Interconnects

96 Épisodes

1. Switched to Claude 3.5

   Publié: 03/07/2024
2. Interviewing Dean Ball on AI policy: CA SB 1047, upcoming AI disaster response, Llama 3 405B, Chinese open-source AI, and scaling laws

   Publié: 27/06/2024
3. RLHF Roundup: Trying to get good at PPO, charting RLHF's impact, RewardBench retrospective, and a reward model competition

   Publié: 26/06/2024
4. Frontiers in synthetic data

   Publié: 21/06/2024
5. Text-to-video AI is already abundant

   Publié: 18/06/2024
6. AI for the rest of us

   Publié: 12/06/2024
7. A realistic path to robotic foundation models

   Publié: 05/06/2024
8. We aren't running out of training data, we are running out of open training data

   Publié: 29/05/2024
9. Name, image, and AI's likeness

   Publié: 22/05/2024
10. OpenAI chases Her

    Publié: 16/05/2024
11. OpenAI's Model (behavior) Spec, RLHF transparency, and personalization questions

    Publié: 13/05/2024
12. RLHF: A thin line between useful and lobotomized

    Publié: 01/05/2024
13. Phi 3 and Arctic: Outlier LMs are hints

    Publié: 30/04/2024
14. AGI is what you want it to be

    Publié: 24/04/2024
15. Llama 3: Scaling open LLMs to AGI

    Publié: 21/04/2024
16. Stop "reinventing" everything to "solve" alignment

    Publié: 17/04/2024
17. The end of the "best open LLM"

    Publié: 15/04/2024
18. Why we disagree on what open-source AI should be

    Publié: 03/04/2024
19. DBRX: The new best open LLM and Databricks' ML strategy

    Publié: 29/03/2024
20. Evaluations: Trust, performance, and price (bonus, announcing RewardBench)

    Publié: 21/03/2024