Home Creators Posts Import Register Favorites Logout

Content

Join me (via Google Hangout) on Saturday, Mar 25th at 9AM PDT [GCal] for a discussion of GPT-4 and its implications for augmented cognition.

Please read the technical report before attending; bring your noticings, wonderings, and ideas.

Comments

Anonymous

Is there any chance you'd be willing to record it for people in different timezones? :)

Anonymous

Not sure if any of these questions are on point for what you have in mind, but sharing in case anything sparks engagement. 1. What are your best metaphors for an LLM? (e.g. transistor, catalyst, calculator, intern, armory, medium). Do LLMS _absorb knowledge by reading sentences_? 2. What are your favorite applications of GPT-4? If you had a million dollars what application would you build? 3. What does Alan Kay think about LLMs? Do you have any thoughts on [adept.ai](https://www.adept.ai/)? 4. What do you think of how khanmigo has used GPT-4? 5. Why does the curse of dimensionality not apply to an LLM? What is the inuition for a model that is bigger (in bytes) than the training data? What do we lose when using much smaller versions like Dolly6B/Alpaca7B?

silka sietsma

I missed the chat, unfortunately. If you recorded, I would love to watch and learn.

Can Sar

Sadly missed this too. Parenting schedule makes weekend events a bit of a gamble. Would love a recording too.