this post was submitted on 28 Jan 2025
98 points (100.0% liked)

Technology

37943 readers
198 users here now

A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.

Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.

Subcommunities on Beehaw:

This community's icon was made by Aaron Schneider, under the CC-BY-NC-SA 4.0 license.

founded 3 years ago
MODERATORS
alyaza@beehaw.org
TheRtRevKaiser@beehaw.org
gyrfalcon@beehaw.org
rs5th@beehaw.org
coldredlight@beehaw.org
Los@beehaw.org
SemioticStandard@beehaw.org
TheRtRevKaiser@kbin.social
remington@beehaw.org
98
‘Sputnik moment’: $1tn wiped off US stocks after Chinese firm unveils AI chatbot (www.theguardian.com)
submitted 1 week ago by schizoidman@lemm.ee to c/technology@beehaw.org
47 comments fedilink hide all child comments
 

cross-posted from: https://lemmy.ml/post/25282200

all 48 comments
sorted by: hot top controversial new old
[–] tangentism@beehaw.org 9 points 6 days ago (2 children)

Last year Dario Amodei, the co-founder of leading AI firm Anthropic, put the current cost of training advanced models at between $100m and $1bn.

So his estimate is so vague as to be fucking useless.

How people haven't seen through this emperors new clothes nonsense & grasped that these are the bullshitting self appointed druids of the neo religion.

That bubble they keep mentioning is where they're all getting high off their own farts

[–] Kichae@lemmy.ca 3 points 6 days ago

Business people will do almost anything to eliminate wages

[–] Midnitte@beehaw.org 2 points 6 days ago

Tbf, that could also just be describing the variability of models.

I'd imagine there's a difference in cost for training Dall-E 3 and GPT-3.5 for example.

[–] arsCynic@beehaw.org 3 points 6 days ago* (last edited 6 days ago)

ELI5 request on lemmy: "You see, dear grandchildren, your grandfather used to have an apple orchard. The fruits were so sweet and nutritious that every town citizen wanted a taste because they thought it was the only possible orchard in the world. Therefore the citizens gave a lot of money to your grandfather because the citizens thought the orchard would give them more apples in return, more than the worth of the money they gave. Little did they know the world was vastly larger than our ever more arid US wasteland. Suddenly an oriental orchard was discovered which was surprisingly cheaper to plant, maintain, and produced more apples. This meant a significant potential loss of money for the inhabitants of the town called Idiocracy. Therefore, many people asked their money back by selling their imaginary not-yet-grown apples to people who think the orchard will still be worth more in the future.

This is called investing, or to those who are honest with themselves: participating in a multi-level marketing pyramid scheme. You see, children, it can make a lot of money, but it destroys the soul and our habitat at the same time, which goes unnoticed by all these people with advanced degrees. So think again when you hear someone speak with fancy words and untamed confidence. Many a times their reasoning falls below the threshold of dog poop. But that’s a story for another time. Sweet dreams."

[–] astrsk@fedia.io 4 points 6 days ago

Sweet, cheaper stocks this week.

[–] fuzzy_feeling@programming.dev 38 points 1 week ago* (last edited 1 week ago) (2 children)

ahh... the sound of a bursting bubble.

- plop -

[–] MagicShel@lemmy.zip 25 points 1 week ago (1 children)

Cheaper AI isn't the pop we want. We want companies to stop trying to use AI for every god damn thing it is terrible at. We don't want cheaper AI that's just going to be baked into more stuff.

[–] LoamImprovement@beehaw.org 15 points 1 week ago (1 children)

I mean, I do want cheaper GenAI in the sense that I want people to see that it's dollar store crap that's not worth the electricity to run the servers to make it and give it up like they did the fucking Juicero and every other smart appliance a couple years ago. God forbid I hold my breath and people wise up and understand that these people are all grifters looking to tape a horn to a horse and sell their "unicorn" to FAANG or whatever the equivalent is these days, I can't be assed to rewrite the new poob acronym.

[–] MagicShel@lemmy.zip 3 points 6 days ago

AI isn't the enemy, though, and we aren't the ones being grifted—that would be companies who think they can make tons of money replacing people with AI. It's a reasonable useful tool/fun toy/interesting curiosity in certain circumstances. And for an end user it doesn't use any more power than a video game. But it's a tool for craftsmen and folks who understand the limitations, not a replacement for workers. And it sure as hell isn't a production feature. Anyone looking to make money baking AI into consumer products is an idiot and going to lose their hat.

[–] artificialfish@programming.dev 2 points 1 week ago (1 children)

Nah, o1 has been out how long? They are already on o3 in the office.

It’s completely normal a year later for someone to copy their work and publish it.

It probably cost them less because they probably just distilled o1 XD. Or might have gotten insider knowledge (but honestly how hard could CoT fine tuning possibly be?)

[–] leisesprecher@feddit.org 16 points 1 week ago (2 children)

Deepseek showed that actually putting thought into the architecture achieves much more than just throwing more hardware at the problem.

This means a) there will be much less demand for hardware, since much more could be run locally on regular consumer devices. And b) the export restrictions don't really work and instead force China to create actually better models.

That means, a lot of the investments into the thousands of AI companies are in jeopardy.

[–] artificialfish@programming.dev 1 points 4 days ago

I think “just writing better code” is a lot harder than you think. You actually have to do research first you know? Our universities and companies do research too. But I guarantee using R1 techniques on more compute would follow the scaling law too. It’s not either or.

[–] bobs_monkey@lemm.ee 9 points 1 week ago (1 children)

Realistically, the CCP is probably throwing a lot of money at developers to get something good going and available, and US companies are whining about how it's not fair. The fact of the matter is that a solid product is available for much cheaper, and US companies are now screaming foul. Guess what, a superior product made of good code (people) beats out just throwing money at hardware, who'd've gone an thunk it.

[–] leisesprecher@feddit.org 12 points 1 week ago (2 children)

What I find really fascinating here is that obviously openAI, Meta, etc. seem to be structurally incapable of actually innovating at this point.

I mean, reducing training costs by literally an order of magnitude just by writing better software is astonishing and shows how complacent the large corporations have gotten.

[–] artificialfish@programming.dev 1 points 4 days ago* (last edited 4 days ago)

Meta? The one that released Llama 3.3? The one that actually publishes its work? What are you talking about?

Why is it so hard to believe that deepseek is just yet another amazing paper in a long line of research done by everyone. Just because it’s Chinese? Everyone will adapt to this amazing innovation and then stagnate and throw compute at it until the next one. That’s how research works.

Not to mention China has billions of more people to establish a research community…

[–] algorithmae@lemmy.sdf.org 1 points 6 days ago

You can write off hardware purchases, paying for skilled devs is like pulling teeth

[–] haerrii@feddit.org 31 points 1 week ago (3 children)

One thing to be kept in mind, though:

https://youtu.be/4RQkl6qcwPY

verified this myself with the 1.1b model

[–] artificialfish@programming.dev 18 points 1 week ago (2 children)
[–] haerrii@feddit.org 15 points 1 week ago (1 children)
  • open weights, as the training dataset is not open/available afaik. But yum :D
[–] artificialfish@programming.dev 12 points 1 week ago* (last edited 1 week ago)

Well the uncensored fine tuning dataset is oss

https://huggingface.co/datasets/Guilherme34/uncensor

[–] Karkitoo@lemmy.ml 3 points 1 week ago

Much better.

Thank you!

[–] Allero@lemmy.today 11 points 1 week ago

continuous development and progress

Lol

[–] zante@slrpnk.net 11 points 1 week ago (2 children)

This is pathetic Sinophobia.

[–] zygo_histo_morpheus@programming.dev 9 points 6 days ago (1 children)

Please explain how this is Sinophobic.

[–] zante@slrpnk.net 6 points 6 days ago (3 children)

Innovation from other countries is not subjected to similar scrutiny

[–] zygo_histo_morpheus@programming.dev 13 points 6 days ago (1 children)

People have scrutinized what chatgpt for example is allowed and not allowed to say by its programmers. I think the difference here is that there is lower hanging fruit to grab because the Chinese state has a different relationship to censorship than a lot of other states.

I also associate Sinophobia with being prejudiced against Chinese people or Chinese culture, however being critical or skeptical of the Chinese state is actually perfectly reasonable. I'm also very critical of the US state and this isn't because I'm "americaphobic" or some nonsense.

[–] gbuttersnaps@programming.dev 2 points 6 days ago

Totally agree. Saying that "any criticism of the Chinese government is sinophobic" is the same as saying that "any criticism of Israel's government is antisemitic."

[–] thelucky8@beehaw.org 7 points 6 days ago

@zante@slrpnk.net

Innovation from other countries is not subjected to similar scrutiny

This is simply not true.

[–] GreyEyedGhost@lemmy.ca 1 points 6 days ago

Welcome to the internet, you must be new. Keep scrolling through new here and you should see some pretty common jokes about the falliability of AI of various flavors. Criticism of the weighting on training models can be found with just slightly more effort.

[–] DdCno1@beehaw.org 18 points 1 week ago (2 children)

So tell me, what happened at that square in 1989?

[–] remotelove@lemmy.ca 14 points 1 week ago

According to Baidu and most of .ml, absolutely nothing. It was a perfectly normal day of getting emulsified by tanks. There are no unhappy people in China, and they have the CCTV recordings to prove it!

[–] zante@slrpnk.net 3 points 1 week ago (1 children)

Point to it on a map.

[–] DdCno1@beehaw.org 19 points 1 week ago (1 children)

Don't distract. I knew where exactly in the Chinese capital this square is when I was six years old. Your accusation that I must be some ignorant American who can't find anything on a world map is wrong on both accounts.

Once again: What happened there? I want an honest answer from you.

[–] zante@slrpnk.net 4 points 1 week ago (1 children)

Don’t distract ?

We’re talking about developments in AI tech, and you want to make it about Tiananmen Square .

[–] thelucky8@beehaw.org 13 points 1 week ago (2 children)

@zante@slrpnk.net

We’re talking about developments in AI tech, and you want to make it about Tiananmen Square .

Please don't use whataboutism to distract from the core issue here. DeepSeek follows the Chinese government's rules, meaning it has biases intentionally built-in to spread misinformation and propaganda by the Chinese Communist Party. You may read my comment and the source in in this thread.

[–] zante@slrpnk.net 3 points 6 days ago (1 children)

As you and I recently discovered in another thread, you don’t know whataboutism is, and entire account is dedicated to sinophobic comments .

As are your other accounts.

[–] thelucky8@beehaw.org 4 points 6 days ago (1 children)

@zante@slrpnk.net

There's nothing sinophobic here. The linked article on China's AI policy cites an official Chinese source. It comes directly from the Chinese government.

[–] DdCno1@beehaw.org 5 points 6 days ago (1 children)

Haven't you heard? The truth is sinophobic. Obligatory:

The party told you to reject the evidence of your eyes and ears. It was their final, most essential command.

[–] thelucky8@beehaw.org 2 points 6 days ago

Ah, yeah, I forgot:

We are different from all the oligarchies of the past, in that we know what we are doing. All the others, even those who resembled ourselves, were- cowards and hypocrites. The German Nazis and the Russian Communists came very close to us in their methods, but they never had the courage to recognize their own motives. They pretended, perhaps they even believed, that they had seized power unwillingly and for a limited time, and that just round the corner there lay a paradise where human beings would be free and equal. We are not like that. We know that no one ever seizes power with the intention of relinquishing it. Power is not a means, it is an end. One does not establish a dictatorship in order to safeguard a revolution; one makes the revolution in order to establish the dictatorship. The object of persecution is persecution. The object of torture is torture. The object of power is power.

[–] RusAD@lemm.ee 3 points 6 days ago (2 children)

Breaking news: AI has the same biases as their creators! More at 11

[–] DdCno1@beehaw.org 6 points 6 days ago (1 children)

Saying that "everyone is doing it" is called hypernormalization and a common propaganda technique used by autocratic regimes like Russia and China. It's meant to instill feelings of hopelessness and indifference. It is also usually extremely dishonest, since it relies on false equivalencies (like in this case) and/or comparing current state crimes with past state crimes of others.

[–] RusAD@lemm.ee 2 points 6 days ago (1 children)

DeepSeek is designed to follow Chinese state propaganda. ChatGPT is designed to follow American oligarchy propaganda. Each side pushes their own biases. Where is the false equivalence?

Also I wasn't meaning to instill hopelessness or indifference. On the contrary, I tried to point out that you should be equally as mad at OpenAI/Google/the rest of the western AI companies as you are at the DeepSeek. The fact that people are only mad at China and not at the megacorporations in the west does, imo, indicate their biases.

[–] thelucky8@beehaw.org 3 points 6 days ago (1 children)

@RusAD@lemm.ee

I tried to point out that you should be equally as mad at OpenAI/Google/the rest of the western AI companies as you are at the DeepSeek

This exactly is whataboutism. The issue here is DeepSeek. There are many articles here where people are "equally as mad" about OpenAI?Google and the rest of Western AI, but in these threads, you never read, "But China, ...". This whataboutism appears to work only in one direction.

[–] RusAD@lemm.ee 1 points 6 days ago

I don't think this really is whataboutism. If I responded with, for instance, "But ChatGPT is way less energy efficient and wastes tons of water", then it would be. But I believe I'm comparing apples to apples. And I'm not singling out specifically ChatGPT or Gemini or what have you, I'm saying that every LLM suffers from the same problem, whether it's from USA, China, Russia, Argentina, Mongolia or Uganda. ChatGPT and Gemini are just big and well-known examples. Also DeepSeek is quite new, at least it's popularity is, so we'll have to wait and see whether these "but DeepSeek..." comments will appear.

[–] thelucky8@beehaw.org 3 points 6 days ago* (last edited 6 days ago) (1 children)

@RusAD@lemm.ee

In addition to @DdCno1@beehaw.org's comment, this is not 'only' about biases we all have. It is about intentionally built-in propaganda supporting Chinese state narratives. As I wrote already in this thread, the Chinese government outlined its plan regarding AI in its so-called "AI Capacity Building and Inclusiveness Plan". It reads:

[Chinese] Government rhetoric draws a direct line between AI exports and existing initiatives to expand China’s influence overseas, such as Xi Jinping’s signature Belt and Road Initiative (BRI) and Global Development Initiative (GDI). In this case, the more influence China has over AI overseas, the more it can dictate the technology’s development in other countries [...]

[According to the Chinese government] AI must not be used to interfere in another country’s internal affairs — language that the PRC has invoked for as long as it has existed, both to bring nations of the global south on board in China’s ongoing efforts to seize Taiwan and to deflect international criticism of its human rights record [...]

The whole article makes a good read.

[Edit typo.]

[–] RusAD@lemm.ee 2 points 6 days ago (1 children)

Yeah, the whole article reads like "China says good things, but actually, they are just bad things in disguise! Chinese AI is cheaper and in some cases better, but actually its all a master plan for world domination! Remember, guys, China bad!!!"

Of course the software created in China would intentionally try to follow Chinese laws. Just as American software follows the laws of the US. However, China made their models open-source, so people can tinker with them and (possibly) make a fork without the censorship, and run it locally without any data being sent who knows where. While ChatGPT, Gemini and all the rest are closed-source, and you have zero control over them, so you can't even hope to circumvent the propaganda that's built into those models.

But yeah, Tiananmen square 1989, China bad, let's all stop thinking about anything

[–] DdCno1@beehaw.org 4 points 6 days ago

let’s all stop thinking about anything

You're way ahead of us on that front.