OpenAI hiring Fidji Simo is an ominous sign. Ads incoming.
She built Instacart Ads. She was in charge of the advertising business and monetizing mobile at Facebook. Before that, she built out advertising at eBay.
When DeepMind publishes research, you can be sure that it is either Nobel Prize worthy or something DeepMind never intends to use in production - no middle ground.
Some thoughts on the Dwarkesh Richard Sutton interview:
1) Richard Sutton has internalized the bitter lesson to a very impressive degree.
2) He doesn't like pretraining because human set the data used in pretraining. He doesn't like post-training because humans set the
Some thoughts on Harvey:
1) Harvey was initially known for having a mediocre product and a lot of people in the legal tech industry predicted that it would fail.
2) But, it's revenue has grown each time that I have gotten info, last time it was at about $100m in revenue, which
In retrospect, all boils down to the fact classic vc tech playbook doesn’t work with deep tech. No you cannot just scale distribution and sort out product later.
right now, it's actually just a bad product for consumers; $20/m for bad rate limits, no voice mode, no image generation. even hobbyist programmers are better off with ChatGPT for rate limits and o1.
i heard a story last night about alec radford; apparently, he never committed to the monorepo; some new researcher, excited to work with him, asked what branch he was on; and Alec sent him a pile of zip files
