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

DeepSeek v4

api-docs.deepseek.com

Post reply on HN
Print view View original

961–970 of 1001 posts

Re: DeepSeek v4

#961

Post by eleventen »

ben_w wrote: post #230
elefanten wrote: post #208

Earlier quoted context omitted.

It’s not remotely hypothetical you’d have to be living under a rock to believe that. And the fusion with a one-party state government that doesn’t tolerate huge swathes of thoughtspace being freely discussed is completely streamlined, not mediated by any guardrails or accountability. This “no harm to me” meme about a foreign totalitarian government (with plenty of incentive to run influence ops on foreigners) hooveri…

As a non-American, everything you wrote other than "one party" applies to the current US regime. Relatively speaking , DeepSeek is less untrustworthy than Grok. When I try ChatGPT on current events from the White House it interprets them as strange hypotheticals rather than news, which is probably more a problem with DC than with GPT, but whatever.

> When I try ChatGPT on current events from the White House it interprets them as strange hypotheticals rather than new

Any specific examples?

Re: DeepSeek v4

#962

Post by eleventen »

regularfry wrote: post #577
amunozo wrote: post #275

For those who rely on open source models but don't want to stop using frontier models, how do you manage it? Do you pay any of the Chinese subscription plans? Do you pay the API directly? After GPT 5.5 release, however good it is, I am a bit tired of this price hiking and reduced quota every week. I am now unemployed and cannot afford more expensive plans for the moment.

I've been on Kimi K2.5 on openrouter for a couple of months for anything I can't run locally. Really is dirt cheap for how good it is. Haven't assessed K2.6 yet but the price is higher so it needs to be more efficient, not just more capable. But more broadly: openrouter solves the problem of making a broad range of models available with a single payment endpoint, so you can just switch around as much as you like.

How do you find the token speed of open router with kimi?

I have tasks that used to take ~3-5min with Sonnet 4.6. With OpenRouter Kimi, the same task takes 10+ min. It's also just obviously slower in opencode sessions. The results are good, and I love the lower cost, but the speed can be frustrating.

Re: DeepSeek v4

#963

Post by hodgehog11 »

wohoef wrote: post #951
hodgehog11 wrote: post #461

Earlier quoted context omitted.

Very cool to hear there is agreement with (probably quite challenging?) coding problems as well. Just ran a couple of them through GPT 5.5, but this is a single attempt, so take any of this with a grain of salt. I'm on the Plus tier with memory off so each chat should have no memory of any other attempt (same goes for other models too). It seems to be getting more of the impressive insights that Gemini got and doing…

Do you an idea of how well these models perform on set theory problems or more niche fields in mathematics? So the model would have to both understand a paper that’s not in its training data, and use this to write proofs.

This is all fairly niche stuff I'm trying it on (well, the first three problems anyway), so yes, it needs me to give it several papers that are not in its training data and use them to write proofs. I would expect my experiences to transfer to set theory problems as well.

Re: DeepSeek v4

#964

Post by SV_BubbleTime »

arrty88 wrote: post #694
fblp wrote: post #13

There's something heartwarming about the developer docs being released before the flashy press release.

now that we can use AI to write the docs , test the docs , proof read the docs , it really isn't that much of a feat right?

If those docs were written by Deepseek, it’s also a pretty positive review of the model.

Re: DeepSeek v4

#965

Post by geon »

stronglikedan wrote: post #748
ifwinterco wrote: post #278

Earlier quoted context omitted.

As a Brit I'm here for it to be honest, I'm tired of America with everything that's going on. China is not perfect but a bit of competition is healthy and needed

> I'm tired of America with everything that's going on. Yeah, me too. All that pesky saving the world stuff that we do on the regular is so exhausting sometimes.

Who has been saved? The US has been doing much more harm than good.

Re: DeepSeek v4

#966

Post by jug »

simonjgreen wrote: post #950
segmondy wrote: post #644

Earlier quoted context omitted.

They have had the best math models for about a year most folks just didn't know about it. You can't find inference on APIs, but I run these at home, this is also the advantage of open models. https://huggingface.co/deepseek-ai/DeepSeek-Math-V2 https://huggingface.co/deepseek-ai/DeepSeek-Prover-V2-671B

You are of course specifically referring to the math optimised models, not the chat ones folks would generally encounter. Not that I’m trying to contradict you, your point is super valid and I agree with you! But I’m supplementing to help anyone following along who may make choices. This is when it happened for anyone interested: https://binaryverseai.com/deepseek-math-v2-benchmarks-review...

Shouldn't one use e.g a Wolfram Alpha MCP endpoint for math in AI? From what I've seen on even premium non-quantized models, I would never ever trust the innate ability of a LLM to calculate.

Re: DeepSeek v4

#967

Post by oofbey »

orbital-decay wrote: post #243

>we implement end-to-end, bitwise batch-invariant, and deterministic kernels with minimal performance overhead Pretty cool, I think they're the first to guarantee determinism with the fixed seed or at the temperature 0. Google came close but never guaranteed it AFAIK. DeepSeek show their roots - it may not strictly be a SotA model, but there's a ton of low-level optimizations nobody else pays attention to.

Nobody does it because it’s expensive. If you remove the requirement for perfect reproducibility you open the door to lots of optimizations. Most people prefer faster cheaper results over perfect reproducibility. When the model is intrinsically statistical the value of perfect reproducibility is … limited.

Re: DeepSeek v4

#969

Post by orbital-decay »

oofbey wrote: post #967
orbital-decay wrote: post #243

>we implement end-to-end, bitwise batch-invariant, and deterministic kernels with minimal performance overhead Pretty cool, I think they're the first to guarantee determinism with the fixed seed or at the temperature 0. Google came close but never guaranteed it AFAIK. DeepSeek show their roots - it may not strictly be a SotA model, but there's a ton of low-level optimizations nobody else pays attention to.

Nobody does it because it’s expensive. If you remove the requirement for perfect reproducibility you open the door to lots of optimizations. Most people prefer faster cheaper results over perfect reproducibility. When the model is intrinsically statistical the value of perfect reproducibility is … limited.

Yeah, of course. Making it cheap/compatible with heavy batching is exactly what they did, that's what I mean. ("with minimal performance overhead")

Re: DeepSeek v4

#970

Post by nwienert »

awongh wrote: post #650
arcticfox wrote: post #632

Earlier quoted context omitted.

Referring to the Dwarkesh interview clearly. Jensen came across as incredibly defensive and intentionally close-minded, shows that even billionaires suffer from "a man can't understand something if his paycheck depends on him not understanding it." Your assertion is silly: did Tesla selling electric cars into China stop them from delivering their own industry? They were going to develop their domestic industry regard…

I thought Jensen’s comparison to Huawei’s cell phone hardware infra (towers and networking) to be an interesting comparison- that shutting them out of a market was one of the causes of their current position in the market. It made them more dominant in the end.

No counterfactual there either though.
Post reply on HN