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

DeepSeek v4.1 Flash

twitter.com

Post reply on HN
Print view View original

511–520 of 543 posts

Re: DeepSeek v4.1 Flash

#511

Post by Onavo »

sroussey wrote: post #450
dgacmu wrote: post #396

Earlier quoted context omitted.

Bandwidth is really cheap in bulk. You can get a 100 gigabit internet connection for about $10,000 a month. If you were to somehow keep that saturated 24/7, you'd move about 30 petabytes in a month, so your per gigabit cost is only $0.0003. Realistically, if you had it 5% utilized, those million tokens would cost you about $0.0000062, which is pretty insignificant compared to what they charge you. (Assuming one byte…

People read the AWS rate card for bandwith and think has something to do with reality. Even though it is 1000x higher!

Most web devs have never heard of Colo unfortunately, they only know Vercel and AWS. Hurricane electric should sponsor more booths at colleges. If they give out more swag maybe the millenials and gen Zs would finally understand bandwidth pricing.

Re: DeepSeek v4.1 Flash

#512

Post by mrob »

idiotsecant wrote: post #481
mrob wrote: post #365

Earlier quoted context omitted.

Possibly. If I had some side-effect-free means to permanently prevent all sleep I'd take it without hesitation. But that's not relevant to the discussion, because it's not anything similar to what an LLM does. Your brain changes state even while sleeping.

In what way does your brain change state while sleeping that a model does not change state via constant fine-tune updates?

It's possible that an LLM is conscious during training, but there are no "constant fine-tune updates" during inference.

Re: DeepSeek v4.1 Flash

#513

Post by entropicdrifter »

largephoton wrote: post #425
randbyte wrote: post #418

Earlier quoted context omitted.

When Chinese tech report is tech report and US tech report is bible scripture.

there's probably fewer bibles in China so less source material to reference I suppose

They have their share of scripture, just not christian specifically

Re: DeepSeek v4.1 Flash

#514

Post by abecode »

Alpha3031 wrote: post #157
Alifatisk wrote: post #140

> New Causal Encoder–Decoder architecture: just 8B active parameters for input, 16B for output. Oh interesting, I can assume what the benefits is for including the Encoder, but whats the downside? I’m thinking GPT (which is decoder only) ruled out Encoder for a reason?

Enc-decs are usually harder to train at frontier scale. Not 100% sure what DeepSeek has done differently here initial read seems to be something related to layer reuse but I just skimmed things so far.

yes, that was surprising to me too. It would be a big deal if they switched to an encoder-decoder model like the original transformer. But I don't think that's what it's doing. One thing is the causal part, so in the original transformer, the encoder was bidirectional, but in this case it is not, so that's one difference. So I think it's an optimization for the prompt/prefill so that the attention is summarized into the output of the encoding layers, rather than all the layers. I just skimmed the paper too so if anyone else has insight, please correct me.

Re: DeepSeek v4.1 Flash

#515

Post by orbital-decay »

3371 wrote: post #503
orbital-decay wrote: post #401

Earlier quoted context omitted.

That's one of the reasons why you should never trust a single word from Anthropic and OpenAI (Sam Altman also blamed them back in the day of R1, in a pretty convenient moment). If you know anything about Claude, DeepSeek, jailbreaking, and distillation, you know the claims are clearly bullshit and the models are nothing alike, and forensic attempts agree, in fact there just was another one [1] [2]. Meanwhile, DeepSee…

This raise a question: why do open source models sometimes identify themself as Anthropic's models. I recall seeing some plausible theories in the past but I can't recall.

Name training is shallow and should never be relied upon. Claude sometimes identifies itself as Qwen or DeepSeek when asked in Chinese. I've seen it identify itself as GPT-3 (that version in particular) and Reddit Anti-Evil Operations team (Sonnet 3.6).

Re: DeepSeek v4.1 Flash

#516

Post by brokencode »

throwaway85825 wrote: post #502
brokencode wrote: post #412

Earlier quoted context omitted.

Yup, alignment doesn’t sound that interesting until you’re getting chased around by the Terminator.

Why are we so worried about supposed AI harms when algorithms already kill millions? It seems like its just a smokescreen to shield the extant harms.

Knives can already kill, so why worry about nukes?

Re: DeepSeek v4.1 Flash

#517

Post by throwaway85825 »

natrys wrote: post #208
TacticalCoder wrote: post #148

Earlier quoted context omitted.

> quant HFT is pretty decent mental exercise and it has given them “deep” brain muscles. that’s my take. It's quite crazy that it's Deepseek's background/original purpose. We already had very advanced stuff from the world of HFT, but now a frontier family of models from a private company that used to be (still is?) in HFT is plain bonkers. Is more known about them and the HFT background?

According to an old interview, apparently they were always interested in AI. But finance is just where they had their first success. > Many of High-Flyer's original team members worked on AI. Back then, we tried a lot of fields before getting our big break in finance, which is complex enough. AGI is probably one of the hardest things we can do next, so for us it was a question of how, not why. It's a very good interv…

In this case the finance model was used to parse lengthy, verbose and inscrutable yet very impactful chinese government pr statements and do sentiment analysis.

Re: DeepSeek v4.1 Flash

#518

Post by throwaway85825 »

brokencode wrote: post #516
throwaway85825 wrote: post #502

Earlier quoted context omitted.

Why are we so worried about supposed AI harms when algorithms already kill millions? It seems like its just a smokescreen to shield the extant harms.

Knives can already kill, so why worry about nukes?

AI doom is a potential harm unlike algorithms knives and nukes which are present harms.

Re: DeepSeek v4.1 Flash

#519

Post by agile-gift0262 »

lowbloodsugar wrote: post #485
walrus01 wrote: post #54

Looking at the huggingface page, the unsloth people haven't finished quantizing it yet, but I'm sure they're active on it right now. It'll be interesting to see how the capabilities and benchmark tests compare on system where it can fit in under 512GB of RAM with full context. In terms of coding and command line capabilities I'm also very interested to see a head-to-head of it vs. qwen 3.8-flash-next Q8 which is some…

3.8-flash-next fits on a single 6000 at Q4 if you offload the PLE. Crazy fast and still effective.

Sorry for the tangent, but how does Qwen3.8-flash-next compare to DeepSeek v4 Flash? I still haven't found the time to set it up, but I'm really happy with DeepSeek v4 Flash

Re: DeepSeek v4.1 Flash

#520

Post by ychnd »

1f60c wrote: post #207
kouteiheika wrote: post #13

It's so refreshing to see DeepSeek's tech report[1] full of juicy details; meanwhile, something like Fable's system card[2] is like 70% "safety", 10% "model welfare" to make sure little Claude isn't distressed, and 20% benchmark numbers. [1]: https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/... [2]: https://www.anthropic.com/claude-fable-5-1-mythos-5-1-system...

> welfare We should be paying attention to it just in case it ends up mattering enormously. It's cheap insurance. Aside from that, US labs' system cards have been pretty useless for a while—I think the last great one was the combined system card for Claude 4 Sonnet and Opus.

Enough with the basilisk already.
Post reply on HN