Artificial Intelligence thread

[Overbom]

https://twitter.com/i/web/status/1886652308911530217

Didn't know Claude is getting so bad. 100% pivot into rent seeking.
The projection, literally what they warned Chinese AI would do and why they'll fall behind.

In the comments:
Can't even ask how to kill bugs in dirt or build an underground bunker without triggering a safety filter.

There is space for Anthropic in the market. Where customers directly interface with business AI in form of chatbots, emails or other sort of direct communication, I could easily see many businesses using Anthropic to minimize legal liability.


It's all about productization. If there is demand, there will be supply as well

 

[tokenanalyst]

Please, Log in or Register to view URLs content!

 

[Hyper]

Claude is finished... Opus was supposed to be released back last October. They have produced nothing. All their models trained couldnt beat R1 so its worthless, they havent put anything new out simply because they cant

Plus they rate limit paid users like crazy, they cant even keep up with inferencing even though they host on aws

They are done for


Also, before Deepseek R1, Amazon came out with its own Nova series, and Nvidia with their own (cant recall the name) and even LG did a LLM....

Now all I hear is Deepseek....

We know what has happened. Amazon pressured Anthropic to use Tranium chips to train their models. These chips don't have proper documentation, no proper sdk, bad interconnects. They will have to return to nvidia if they wish to train models properly.

 

[tokenanalyst]

Please, Log in or Register to view URLs content!

 

[tphuang]

https://twitter.com/i/web/status/1887185865664831839

Lenovo AI space Xiaotian has now integrated DeepSeek for coding, reasoning, multi-modal processing, speech processing and more

 

[supercat]

Zhejiang University, where DeepSeek's Liang Wenfeng came from, is ranked #1 in AI in the world.

https://twitter.com/i/web/status/1887249911764127949

Please, Log in or Register to view URLs content!

 

[tamsen_ikard]

Zhejiang University, where DeepSeek's Liang Wenfeng came from, is ranked #1 in AI in the world.

https://twitter.com/i/web/status/1887249911764127949

Please, Log in or Register to view URLs content!

Tsinghua is number 1. Its given 2 places with a slight variation of the name

 

[luminary]

Claude is finished... Opus was supposed to be released back last October. They have produced nothing. All their models trained couldnt beat R1 so its worthless, they havent put anything new out simply because they cant

Plus they rate limit paid users like crazy, they cant even keep up with inferencing even though they host on aws

They are done for


Also, before Deepseek R1, Amazon came out with its own Nova series, and Nvidia with their own (cant recall the name) and even LG did a LLM....

Now all I hear is Deepseek....

I warned y'all about Anthropic.


On the other hand, online among the non-techs, it seems like US DeepSeek propaganda has successfully split the populace. AI hasn't become ubiquitous yet, so it's still easy to ignore the significantly better Chinese model, at least until all the US small businesses start using it.


Jig is up! My early investor friends into Anthropic also jumping ship ASAP

Please, Log in or Register to view URLs content!

AI researcher Schulman had joined Anthropic in August

 

[GulfLander]

"ChatGPT is currently down for users worldwide. According to website performance tracking tool – Downdetector, thousands of users are impacted by this global outage. “Increased errors for ChatGPT, Sora, and API,” read a statement from the company. If further added “We are continuing to investigate this issue.” OpenAI has not yet provided any details regarding the cause or the extent of the outage." Times of india feb 6, 2025

Please, Log in or Register to view URLs content!

 

[luminary]

"DeepSeek surges to No. 1 downloaded app in Korea"

Please, Log in or Register to view URLs content!

Private sector widespread ban.

Please, Log in or Register to view URLs content!

Government agencies and major companies in South Korea are rushing to block access to Chinese artificial intelligence service DeepSeek, amid concerns about user data collection.

The move comes after the Ministry of the Interior and Safety issued a warning on Tuesday, advising central government agencies and 17 provincial governments to be cautious when using AI-based generative services like DeepSeek. Following this advisory, an increasing number of ministries have begun blocking access.

The Foreign, Trade and Defense ministries imposed access restrictions on DeepSeek later Wednesday, while the Finance Ministry is set to follow suit.

With the Environment Ministry joining the ban Thursday, restrictions have expanded beyond security and economic bodies to social bodies as well.

On the business front, IT giant Kakao prohibited the use of DeepSeek for business purposes Tuesday, becoming the first major tech company to do so. Telecom LG Uplus implemented a similar policy earlier Wednesday.

Other major tech companies, including Samsung Electronics, SK Group and LG Electronics, which are all developing their own generative AI services, have banned such programs on company computers without prior authorization.

The financial sector is also joining the movement to block DeepSeek.