All of pavoras's Comments + Replies

As usual, beautifully animated and made accessible for a wide audience!

Thank you for writing this. I understand the feeling of worry and being overwhelmed in regards to AI. It sure does feel unlikely to get global politics to listen and, ideally, on board for this...

I studied chemistry and worked in Pharma research for half a year (Germany, 2018). Standard safety procedures are there, but it's no biology lab. Stuff gets spilled, you breath in vapors of solvents, and the lunch break happens in the office next to the lab, after a quick hand wash.

I am pretty sure up to a microgram of almost anything we made ended up in somebody... Usually we handled 10mg up to 10g batches. Surprisingly, the lab Assistents are overall pretty fine and don't appear to suffer any consequences more than the average population well into their ... (read more)

Thanks! I am not "really" new though. Have just been reading for the last few months but I suppose this actually was my first comment 😅

Also, if you already know LW, the channel screams "Hey, this is LW stuff!". But I don't think I would have made the connection if I encountered the channel first. It does a hell of a job getting the "average" nerd informed and interested in these topics though, which is really nice.

Not sure how the average population fares. I suppose the videos have way too much content to appeal to a mainstream audience... Which is good, in a way. Somebody has to do the "informing the public"-chore at some point though, but it might as well be someone else ^^

Thank you for putting in the effort to make this Video! Even though the comments went into a "lewd=bad?" Discussion pretty quickly, in my opinion it does a very good job of transmitting, that misalignment is something that can actually happen in real life and we should be concerned about. It is valuable, because it connects the theory to real life experience and a true story (also good clickbait). "Catastrophic AI risk" is widely discussed in other videos anyways, so based on transfer learning many viewers will probably get the point anyways.

Also I wanted ... (read more)