Free Republic
Browse · Search
General/Chat
Topics · Post Article

Skip to comments.

Meta (Facebook) creates four 'war rooms' to unravel how DeepSeek is outperforming rivals at lower costs
neowin.net ^ | Jan 28, 2025 08:08 EST | Sagar Naresh Bhavsar @@SNB3112 ·

Posted on 01/28/2025 10:26:31 AM PST by dennisw

DeepSeek AI has disrupted the AI landscape in the US. In just a few weeks after the launch of its AI model, DeepSeek overtook ChatGPT to become the number one free app on the App Store. Not only this, DeepSeek's rise in popularity sent shockwaves to the tech industry, leading to a $400 billion in market cap loss for NVIDIA in the US.

Recently, DeepSeek launched its Janus-Pro 7B, a groundbreaking image generation model that started making headlines, as it outperformed the likes of OpenAI's DALL-E, Stability AI's Stable Diffusion, and other image generation models in several benchmarks.

The popularity of DeepSeek has caught the attention of Meta, and to understand the success of this Chinese AI startup, Mark Zuckerberg's Meta has reportedly assembled four specialed teams, referred to as "war rooms," consisting of engineers to understand how a Chinese AI startup backed by High-Flyer Capital Management has managed to achieve performance on par with or exceeding that of top competitors like ChatGPT at a fraction of the cost.

Notably, DeepSeek gained popularity after it launched the R1 model, an AI chatbot that beat ChatGPT. The company claims that it invested less than $6 million to train its model, as compared to over $100 million invested by OpenAI to train ChatGPT. Meta's war rooms will be brainstorming to find ways how to address the potential threat posed by DeepSeek's breakthrough.

Two of the four war rooms will be dedicated to understanding how DeepSeek managed to cut costs in developing and running R1 models, with hopes of applying the same strategy to Meta's own AI model, Llama. Another team will be investigating the training data that DeepSeek used. The last team will be focussing on exploring ways to redesign Llama's architecture to compete with Chinese AI technology.

Although Meta did not comment on this development, a Meta spokesperson said in a statement to The Information that:

We regularly evaluate all competitive models in our development process and have done so since [the company’s] Gen Al [group] was formed. Llama has been foundational in establishing the ecosystem for open-source AI models and we couldn’t be more excited to extend this leadership with the upcoming release of Llama 4.

Meta is on high alert because Meta AI infrastructure director Mathew Oldham has told colleagues that DeepSeek’s newest model could outperform even the upcoming Llama AI, expected to launch in early 2025. Even OpenAI's CEO Sam Altman has responded to DeepSeek's rise and called it impressive. NVIDIA, which is one of the biggest sufferers of the sudden popularity of DeepSeek, also commended the Chinese AI and also highlighted how NVIDIA GPUs were used for DeepSeek's software.


TOPICS: Business/Economy; Computers/Internet; Gardening
KEYWORDS: china; deepseek; facebook; fakenews; hype; magnificent7; markzuckerberg; meta; propaganda; redchina; theskyisfalling

1 posted on 01/28/2025 10:26:31 AM PST by dennisw
[ Post Reply | Private Reply | View Replies]

To: dennisw

I would say it’s not a good idea to download that communist chinese crap. They will use it for spying and control


2 posted on 01/28/2025 10:28:30 AM PST by Strict9
[ Post Reply | Private Reply | To 1 | View Replies]

To: dennisw

Four war rooms! This looks like DEFCON 5 for scared shtyeless nerds at Meta -— owner of Facebook and Instagram


3 posted on 01/28/2025 10:28:46 AM PST by dennisw (DËMÔNràts - Truth is hate to people who hate truth.)
[ Post Reply | Private Reply | To 1 | View Replies]

To: dennisw

It’s competition, somebody will eventually exceed them.

Welcome to “Creative Destruction”.


4 posted on 01/28/2025 10:28:51 AM PST by dfwgator (Endut! Hoch Hech!)
[ Post Reply | Private Reply | To 1 | View Replies]

To: dennisw

But is DeepSeek “better”?

It won’t give answers about Tienanmen Square, so it has built-in blind spots and/or prejudices.

I thought the whole concern was that it was developed super cheap and super fast and used a remarkably small amount of energy.

And we “know” this because China said so.

I think DeepSeek is snakeoil.


5 posted on 01/28/2025 10:40:46 AM PST by ClearCase_guy
[ Post Reply | Private Reply | To 1 | View Replies]

To: dennisw

The CCP released the source code for this project. I would believe that this would allow a WESTERN company to use it to build the same quality of language model, without the Chinese filter.


6 posted on 01/28/2025 10:54:51 AM PST by FrankRizzo890
[ Post Reply | Private Reply | To 1 | View Replies]

To: dennisw
I think it is so indicative of people's fecklessness and gullibility that everyone just believed that the Chinese version of ChatGPT was so much better than the dozen other versions out there.

First of all, it would take a LONG time to assess just what it knew and what it didn't.

Second, it is programmed to evade or erase or conceal so much about Communism and the realities of world history, politics, economics, etc., that it is going to finish last straight out of the gate, anyway you cut it.

They just did this psyop on us for the hell of it, and it worked!

So amazing that geeks and stockholders are so dumb.

7 posted on 01/28/2025 10:56:22 AM PST by caddie
[ Post Reply | Private Reply | To 1 | View Replies]

To: FrankRizzo890

“””The CCP released the source code for this project. I would believe that this would allow a WESTERN company to use it to build the same quality of language model, without the Chinese filter.”””

Correct and we are. It’s open source:)

You still need GPU’s so I’m not sure why NVIDIA is tanking.

It’s faster and has better responses than Chat from what we have seen.


8 posted on 01/28/2025 10:57:35 AM PST by isthisnickcool (1218 - NEVER FORGET!)
[ Post Reply | Private Reply | To 6 | View Replies]

To: caddie

Yes, the outcome of the “investigation” will be most interesting. It might even lead to a better breakthrough.


9 posted on 01/28/2025 11:04:53 AM PST by SaxxonWoods (Black guy upon receiving a MAGA hat: "MURICA!")
[ Post Reply | Private Reply | To 7 | View Replies]

To: dennisw

They didn’t. They lied. They had thousands of “banned” Nvidia cards. It’s freaking obvious, as Musk and Wang (ScaleAI) have asserted.


10 posted on 01/28/2025 11:07:12 AM PST by montag813
[ Post Reply | Private Reply | To 1 | View Replies]

To: ClearCase_guy
I think DeepSeek is snakeoil.

It's also slow as shit, and was two weeks ago, when I tested with it before the rush.

The only decent way to run their R1 model is on your home computer, as our own resident AI genius bobcat62 is doing.

11 posted on 01/28/2025 11:09:47 AM PST by montag813
[ Post Reply | Private Reply | To 5 | View Replies]

To: dennisw

There is the possibility that High Flyer, the Chinese hedge fund that owns Deep Seek, shorted the US market and Nvidia before it released its “breakthrough” AI system. Indeed, an impressive $6 million investment cost is a minor investment to disrupt the speculative market. No one can immediately disprove the algorithm so the tech market crashes. Billions made on the short.

Feom a person I know.


12 posted on 01/28/2025 11:22:04 AM PST by HYPOCRACY (Democracy is dead. Long live the Republic!)
[ Post Reply | Private Reply | To 1 | View Replies]

To: HYPOCRACY
Indeed, an impressive $6 million investment cost is a minor investment to disrupt the speculative market.

Imagine how much a mere $6 million in NVDA put options bought at Friday's close would have been worth at 3:59 yesterday.

13 posted on 01/28/2025 11:40:36 AM PST by montag813
[ Post Reply | Private Reply | To 12 | View Replies]

To: montag813

Billion$


14 posted on 01/28/2025 12:17:59 PM PST by HYPOCRACY (Democracy is dead. Long live the Republic!)
[ Post Reply | Private Reply | To 13 | View Replies]

To: dennisw

I had read last night that China lied about the cost.


15 posted on 01/28/2025 12:33:16 PM PST by roving (Deplorable MAGA Garbage )
[ Post Reply | Private Reply | To 1 | View Replies]

To: dennisw

Copy the terms of deep seek and ask your favorite AI for a summary.
You will quickly see the true purpose of deep seek


16 posted on 01/28/2025 2:06:17 PM PST by NoLibZone (Scary that a party can "run" a candidate that doesn't feel any need to campaign.)
[ Post Reply | Private Reply | To 1 | View Replies]

To: isthisnickcool

Exactly what I was thinking. “The Nvidia chips run the existing bigger/slower products at this speed. The newer faster, smaller code should be even FASTER on them”.


17 posted on 01/28/2025 3:57:15 PM PST by FrankRizzo890
[ Post Reply | Private Reply | To 8 | View Replies]

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.

Free Republic
Browse · Search
General/Chat
Topics · Post Article

FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson