O'Malley
@omalleyhh.bsky.social
Reposted by O'Malley
Reminder that these protests work:
- Evidence from Greece:
www.cambridge.org/core/journal...
- Evidence from Italy: academic.oup.com/poq/article-...
- Evidence from France: academic.oup.com/ej/advance-a...
- Evidence from Greece:
www.cambridge.org/core/journal...
- Evidence from Italy: academic.oup.com/poq/article-...
- Evidence from France: academic.oup.com/ej/advance-a...
February 2, 2025 at 5:42 PM
Reminder that these protests work:
- Evidence from Greece:
www.cambridge.org/core/journal...
- Evidence from Italy: academic.oup.com/poq/article-...
- Evidence from France: academic.oup.com/ej/advance-a...
- Evidence from Greece:
www.cambridge.org/core/journal...
- Evidence from Italy: academic.oup.com/poq/article-...
- Evidence from France: academic.oup.com/ej/advance-a...
Reposted by O'Malley
gentle reminder that deepseek is an open-weight model and NOT open source. there's nothing open about it; no information on how its trained or data trained on, just like its western counterparts
Any company whose mission is to “to ensure that artificial general intelligence benefits all of humanity” would be ecstatic about a capable open source model like Deepseek released.
Any company which has a mission of making money off closed-source models would be devastated tho.
Any company which has a mission of making money off closed-source models would be devastated tho.
January 26, 2025 at 1:58 PM
gentle reminder that deepseek is an open-weight model and NOT open source. there's nothing open about it; no information on how its trained or data trained on, just like its western counterparts
Reposted by O'Malley
Why is it making waves?
- it's open source
- “profited from open research (PyTorch and Llama)” LeCun, LinkedIn
- Beats o1 with only $5.6 million to train (www.wsj.com/tech/ai/chin...)
- Destabilizes US corp edge on AI in spite of US sanctions on chip sales to China
techcrunch.com/2025/01/26/d...
- it's open source
- “profited from open research (PyTorch and Llama)” LeCun, LinkedIn
- Beats o1 with only $5.6 million to train (www.wsj.com/tech/ai/chin...)
- Destabilizes US corp edge on AI in spite of US sanctions on chip sales to China
techcrunch.com/2025/01/26/d...
Silicon Valley Is Raving About a Made-in-China AI Model
DeepSeek is called ‘amazing and impressive’ despite working with less-advanced chips.
www.wsj.com
January 27, 2025 at 9:02 PM
Why is it making waves?
- it's open source
- “profited from open research (PyTorch and Llama)” LeCun, LinkedIn
- Beats o1 with only $5.6 million to train (www.wsj.com/tech/ai/chin...)
- Destabilizes US corp edge on AI in spite of US sanctions on chip sales to China
techcrunch.com/2025/01/26/d...
- it's open source
- “profited from open research (PyTorch and Llama)” LeCun, LinkedIn
- Beats o1 with only $5.6 million to train (www.wsj.com/tech/ai/chin...)
- Destabilizes US corp edge on AI in spite of US sanctions on chip sales to China
techcrunch.com/2025/01/26/d...
Reposted by O'Malley
As someone who has reported on AI for 7 years and covered China tech as well, I think the biggest lesson to be drawn from DeepSeek is the huge cracks it illustrates with the current dominant paradigm of AI development. A long thread. 1/
January 27, 2025 at 2:12 PM
As someone who has reported on AI for 7 years and covered China tech as well, I think the biggest lesson to be drawn from DeepSeek is the huge cracks it illustrates with the current dominant paradigm of AI development. A long thread. 1/