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

DeepSeek R2 launch stalled as CEO balks at progress

reuters.com

Post reply on HN
Print view View original

61–70 of 186 posts

Re: DeepSeek R2 launch stalled as CEO balks at progress

#61

Post by Thlom »

throwaway290 wrote: post #40
spaceman_2020 wrote: post #19

Earlier quoted context omitted.

I'm not American. Ever since I've been old enough to understand the world, the only country constantly at war everywhere is America. An all-powerful American AI is scarier to me than an open source Chinese one

As Russian I only recently started to understand that russian government was at wars for a lot of its existence from USSR times: https://en.wikipedia.org/wiki/List_of_wars_involving_Russia#... . Many invasions and wars in places Russia should have no business in. Most of them not publicized in the country. Unlike US it was not spreading liberal values of individual freedom and against violent dictatorships, actually…

The US is not at perpetual war to spread "liberal values".

Re: DeepSeek R2 launch stalled as CEO balks at progress

#63

Post by dachworker »

jamesblonde wrote: post #14
qwertox wrote: post #9

"We had difficulties accessing OpenAI, our data provider." /s

Rumour was that DeepSeek used the outputs of the thinking steps in OpenAI's reasoning model (o1 at the time) to traing DeepSeek's Large Reasoning Model R1.

Maybe they also do that, but I work with a class of problems* that no other model has managed to crack, except for R1 and that is still the case today.

Remember that DeepSeek is the offshoot of a hedge fund that was already using machine learning extensively, so they probably have troves of high quality datasets and source code repos to throw at it. Plus, they might have higher quality data for the Chinese side of the internet.

* Of course I won't detail my class of problems else my benchmark would quickly stop being useful. I'll just say that it is a task at the undergraduate level of CS, that requires quite a bit of deductive reasoning.

Re: DeepSeek R2 launch stalled as CEO balks at progress

#64

Post by WiSaGaN »

jamesblonde wrote: post #14
qwertox wrote: post #9

"We had difficulties accessing OpenAI, our data provider." /s

Rumour was that DeepSeek used the outputs of the thinking steps in OpenAI's reasoning model (o1 at the time) to traing DeepSeek's Large Reasoning Model R1.

Deepseek published thinking trace before OpenAI did, not after.

Re: DeepSeek R2 launch stalled as CEO balks at progress

#65

Post by orbital-decay »

astar1 wrote: post #45
ozgune wrote: post #17

Earlier quoted context omitted.

I think that's unlikely. DeepSeek-R1 0528 performs almost as well as o3 in AI quality benchmarks. So, either OpenAI didn't restrict access, DeepSeek wasn't using OpenAI's output, or using OpenAI's output doesn't have a material impact in DeepSeek's performance. https://artificialanalysis.ai/?models=gpt-4-1%2Co4-mini%2Co3...

almost as well as o3? kind of like gemini 2.5? I dug deeper and surprise surprise: https://techcrunch.com/2025/06/03/deepseek-may-have-used-goo... I am not at all surprised, the CCP views AI race as absolutely critical for their own survival...

Not everything that's written is worth reading, let alone drawing conclusions from. That benchmark shows different trees each time the author runs it, which should tell you something about it. It also stacks grok-3-beta together with gpt-4.5-preview in the GPT family, making the former appear to be trained on the latter. This doesn't make sense if you check the release dates. And previously it classified gpt-4.5-preview to be in a completely different branch than 4o (which does make some sense but now it's different).

EQBench, another "slop benchmark" from the same author, is equally dubious, as is most of his work, e.g. antislop sampler which is trying to solve an NLP task in a programmatic manner.

Re: DeepSeek R2 launch stalled as CEO balks at progress

#66

Post by sschueller »

teruakohatu wrote: post #3

The title of the article is "DeepSeek R2 launch stalled as CEO balks at progress" but the body of the article says launch stalled because there is a lack of GPU capacity due to export restrictions, not because a lack of progress. The body does not even mention the word "progress". I can't imagine demand would be greater for R2 than for R1 unless it was a major leap ahead. Maybe R2 is going to be a larger/less perform…

> lack of GPU capacity due to export restrictions

Human progress that benefits everyone being stalled by the few and powerful who want to keep their moats. Sad world we live in.

Re: DeepSeek R2 launch stalled as CEO balks at progress

#67

Post by layer8 »

Davidzheng wrote: post #44
layer8 wrote: post #29

Earlier quoted context omitted.

Are you saying that large-model capabilities would make a substantial difference in a military conflict within the next five years? Because we aren’t seeing any signs of that in, say, the Ukraine war.

small scale drones are in use in that conflict. On device AI would be a game-changer no?

It’s not impossible, but also highly nontrivial. Apart from the actual AI implementation, power supply might be a challenge. And there is a multitude of anti-drone technology being continuously developed. Already today, an autonomous drone would have to deal with RF jamming and GPS jamming, which means it’s easily defeated unless it has the ability to navigate purely visually. Drones also tend to be limited to good weather conditions and daytime.

Re: DeepSeek R2 launch stalled as CEO balks at progress

#68

Post by rescbr »

throwaway290 wrote: post #40
spaceman_2020 wrote: post #19

Earlier quoted context omitted.

I'm not American. Ever since I've been old enough to understand the world, the only country constantly at war everywhere is America. An all-powerful American AI is scarier to me than an open source Chinese one

As Russian I only recently started to understand that russian government was at wars for a lot of its existence from USSR times: https://en.wikipedia.org/wiki/List_of_wars_involving_Russia#... . Many invasions and wars in places Russia should have no business in. Most of them not publicized in the country. Unlike US it was not spreading liberal values of individual freedom and against violent dictatorships, actually…

> against violent dictatorships

Then look up Latin America’s history, where the US actively worked to install and support such violent dictatorships.

Some under the guise of protecting countries from the threat of communism - like Brazil, Argentina and Chile, and some explicitly to protect US company’s interests - like in Guatemala

Re: DeepSeek R2 launch stalled as CEO balks at progress

#69

Post by Davidzheng »

archon1410 wrote: post #48
roenxi wrote: post #32

Earlier quoted context omitted.

Presumably there is a CEO statement somewhere. If DeepSeek said May, but it is almost July, that would call for some comment from them. Although I'd like to know the source for the "this is because of chip sanctions" angle. SMIC is claiming they can manufacture at 5nm and a large number of chips at 7nm can get get the same amount of compute of anything Nvidia produces. It wouldn't be market-leading competitive but de…

> If DeepSeek said May It is pretty strange that DeepSeek didn't say May anywhere, that was also a Reuters report based on "three people familiar with the company".[1] DeepSeek itself did not respond and did not make any claims about the timeline, ever. [1]: https://www.reuters.com/technology/artificial-intelligence/d...

Actually I think one of the researchers at Deepseek did say on Twitter but I think that tweet has since been deleted.

Re: DeepSeek R2 launch stalled as CEO balks at progress

#70

Post by tazjin »

diggan wrote: post #26
energy123 wrote: post #20

Earlier quoted context omitted.

My experience with R1-0528 for python code generation was awful. But I was using a context length of 100k tokens, so that might be why. It scores decently in the lmarena code leaderboard, where context length is short.

Would love to see the system/user prompts involved, if possible. Personally I get it to write the same code I'd produce, which obviously I think is OK code, but seems other's experience differs a lot from my own so curious to understand why. I've iterated a lot on my system prompt so could be as easy as that.

Do you use the DeepSeek hosted R1, or a custom one?

The published model has a note strongly recommending that you should not use system prompts at all, and that all instructions should be sent as user messages, so I'm just curious about whether you use system prompts and what your experience with them is.

Maybe the hosted service rewrites them into user ones transparently ...

Post reply on HN