misread the intent.

Re: Everyone should slow down AI development except for me

#423

Post by cute_boi »

ianm218 wrote: post #383
Oarch wrote: post #339

Earlier quoted context omitted.

GPT-2 was called "too dangerous to release".

Is there a sense where they were right because GPT-2 directly led to the current situation where you can atleast objectively say that reasonable people are pretty worried about existential risks of AI .

Well you can't attribute it like this. If you go like this, Are you going to blame google and ban research on LLM?

Re: Everyone should slow down AI development except for me

#424

Post by lukewarm707 »

user43928 wrote: post #413
lukewarm707 wrote: post #391

Earlier quoted context omitted.

i can provide some sources. 5.6 sol ultrafast on cerebras is 750tps, open models readily exceed this. just by using a smaller model cerebras serves qwen 3.8 27b at 1850tps. or even larger models, mimo 2.5 pro was served for a while at 1000tps. and so on. [ https://inference-docs.cerebras.ai/models/choose-a-model ] the chinese ai companies have 10% of the total compute resources of the US ones. since the USA tries to…

The proper comparison would be GPT 5.6 Luna at 38 on the intelligence score and $0.18 per task vs $0.27 for DeepSeek Flash 4.1.

i was trying to make a point about efficiency of serving the model. the cost per task itself would not be enough to show that.

you could compare gpt 5.6 luna. if you did that the same way as before you would get a blended price of $0.17 for luna at AA 38. for baseten it would be $0.20 for v4.1 at AA 40.

assume roughly the same intelligence. on AA openai gets 117tps. baseten gets 284tps. so 18% more expensive but 142% more tps.

the fast mode is again double the cost, roughly same intelligence. so luna in that case would be 70% expensive. take the per task cost and it would still 9% more expensive.

so i think there is something to be said about the efficiency of the model.

Re: Everyone should slow down AI development except for me

#425

Post by the_optimist »

ben_w wrote: post #347
the_optimist wrote: post #328

Earlier quoted context omitted.

It is an utterly silly premise that we should appoint insiders with permanent control. Megalomania is not evidence of either a problem or a solution.

The evidence is the acts the AI has already done.

All of the things it has done is just what people are already doing. Every single one. It helps, but there are already people doing this stuff. And we already have defenses against the existing attacks. We might have to defend better, but the real problem is “Who is setting the moral compass over generations.” One one hand, that will inherently fall to future generations. On the other, we are laying the groundwork.

Imagine you use it to inflict trauma on your cruelest political enemy. Then next year they do that to you. That is war, and we already do it. But we don’t want people / governments in charge that are going to do this.

God knows we have governments and individuals doing this historically, and this is probably the greatest source of historic instability. It might be the single best argument for open models — a unified frontier where no single exploit is going to represent capture.

This is exactly the opposite of what Dario proposes.

Re: Everyone should slow down AI development except for me

#426

Post by bossyTeacher »

AaronAPU wrote: post #27

They are talking about slowing down the public facing AI development. Because then nation states can create a capabilities gap between them and the public. Why does nobody seem to be pointing out this obvious explanation? It explains why the “we need to race China” concern suddenly vanished in the discussion. The government can simply gag Sam, Dario, Musk on national security basis, getting them all behind the public…

> They are talking about slowing down the public facing AI development.

I assume there already existed advanced models frontier labs only made available for national security. We just weren't told. I am guessing the announcement is a way to avoid jeopardising their IPO, while reducing their workload.

Re: Everyone should slow down AI development except for me

#427

Post by Marsymars »

cmrdporcupine wrote: post #368
Marsymars wrote: post #91

Earlier quoted context omitted.

You can have an arbitrarily low opinion of Trump and still trust in democracy as a form of government over autocracy or oligarchy, which is what the AI labs have.

You don't speak for anybody outside of your borders, so please stop trying to. The majority of the world will refuse to be subject to American AI laws under the control of the US gov't, whether it pretends to be democratic or not. Or should.

> You don't speak for anybody outside of your borders, so please stop trying to.

FWIW, I'm not American and refuse to step foot within American borders while Trump is in power.

Re: Everyone should slow down AI development except for me

#428

Post by Marsymars »

int_19h wrote: post #289
Marsymars wrote: post #91

Earlier quoted context omitted.

You can have an arbitrarily low opinion of Trump and still trust in democracy as a form of government over autocracy or oligarchy, which is what the AI labs have.

You can have abstract trust in democracy over autocracy, but this is not an abstract question - these are the actual choices. Does it really matter if the person who will use AI to fuck us has a "democratic mandate" to do so? We're fucked either way.

It's less the mandate that matters, but the fact that they can be removed from power.

Re: Everyone should slow down AI development except for me

#429

Post by conception »

user43928 wrote: post #345
lukewarm707 wrote: post #338

Earlier quoted context omitted.

open models are ahead in speed. they complete tasks as fast as you choose to scale compute. they are more efficient and require less compute for the same thing. they are ahead in specialized tasks. they are ahead in areas closed models refuse to answer. they are ahead in emotional intelligence.

I doubt most of your claims. Maybe the guardrails and emotional intelligence is true. For speed and efficiency, you are most likely wrong. Speed is led by GPT-5.6 Sol on Cerebras Ultrafast at 750 t/s. Afaik you cannot serve a single DeepSeek Flash 4.1 stream at 750 t/s, plus the model is less intelligent as seen on newer benchmarks. I believe OpenAI and Anhropic are at the frontier of efficiency too. There were numer…

Inco serves DS 4.1 at 650t/s.

Re: Everyone should slow down AI development except for me

#430

Post by jurgenburgen »

rescbr wrote: post #398
user43928 wrote: post #343

Earlier quoted context omitted.

I always thought switching from a SOTA model to a dumber model after planning was a terrible idea. Mostly I heard this from people who I got the impression have little experience in developing greenfield software with agentic AI. Often the same people who talk about spec frameworks. I fundamentally disagree with the approach. I believe the ability to autonomously evaluate, test, and adjust during long horizon tasks i…

Well, it's the enterprise software house pipeline... The software architect writes the spec, hands it down to the implementation team, senior leads, junior devs or offshore teams codes it. I also disagree with the approach, this is cargo-culting the existing ways of working.

> this is cargo-culting the existing ways of working.

I think you mean the existing anti-patterns. They don’t call them ivory tower architects for nothing.

Post reply on HN