Do frontier AI labs abandon or neglect products because their focus is on scaling toward AGI?

Aidan Gomez · Aug 19, 2024 · hedged

The new OpenAI is fundamentally a consumer product company and its AGI effort is taking a backseat to product

They are hardcore building a consumer product that people love and that is now a household name, they have to become a product company to foot the bill for what they want to build, and the pattern of recent departures suggests AGI work is being deprioritised

33:24 Openai has become a product focused consumer company with agi taking a backseat

Ethan Mollick · Jul 31, 2024

Silicon Valley labs neglect usability and documentation because they believe scale solves everything, so any smart person or compute cycle not spent on scaling is considered wasted

If you could build a machine god you win, so the plan is to come back and figure out usage later; and past application work does get obsoleted — a large financial institution's GPT-3 sales assistant was instantly obsolete when ChatGPT shipped

16:16 Scale obsession crowds out usability and documentation investment

Ethan Mollick · Jul 31, 2024

OpenAI abandons products constantly because its products are individual passion projects while all serious talent and compute are directed at building AGI

If you have talented staff or compute, you throw them at the next technology for AGI; their revenue is almost incidental and there isn't really a product beyond the chatbot and the API

25:20 Products are side projects since talent and compute chase agi

Aravind Srinivas · Jun 5, 2024

Perplexity should not be seen as an OpenAI competitor at all, only as a product competing with ChatGPT for user mindshare

OpenAI is an AGI lab whose leadership is explicitly indifferent to profit, whereas Perplexity is a product business that must be profitable

43:09 Agi labs like openai operate on different incentives indifferent to profit than product companies like perplexity so they arent true competitors