https://twitter.com/AndrewYNg/status/1736577228828496179

This is one of those perspectives that feels wise but I quite disagree with. Consider Covid in Jan 2020. At that time, was it right to focus on (i) actual current realized harm, or (ii) hypotheticals based on projecting exponential functions? Clearly (ii).

And I do think AI is similar. If we focus on actual harm, then honestly AI has been much less harmful than even I predicted. Like, if you told someone in 2018 about the capabilities of GPT4, SD, etc, I expect they would predict: mass unemployment, mass election interference, mass something involving social biases...

Does this reality affect your priors around predicting future harm?

Building /bright-moments - minting onchain art IRL | /purple #15 | FID 129 #magikarp | getting paid to post with @ampsfun

If I didn't perceive the reality that way I would definitely be quite a bit closer to being a "doomer"!