There are 56 people in the bootcamp. The bottom half of the program is shockingly bad, and I don’t think they’ll get much better.
I think only 5-10 people have actual potential to influence AI safety in some way. I’m not sure this is a good use of money for the rest, but perhaps the funders are thinking of this as a hits businesses, or that future cohorts will improve in quality.
I have been doing something similar since the beginning of June, posting daily on TikTok. Some AI safety related content, and a lot of other stuff.
I can definitely see lots of mistakes the 10 top creators you listed are making. I’m pretty sure I’m at least a better content creator now than the average of those 10 people. I’m also pretty sure I was significantly worse than all of them on day one, probably what you’d call shockingly bad, 36 days ago! So don’t discount the effect of daily practice!
I’m too embarrassed to share my TikTok :/ If you want an example of mistakes that I saw, one is that introductions like “Hi, my name is X” or “Today we’re going to talk about X” are algorithmic poison on short form video sites.
Of particular interest to people sharing AI safety content is the fact that people will skip any videos that aren’t immediately and clearly going to provide value to their lives. “You’re going to die maybe in a few years” doesn’t count. So are you very obviously telling a joke that’s going to be funny? Are you telling a compelling story they can’t help but listen to? Are you giving them an identity as an insider who knows more than those stupid outsiders? Are you teaching them something about AI they can use right away to do better at work? The value has to be clear and evident in the first 1 second.
Anyway, I plan to make a post about what I’ve learned on my own journey after the 2 month mark.
I would be very surprised if the thinking was ‘and this gets us 56 good new people’ vs ‘this gets us 5-10 counterfactual people, either by helping existing creators shift focus, or by giving value-aligned people a structure to get going.’
also my sense is the per-participant cost is << one FTE doing the same work.
plzdontkillus is an AI safety content creator bootcamp, and I watched all of the videos from the first five days.
Here are my top 10 creators with the most potential, based on:
presentation style
engagement provided by video
how likely I would be to be influenced by them about AI safety were I in their general audience.
Note that I removed people with standing experience as content creators, like Avisha NessAiver and Christian González Capizzi.
Alexander Highcliff
broomhead_ai
Romy
Verónica Reinero
Brenda Godinez
Darknetdoll
Josie Camarillo
Tyler Alterman
Chefboy_AI
General thoughts:
There are 56 people in the bootcamp. The bottom half of the program is shockingly bad, and I don’t think they’ll get much better.
I think only 5-10 people have actual potential to influence AI safety in some way. I’m not sure this is a good use of money for the rest, but perhaps the funders are thinking of this as a hits businesses, or that future cohorts will improve in quality.
Here’s the full video feed, let me know if you disagree!
I have been doing something similar since the beginning of June, posting daily on TikTok. Some AI safety related content, and a lot of other stuff.
I can definitely see lots of mistakes the 10 top creators you listed are making. I’m pretty sure I’m at least a better content creator now than the average of those 10 people. I’m also pretty sure I was significantly worse than all of them on day one, probably what you’d call shockingly bad, 36 days ago! So don’t discount the effect of daily practice!
I’m too embarrassed to share my TikTok :/ If you want an example of mistakes that I saw, one is that introductions like “Hi, my name is X” or “Today we’re going to talk about X” are algorithmic poison on short form video sites.
Of particular interest to people sharing AI safety content is the fact that people will skip any videos that aren’t immediately and clearly going to provide value to their lives. “You’re going to die maybe in a few years” doesn’t count. So are you very obviously telling a joke that’s going to be funny? Are you telling a compelling story they can’t help but listen to? Are you giving them an identity as an insider who knows more than those stupid outsiders? Are you teaching them something about AI they can use right away to do better at work? The value has to be clear and evident in the first 1 second.
Anyway, I plan to make a post about what I’ve learned on my own journey after the 2 month mark.
I would be very surprised if the thinking was ‘and this gets us 56 good new people’ vs ‘this gets us 5-10 counterfactual people, either by helping existing creators shift focus, or by giving value-aligned people a structure to get going.’
also my sense is the per-participant cost is << one FTE doing the same work.