← Blog | Thoughts on GPT-3

Thoughts on GPT-3

Published: 2020

Reference: https://twitter.com/rrichrs/status/1284939006195593217?s=20

Roam Source: https://roamresearch.com/#/app/ryanrich/page/jOb93iIMo

GPT-3 summer 2020 is like crypto summer 2017.

Companies being funded over a tremendous amount of hype. These companies, like their ICO kin, will not survive the cycle. Aside from Coinbase and other exchanges, have we seen any real breakout crypto startups? We’ve certainly seen billions in capital raised, but very few results — I really can’t think of an example of a pure crypto play that has a significant user base let alone revenue.

GPT is real, just like crypto is real, but maturity is still 10 years out. I fear investors that have little understanding of the technology are hyping it up (GPT-3 isn’t even generally available yet, and may never be).

10 years is an arbitrary amount of time. If we still don’t have AGI or decentralized infrastructure in 10 years I would not be surprised at all. Societal changes need a strong forcing function. Strong forcing functions take time. Even the effects of COVID-19 are beginning to wear off (companies are abandoning remote work, moving back into offices) and that’s one of the most significant events a person can go through in a lifetime.

GPT-X but pre-AGI will improve work we already mostly automate today. Data entry, mech turk, even copywriting (think auto-complete today), but it won’t be eliminating entire industries overnight.