Live data from Hacker News

Anthropic's Safety Superpower

stratechery.com

201–206 of 206 posts

Re: Anthropic's Safety Superpower

#201

Earlier quoted context omitted.

Do you have more info about this? I can't tell if you're being misled by the unfortunate "Mixture of Experts" terminology (which don't work the way you're describing), or alluding to something different. Or, maybe I'm wrong, but my understanding is: MoE is just an architecture to keep the activated weights smaller per token. The experts get routed basically token-by-token, and the "experts" themselves don't have a se…

No, this is an agent-level thing, not a feature of the model (ish, unsure for Fable). You talk to a smart, heavy model to build a plan composed of smaller steps. Then you have the heavy model spin up smaller, cheaper LLMs to actually implement the tasks. The heavy model is basically read-only in that mode. It can read files, execute tests, etc, but it can’t write code. It just tracks what needs to be done, offloads t…

The AWS Kiro (https://kiro.dev) spec-driven coding harness operates this way in Auto mode which offers the base token rate.

Manually-specifying Sonnet or Opus is a multiplier on the base token rate; specifying Qwen fractions it. Left to its own, it presumably uses the heavier models to create the plan and orchestrate the work; the bite-sized task definitions are delegated to smaller models.

Re: Anthropic's Safety Superpower

#202
post #135

Earlier quoted context omitted.

If you’re implying that the government is in on it and is doing this stuff intentionally in order to boost Anthropic, that’s ridiculous.

Yeah that would be ridiculous, this administration pumping stocks.

Come on, these conspiracy theories make no sense, Anthropic isn’t even a publicly traded company.

Re: Anthropic's Safety Superpower

#203
The thing about all this is that there's not a chance in hell that Anthropic can retain control against the wishes of the USG. Like, the USG has the guns, simple as. They're not going to tolerate a private company controlling this technology.

Anthropic likely knows this and is merely performing a song and dance. They're auditioning to be THE frontier AI lab.

Re: Anthropic's Safety Superpower

#204

Earlier quoted context omitted.

> Anthropic and Claude remains very popular among the people who use LLMs Only because someone else is paying the bills. I use Claude Opus at work because my employer pays for the tokens and encourages me to do it. At home, I use DeepSeek Flash. It's not as good, but it's maybe 0.7 quality for 0.001 cost.

What's the speed on DeepSeek Flash? And what provider?

ds4: prefill: 150.18 t/s, generation: 48.47 t/s together with https://github.com/antirez/ds4 + RTX Pro 6000, local provider :) I guess you could rent one too, get similar speed. https://gist.github.com/embedding-shapes/eceefa1b2699dca448f...

The weights I've tried is the "DeepSeek-V4-Flash-IQ2XXS-w2Q2K-AProjQ8-SExpQ8-OutQ8-chat-v2-imatrix" ones which fits just about within 96GB VRAM. With some tuning, I've managed to get it to get up to ~60 t/s, I'm sure there is other things to do there too :)

Re: Anthropic's Safety Superpower

#205

Earlier quoted context omitted.

I don’t think so. I’d want more than just your word on it.

Thanks for pointing that out! I'm not in the US and I guess it's not illegal in China (given that Deepseek was more than happy to do it). That does raise an interesting question, what kind of laws should LLMs (attempt to) follow? It's easy enough to spoof the country in the system prompt. I wonder how ChatGPT would respond if I told it I was located in a developing country without any piracy laws.

> ... what kind of laws should LLMs (attempt to) follow?

I think being able to align your text generators with a specific set of laws, speech codes, tonality, etc. is very useful. However, all of this should be a user choice.

Your spell checker does not stop checking your spelling just because you are describing a murder scene either.

Safety slop and boilerplating should be configurable as well.

Re: Anthropic's Safety Superpower

#206

Earlier quoted context omitted.

> I don’t even think the subscriptions are losing money, I suspect they’re around break even, maybe small loses whats the basis for this thought

What's the basis for the opposite?

ask the person that proposed that. not me
Post reply on HN