So Zuck got all these people, but does he have a differentiated strategy for a foundation model? Outside of the access to Meta’s data hoard.

Classic commoditize your complement. 

If you have an open source model that’s at the frontier, time spent in other models / apps / etc. decreases.

Classic commoditize your complement. 

If you have an open source model that’s at the frontier, time spent in other models / apps / etc. decreases. 

https://www.joelonsoftware.com/2002/06/12/strategy-letter-v/

Will read this tonight. 

Was Llama 3 a failure because it just couldn’t compete? Or was it something else? Like the fact that ChatGPT has mindshare and “use case dominance”

Chasing Curiosity. People, Nature, Machines. Hakuna Matata, "Its a problem-free philosophy.”

Fair, assumed it was more of a failure of product than it was a foundational model problem. 

But tbh I’m not watching this that closely.

Do you know anyone who uses Llama (besides the weird Meta character chats)?

OpenAI, Anthropic, Google and maybe Grok are the relevant consumer products / brands.