Interconnects

109 Episodes

1. We aren't running out of training data, we are running out of open training data

   Published: 5/29/2024
2. Name, image, and AI's likeness

   Published: 5/22/2024
3. OpenAI chases Her

   Published: 5/16/2024
4. OpenAI's Model (behavior) Spec, RLHF transparency, and personalization questions

   Published: 5/13/2024
5. RLHF: A thin line between useful and lobotomized

   Published: 5/1/2024
6. Phi 3 and Arctic: Outlier LMs are hints

   Published: 4/30/2024
7. AGI is what you want it to be

   Published: 4/24/2024
8. Llama 3: Scaling open LLMs to AGI

   Published: 4/21/2024
9. Stop "reinventing" everything to "solve" alignment

   Published: 4/17/2024
10. The end of the "best open LLM"

    Published: 4/15/2024
11. Why we disagree on what open-source AI should be

    Published: 4/3/2024
12. DBRX: The new best open LLM and Databricks' ML strategy

    Published: 3/29/2024
13. Evaluations: Trust, performance, and price (bonus, announcing RewardBench)

    Published: 3/21/2024
14. Model commoditization and product moats

    Published: 3/13/2024
15. The koan of an open-source LLM

    Published: 3/6/2024
16. Interviewing Louis Castricato of Synth Labs and Eleuther AI on RLHF, Gemini Drama, DPO, founding Carper AI, preference data, reward models, and everything in between

    Published: 3/4/2024
17. How to cultivate a high-signal AI feed

    Published: 2/28/2024
18. Google ships it: Gemma open LLMs and Gemini backlash

    Published: 2/22/2024
19. 10 Sora and Gemini 1.5 follow-ups: code-base in context, deepfakes, pixel-peeping, inference costs, and more

    Published: 2/20/2024
20. Releases! OpenAI’s Sora for video, Gemini 1.5's infinite context, and a secret Mistral model

    Published: 2/16/2024