this post was submitted on 27 Jan 2025
136 points (93.0% liked)

Technology

61227 readers
5182 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each other!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed
  10. Accounts 7 days and younger will have their posts automatically removed.

Approved Bots


founded 2 years ago
MODERATORS
 

The milestone highlights how DeepSeek has left a deep impression on Silicon Valley, upending widely held views about U.S. primacy in AI and the effectiveness of Washington's export controls targeting China's advanced chip and AI capabilities.

Mirror: https://archive.is/2025.01.27-062326/https://www.reuters.com/technology/artificial-intelligence/chinese-ai-startup-deepseek-overtakes-chatgpt-apple-app-store-2025-01-27/

top 28 comments
sorted by: hot top controversial new old
[–] [email protected] 48 points 4 days ago* (last edited 4 days ago) (4 children)

has prompted U.S. tech executives to question the effectiveness of tech export controls.

It worked well against Russia and the Soviet Union, but it's been my claim for years now, that this won't work against China anymore. Because China is way more advanced now than the Soviet Union ever was by comparison for the time. And all it does is delay China while they build their own competing supply industries. And the end result will be that China will just dominate the backbone industries too.

The difference is in raw power, China has 10 times the people Russia has, and an economy that in real value more than matches USA now. In short China has all the advantages for long term developments. Except for not being ahead on chip manufacturing already.

China is already competing at top level in almost all key industries, and has absolutely unparalleled manufacturing capacity. but also technologically they are in the top of EV, Batteries, solar panels, and 44 other technology areas.

https://www.theguardian.com/world/2023/mar/02/china-leading-us-in-technology-race-in-all-but-a-few-fields-thinktank-finds

all of the world’s top 10 leading research institutions are based in China and are collectively generating nine times more high-impact research papers than the second-ranked country (most often the US).”

Trying to hold back China is simply not a viable option anymore.

[–] [email protected] 41 points 4 days ago (2 children)

It's not just people, it's also that China invests heavily in the building blocks of effective industries. They're serious about education and healthcare as public services, they invest heavily in transit. You can't expect life to bloom in a desert, but the US keeps acting like stripping away everything necessary for industries to prosper will somehow make them prosper harder.

Requisite disclaimer; The CCP are an appalling autocracy with a litany of crimes against humanity to their name. Just because they do some things well doesn't make them worthy of admiration. Yes, I'm looking at you, .ml

[–] [email protected] 14 points 3 days ago (1 children)

sadly "autocracy with litany of crimes against humanity" is rapidly approaching background noise when comparing nations

[–] [email protected] 3 points 3 days ago

At least they're doing some things right

[–] [email protected] 8 points 3 days ago (1 children)

Maybe not in a vacuum, but compared to pretty much any other government in the world right now, they are still very much worthy of admiration lol

[–] [email protected] 0 points 2 days ago

How much shoe polish do you go through monthly keeping a sheen on those oversized shoes?

There are plenty of national governments not engaged in genocide or imperialism. There are tons of governments better than China.

[–] [email protected] 21 points 4 days ago (2 children)

I think you can effectively compete with China but you would need a giant coalition of democratic countries that would move fast and stay focused.

With oligarchs completely taking over the US, this is unlikely to happen. But that's hardly the sole roadblock to such a coalition.

[–] [email protected] 8 points 3 days ago* (last edited 3 days ago)

I think you can effectively compete with China but you would need a giant coalition of democratic countries that would move fast and stay focused.

Exactly, like EU and USA working together, but then Americans go ahead and elect Trump!?
As it's going we were already losing, when China was behind. There is no way we can beat them when they are ahead, unless we dramatically up our game.

I've been cursing our (Denmark) government for cutting education and reducing taxes for 20 years. Instead of increasing education, and focus and invest in the sectors we are good at.
That's what made us rich in the past, it's like they think we don't need to invest anymore, because we are economically doing fine cutting taxes and undermining the quality society we built in the past.

[–] [email protected] 3 points 4 days ago

With oligarchs completely taking over the US, this is unlikely to happen.

It could still happen, it’s just that oligarchs will be the only ones to profit and it might require even more coercion than usual.

[–] [email protected] 5 points 4 days ago (1 children)

I had heard that DeepSeek only took less than a hundredth of the power to train it, so it wasn't them getting their hands on a bunch of new chips or anything.

[–] [email protected] 5 points 4 days ago* (last edited 4 days ago) (1 children)

I dont fundamentally disagree with you, but it should be pointed out that the USSR and its proxies in the Warsaw pact had about 0.4 billion people in 1980 to the PRC's 1 billion. So it wasnt anywhere near 10x the size. The important difference IMO is that china has opened up economically, whereas the USSR stayed closed, and are acting out Lenin's maxim of "the capitalists will sell us the rope to hang them".

[–] [email protected] 2 points 3 days ago (1 children)

I clearly write: "China has 10 times the people Russia has,"

So why are you arguing a strawman?

[–] [email protected] 0 points 3 days ago* (last edited 3 days ago) (1 children)

Because you used Russia and USSR interchangeably in your argument

It worked well against Russia and the Soviet Union, but it’s been my claim for years now, that this won’t work against China anymore. Because China is way more advanced now than the Soviet Union ever was by comparison for the time

...

China has 10 times the people Russia has

You first say what worked again Russia and the USSR (which is a bit weird, like writing new york and the USA) wont work against China, then say its because China has 10 times the pop of Russia.

[–] [email protected] 2 points 3 days ago* (last edited 3 days ago)

Because you used Russia and USSR interchangeably in your argument

No I did not, I very consciously made some comparisons to both, and some separately because USA used this tactic successfully against both Russia and the Soviet Union.

Soviet Union and Russia are NOT the same. But the tactic worked for half a century against them. And can still work against Russia, except if China goes all in on supporting Russia.

[–] [email protected] 33 points 4 days ago

How long before deepseek is banned by US gov for being too good—-wait sorry I mean a security risk?

[–] [email protected] 4 points 3 days ago (1 children)

So I'm still on the fence about the AI arms race in general. However, reading up on DeepSeek it feels like they built a model specifically to work well on the benchmarks.

I say this cause it's a Mixture of Experts approach, so only parts of the model are used at any given point. The drawback is generalization.

Additionally, it isn't a multimodal model and the only place I've seen real opportunity for workflows automation is using the multimodal models. I guess you could use a combination of models, but that's definitely a step back from the grand promise of these foundational models.

Overall, I'm just not sure if this is lay people getting caught up in hype or actually a significant change in the landscape.

[–] [email protected] 8 points 3 days ago (1 children)

they built a model specifically to work well on the benchmarks.

To be fair, I'm pretty sure that's what everyone is doing. If you're not measuring against something, there's no way to tell if you're doing anything at all.

[–] [email protected] 1 points 3 days ago

My point was a mixture of Experts model could suffer from generalization. Although in reading more I'm not sure if it's the newer R model that had the MoE element.