The most prestigious law school admissions discussion board in the world.
|
 |
 
|
Deepseek (chinese) releases open weight LLM that is ChatGPT quality
| magadood | 12/30/24 |
| the ordeal of tranquility | 12/30/24 |
| magadood | 12/30/24 |
| the ordeal of tranquility | 12/30/24 |
Poast new message in this thread
Date: December 30th, 2024 2:12 PM Author: magadood https://www.maginative.com/article/deepseek-v3-achieves-frontier-ai-performance-at-a-fraction-of-the-cost/
only cost $5.6 million to train. OpenAI's business model is getting fucked from every direction. Gemini and Claude surpass 4o, Sora sucks compared to Veo 2 and now competitors are releasing cheap, open weight LLM models that are essentially the same quality (and it's the chinese who are supposed to be cut off from the decent hardware). bunch of stories about how GPT-5 is behind schedule and only a marginal improvement over 4o. whatever edge they had is gone. Sam Altman is not going to be as rich as he hoped. |
Date: December 30th, 2024 2:15 PM Author: the ordeal of tranquility Yeah it's really good news
AI is going to end up being an incredible force multiplier for extremely capable individuals, relative to large orgs/institutions/government
These people think they're building up a totalitarian Elysium for themselves but it's going to look a lot more like the ultimate Hobbesian Total War battlefield |
 |
Date: December 30th, 2024 2:18 PM Author: magadood i'd certainly rather see this than let a psychopath like Altman control everything. pretty clear now that OpenAI is not benevolent. (http://www.autoadmit.com/thread.php?thread_id=5656826&forum_id=2:#48501817) |
|
Date: December 30th, 2024 2:30 PM Author: the ordeal of tranquility (http://www.autoadmit.com/thread.php?thread_id=5656826&forum_id=2:#48501858) |