DeepSea: Redefining AI Efficiency

Jonathan Ross is the Co-Founder and CEO of Groq, providing fast AI inference. Prior to founding Groq, Jonathan started Google’s TPU effort where he designed and implemented the core elements of the original chip. Jonathan then joined Google X’s Rapid Eval Team, the initial stage of the famed “Moonshots factory,” where he devised and incubated new Bets (Units) for Alphabet. 

The 10 Most Important Questions on Deepseek:

1. How did Deepseek innovate in a way that no other model provider has done?

2. Do we believe that they only spent $6M to train R1?

3. Should we doubt their claims on limited H100 usage? Is Josh Kushner right that this is a potential violation of US export laws?

4. Is Deepseek an instrument used by the CCP to acquire US consumer data?

5. How does Deepseek being open-source change the nature of this discussion?

6. What should OpenAI do now? What should they not do?

7. Does Deepseek hurt or help Meta who already have their open-source efforts with Lama?

8. Will this market follow Satya Nadella’s suggestion of Jevon’s Paradox?

9. How much more efficient will foundation models become?

10. What does this mean for the $500BN Stargate project announced last week?