Miscellaneous News

T

taxiya

Brigadier
Registered Member
Eventine said:
A wise man once said, allow nothing get in the way of a great story; and Deep Seek is a great story. But end of the day, we need to sit down and talk about facts.

Deep Seek is not a super hero team of 200 IQ geniuses. They didn't "do the impossible." The reason they did what Big Tech. did with a fraction of the costs is because China does everything the US does at a fraction of the costs. It's the same reason BYD can sell an EV brand new for $11,500, but a "domestically built EV" cost $50,000+ in the US. It's why DJI can sell a drone for $500, but Skydio sells it at $2,000. It's why a medical procedure in China cost $500, but the same medical procedure in the US costs $5000.

It's the currency, stupid. Chinese labor, energy, infrastructure, food, medicine, everything seems absurdly cheap when converted to dollars. And that's what Deep Seek's costs were evaluated in - dollars. Nobody ever talks about the amount of energy they used to train the model and how that compares to the energy spent by other AI companies. No, we always talk about dollars, dollars, dollars - and that's the same story across every industry - China does it cheap. AI is just the latest to be hit by it.

Yes, there's a bit more to it than that. Institutional bloat - which doesn't just affect US Big Tech., but also Chinese Big Tech. companies like Baidu, Tencent, Alibaba, etc. - that also has tremendous costs. It's why Byte Dance is investing billions of dollars into the effort, while Deep Seek with its relatively small team with minimal middle management, could get away with a lot less. Fact is, if you have thirty Directors, Senior Directors, and Vice Presidents making millions of dollars a year each, it's going to cost you billions on compensation alone especially with all the usual corporate alignment, meetings, delays, etc.

But that's not just an US problem. That's a big corporation problem. It's why nobody in Chinese Big Tech. is claiming it's cheap. Because it isn't, when your corporation becomes that large.

Deep Seek's advantage is the advantage enjoyed by every Chinese company in competition vs. the US - cost. Their disadvantage will also be the same - it'll be hard for Deep Seek to keep hold of its talent in the coming days, as recruiters start reaching out to poach them for ten times the salary. But maybe that's okay. As long as China maintains a talent advantage vs. the US, and there's no shortage of young Chinese AI graduates hungry for fame and fortune, companies like Deep Seek can continue to surprise the world about Chinese costs in dollars.
Click to expand...
I can summerize your key point that is China's AI advantage is only due to lower input cost such as lower salary and engergy cost. This is the very convenient reason to explain every advantage that China has. It also gives a confort to the western mind "China is good only because cheap labour". However very often that is copium.

If you have ever worked in computer related industry, you should know there are many other things that lead to lower overall cost. For one, if you have a badly designed algorithm and coding, you could have made 10 to 100 times loops that can be done in just one loop. That is 10 to 100 times time and energy cost. An extra day spent on the drawing board will save hundreds of hours every year on every machine deployed.

I work in a world class company selling software product in the US which you live with every minute. I know what we did and there are lots of these sub optimized loops. These problems come from short cut in system design due to high time pressure from customer, badly trained engineers, and rush/hush R&D methodology aka "agile and lean" which leads to the egineers making new feature without overall idea of why the baseline was done in the way it was done, so people just patch things up to fit their new demand even though the system is better to be changed in broader scale.

In short, Deepseek may simply has been developed by smarter and more experience people who is willing to spend some extra time to think things over than rushing to the market for "who comes first in the shortest time" which is what US is "good" at. In a little bit longer time, Deepseek got a more robust, optimized and coherent body while others got frankenstein monsters.
 
Last edited:
I

iewgnem

Senior Member
Registered Member
Eventine said:
A wise man once said, allow nothing get in the way of a great story; and Deep Seek is a great story. But end of the day, we need to sit down and talk about facts.

Deep Seek is not a super hero team of 200 IQ geniuses. They didn't "do the impossible." The reason they did what Big Tech. did with a fraction of the costs is because China does everything the US does at a fraction of the costs. It's the same reason BYD can sell an EV brand new for $11,500, but a "domestically built EV" cost $50,000+ in the US. It's why DJI can sell a drone for $500, but Skydio sells it at $2,000. It's why a medical procedure in China cost $500, but the same medical procedure in the US costs $5000.

It's the currency, stupid. Chinese labor, energy, infrastructure, food, medicine, everything seems absurdly cheap when converted to dollars. And that's what Deep Seek's costs were evaluated in - dollars. Nobody ever talks about the amount of energy they used to train the model and how that compares to the energy spent by other AI companies. No, we always talk about dollars, dollars, dollars - and that's the same story across every industry - China does it cheap. AI is just the latest to be hit by it.

Yes, there's a bit more to it than that. Institutional bloat - which doesn't just affect US Big Tech., but also Chinese Big Tech. companies like Baidu, Tencent, Alibaba, etc. - that also has tremendous costs. It's why Byte Dance is investing billions of dollars into the effort, while Deep Seek with its relatively small team with minimal middle management, could get away with a lot less. Fact is, if you have thirty Directors, Senior Directors, and Vice Presidents making millions of dollars a year each, it's going to cost you billions on compensation alone especially with all the usual corporate alignment, meetings, delays, etc.

But that's not just an US problem. That's a big corporation problem. It's why nobody in Chinese Big Tech. is claiming it's cheap. Because it isn't, when your corporation becomes that large.

Deep Seek's advantage is the advantage enjoyed by every Chinese company in competition vs. the US - cost. Their disadvantage will also be the same - it'll be hard for Deep Seek to keep hold of its talent in the coming days, as recruiters start reaching out to poach them for ten times the salary. But maybe that's okay. As long as China maintains a talent advantage vs. the US, and there's no shortage of young Chinese AI graduates hungry for fame and fortune, companies like Deep Seek can continue to surprise the world about Chinese costs in dollars.
Click to expand...
Nope, you can give OpenAI billions of $ (and people have) and they still wouldn't be able to run O1 on a Raspberry Pi
Millennia of meritocratic tradition isn't something you can buy, China did not become the world's oldest and most successful civilization by simply being less corrupt.
Being institutionally and culturally corrupt to the bone is certainly not helping Americans, but at end of the day even if they were not corrupt they'll still not be China's equal, they need to be around for at least another 2000 years to begin to even think about it.
 
I

iewgnem

Senior Member
Registered Member
Eventine said:
The fore runners on the LLM front are now Open AI (Microsoft), Google, Meta, Deep Seek, Anthropic (Amazon), Alibaba, and Byte Dance.

There's not enough market for all of these companies. The Western companies are mainly held up by scatter shot investment, where investors are hedging their bets by throwing money at all of them - similar to what they were doing during the internet bubble and bit coin bubble - so that they don't miss out on the winner. Open AI is currently the government favorite (Sam in the White House etc.), but Google probably has better long term fundamentals. There's also Musk's x AI, which could get a boost from Musk being in the government, but they've yet to show they're at the frontier.

If I had to bet, I would say that Google is most likely the eventual winner in the West, though Open AI is a close contender. The main reason I favor Google is because, despite both companies having great research teams, Google has the benefit of dominating search despite Microsoft's best attempts, and having that reliable, synergistic revenue stream outside of just selling chat bots makes their project much more sustainable. Nonetheless, Microsoft may come to dominate the LLM coding space, simply because they have existing leverage there with Visual Studio and GitHub. I don't much like Meta and Anthropic's chances - the chat bot business model sucks.

The situation is different in China. There is no "Chinese Google" because, we should be honest, Baidu sucks. The chances of them winning the LLM race are slim, despite the search advantage; and in any case, being the leading search engine in China is not the same, because the Chinese internet is fundamentally different from the English internet in that the former is much more diffused between various media formats, while the latter is arguably still concentrated in the text-based World Wide Web.

Productivity enhancement is going to be much more important, as such, in the Chinese market. Companies will favor the model that give them the most bang for the buck. LLM companies are currently competing on price, and in that domain, companies that have deep pockets and which can sustain low or even negative margins for a long time, has the advantage. Deep Seek has the benefit of being backed by their own hedge fund, but Byte Dance and Alibaba both also have their own revenue streams outside of AI. The question in this scenario is more around integrations - who will be able to capture the most third party partners and build the best ecosystem? So far it's not clear, but I think we will get much better clarity in 2025, specifically through answering questions like: 1) is Deep Seek going to beat the rest to o3, 2) will the next Qwen release be successful or a bust, and 3) how fast can Byte Dance leverage multi-modal capabilities and stand out from the rest?

2025 will prove to be an exciting year, and I hold onto my previous argument, which is that China doesn't have to win the AI race, it just needs to not lose it. Being a fast follower is fine - and more profitable - as long as you're close enough.
Click to expand...
I don't think you've fully comprehended what it means to be 100x-ed
The only fore-runners right now are Deepseek and ByteDance (Doubao)
Eventually winner in the west will be whoever hosts Deepseek.
100x means even if you convince some companies to use OpenAI, they will be destroyed by companies who uses open source.
 
proelite

proelite

Junior Member
9dashline said:
https://twitter.com/i/web/status/1883209016261874010


Deepseek is pulling a rednote guys
Click to expand...

In order to maintain an open internet of free commie influence we need to ban all Chinese models and apps, and any models and apps with Chinese influence. I propose we silo our internet from the rest of the corrupted world and make sure there are no pro-China and anti-Israel sentiments. This is the only way to maintain a free and open internet as envisioned and invented by America 50 years ago.
 
AssassinsMace

AssassinsMace

Lieutenant General
Love how the masturbators to the US are upset over the cost of Chinese AI. Does that make any difference to the fact that everyone agrees that China’s model works? That’s the illusion they created for themselves hence why the problem they face and no one else. Yes they created a bubble because they were lying so they can exploit and make as much money as they can. From neutral eyes they were trying perpetrate the biggest crime in the history of the world. Is that China's fault? China didn't help maintain the illusion. Want to be as petty as their logic? "OpenAI" is actually "ClosedAI" since it's not open-source and because their AI doesn't point that out, it fails as an AI.
 
You must log in or register to reply here.
Top