Hey, I was a fellow at pdku. You did a great job capturing a bit of the fever dream that was this past month. Shame we didn't get a chance to talk though.
I will say on the disagreement front that while the overwhelming majority of fellows were "doomers" not all of us were. Essentially all of my AI safety related videos were anti-doom (published in the first week or so before you arrived iirc). I also filmed an entire longform podcast with Nate Soares where we debated the doom question!
But your overall assessment was definitely correct. While the organizers did explicitly welcome people with different opinions, and tried really hard to enforce norms around this, most of the fellows ended up (or were self-selected for) being more doomer.
As to whether pdku was successful, I think this the answer is moderately yes but still tbd. I think between all of us we got something like 100m views. The majority wasn't ai safety related but I still think that's a good start. A lot of us figured out better ways of doing safety comms (not being so on the nose, honors our already existing niches) a little later.
But my read of the situation is we won't see the real effects of pdku for at least a few months. Several fellows are in the process of getting hired or funded by various safety-related orgs. Add to that the possibility of 3 more cohorts next year and it's hard to predict the ripple effects of there finally being a rationalist/EA/safety related network that's short form-native. We're also, as you described, thinking about doing a content house once many of us move to the Bay Area + pending funding. This could serve as a hub in the network. So we'll see!
The biggest success of the program for me, however, was as an experiment in living. Getting 55 weirdo creative types to live together at a "creator commune" was beyond fun and Lighthaven was the perfect venue for it. A lot of us walked away from this wanting to recreate similar living conditions albeit in a more modest fashion. Everyone's bedtime slowly drifted to like 4am by the end of the program just because no one wanted to stop talking by the fire pits. It was a special experience.
Glad you liked it! (and sorry to have missed you IRL -- simply too many fellows, too little time)
Totally agree re: not everyone being a Doomer(tm), but I think the narrative convergence was inevitable given the organizers'/funders' beliefs + self-selection bias. I'm curious to see what happens in the coming months!
I'm nonbinary and an AIS tech gov researcher. This is hilarious. I have not met a single other person with dyed hair in the field so far, this must have been an experience. I should start bringing xrisk quotes into my goth DJ sets
Thanks for pointing this out! I was at the July protest and agree that the diversity (especially across speakers) was much improved. Like CJ said, our conversation was specifically about MIRI/PDKU, not Stop the AI Race as an org.
Hi Celia, thank you so much for this detailed reporting of pdku, we need more content like this!
Regarding the ick quote on the "AI Race" protest: I have had the opposite reaction, from two people that are deep into Bay Area activism who dislike the AI Safety community, who were really happy about how this bigger tent protest went, especially about the diversity of the coalition, speakers and people attending. I'm saying this because these are people who explictly mentioned to me having the ick when it comes to being too close to sillicon valley, AI Safety transhumanity vibes, and they were like "oh yeah michael i'm so glad you fixed that at this protest". Happy to connect!
> "What most people — likely including Becca’s friends — actually resent is Silicon Valley overreaching into everyone’s lives without consent."
Well, it turns out that building technology without our consent was at the core of the coalition message on the protest website (https://theprotest.ai/). Under "who marched" it says "Labor displacement. Surveillance. Bias. Stolen work. Existential risk. Different harms, different strategies, one demand: stop developing more advanced AI without our consent."
> tapping into this resentment might mean letting people who don’t care about x-risk (yet) take the wheel [...] The advice CJ keeps giving them: “You need to be losing control of the narrative.” But, they added, “I’m not convinced that they’re willing to let go of being right.”
I'm assuming you're talking about pdku & the MIRI narrative here, but since it's directly after the paragraph on the protest, I wanted to flag that as part of "The AI Protest" on July 11 we've had many 1o1 meetings, and multiple coalition meetings, with several different orgs who didn't care about x-risk very much, which eventually lead to the "stop developing AI without our consent" bit.
speaking as the CJ quoted I think the Stop The AI Race protests have been a really positive example of Big Tent-ing in a good way. Keep up the good work
Okay so activism to these people means throwing a month-long party with a bunch of edgy influencers? And this is supposed to make the general public take their fears more seriously? LOL.
I mean compare that to the climate activists. That's an actual grassroots movement where the activists aren't waterboarding each other for the lulz, they're going to great personal sacrifice - even risking jail time - to make people talk more about climate change. They aren't paid by shady companies, they have to ask ordinary people to donate to them. If you talk to any one of them it's pretty obvious that they care deeply about doing the right thing and aren't motivated by desire to acquire status or whatever. (Besides, some of their actions are deeply unpopular, so that would be a poor idea.)
Which is entirely different to how this crowd operates. You would think that if AI was actually an existential risk they'd behave a little more seriously rather than making so much noise about how ✨ cool ✨ and ✨ weird ✨ they all are. No, instead they admit to running a psyop!
This is nothing more than a month-long networking event for trend-hopping losers all united by the fact that they want attention from people on the internet to make up for their utter lack of integrity, morals, or any other once-lauded personality trait that has been made irrelevant in the meaningless hell that is late stage capitalism. God is dead.
Great piece that shows while PDKU tries not to indoctrinate its fellows, the same x-risk narrative largely reproduces itself anyway. Glad there were some non-doomers there.
Hey, I was a fellow at pdku. You did a great job capturing a bit of the fever dream that was this past month. Shame we didn't get a chance to talk though.
I will say on the disagreement front that while the overwhelming majority of fellows were "doomers" not all of us were. Essentially all of my AI safety related videos were anti-doom (published in the first week or so before you arrived iirc). I also filmed an entire longform podcast with Nate Soares where we debated the doom question!
But your overall assessment was definitely correct. While the organizers did explicitly welcome people with different opinions, and tried really hard to enforce norms around this, most of the fellows ended up (or were self-selected for) being more doomer.
As to whether pdku was successful, I think this the answer is moderately yes but still tbd. I think between all of us we got something like 100m views. The majority wasn't ai safety related but I still think that's a good start. A lot of us figured out better ways of doing safety comms (not being so on the nose, honors our already existing niches) a little later.
But my read of the situation is we won't see the real effects of pdku for at least a few months. Several fellows are in the process of getting hired or funded by various safety-related orgs. Add to that the possibility of 3 more cohorts next year and it's hard to predict the ripple effects of there finally being a rationalist/EA/safety related network that's short form-native. We're also, as you described, thinking about doing a content house once many of us move to the Bay Area + pending funding. This could serve as a hub in the network. So we'll see!
The biggest success of the program for me, however, was as an experiment in living. Getting 55 weirdo creative types to live together at a "creator commune" was beyond fun and Lighthaven was the perfect venue for it. A lot of us walked away from this wanting to recreate similar living conditions albeit in a more modest fashion. Everyone's bedtime slowly drifted to like 4am by the end of the program just because no one wanted to stop talking by the fire pits. It was a special experience.
Anyways, excellent piece!
Glad you liked it! (and sorry to have missed you IRL -- simply too many fellows, too little time)
Totally agree re: not everyone being a Doomer(tm), but I think the narrative convergence was inevitable given the organizers'/funders' beliefs + self-selection bias. I'm curious to see what happens in the coming months!
I was also a pdku fellow and I don't think the overwhelming majority of fellows were doomers - I'd guess that doomers were a slight majority
I'm nonbinary and an AIS tech gov researcher. This is hilarious. I have not met a single other person with dyed hair in the field so far, this must have been an experience. I should start bringing xrisk quotes into my goth DJ sets
Buck shelgeris I am pretty sure has trademark dyed hair!
An icon, it's true!
Dyed hair still wasn't the majority, but definitely nonzero!
batshit insane. perfect piece.
Thanks for pointing this out! I was at the July protest and agree that the diversity (especially across speakers) was much improved. Like CJ said, our conversation was specifically about MIRI/PDKU, not Stop the AI Race as an org.
Hi Celia, thank you so much for this detailed reporting of pdku, we need more content like this!
Regarding the ick quote on the "AI Race" protest: I have had the opposite reaction, from two people that are deep into Bay Area activism who dislike the AI Safety community, who were really happy about how this bigger tent protest went, especially about the diversity of the coalition, speakers and people attending. I'm saying this because these are people who explictly mentioned to me having the ick when it comes to being too close to sillicon valley, AI Safety transhumanity vibes, and they were like "oh yeah michael i'm so glad you fixed that at this protest". Happy to connect!
> "What most people — likely including Becca’s friends — actually resent is Silicon Valley overreaching into everyone’s lives without consent."
Well, it turns out that building technology without our consent was at the core of the coalition message on the protest website (https://theprotest.ai/). Under "who marched" it says "Labor displacement. Surveillance. Bias. Stolen work. Existential risk. Different harms, different strategies, one demand: stop developing more advanced AI without our consent."
> tapping into this resentment might mean letting people who don’t care about x-risk (yet) take the wheel [...] The advice CJ keeps giving them: “You need to be losing control of the narrative.” But, they added, “I’m not convinced that they’re willing to let go of being right.”
I'm assuming you're talking about pdku & the MIRI narrative here, but since it's directly after the paragraph on the protest, I wanted to flag that as part of "The AI Protest" on July 11 we've had many 1o1 meetings, and multiple coalition meetings, with several different orgs who didn't care about x-risk very much, which eventually lead to the "stop developing AI without our consent" bit.
speaking as the CJ quoted I think the Stop The AI Race protests have been a really positive example of Big Tent-ing in a good way. Keep up the good work
Okay so activism to these people means throwing a month-long party with a bunch of edgy influencers? And this is supposed to make the general public take their fears more seriously? LOL.
I mean compare that to the climate activists. That's an actual grassroots movement where the activists aren't waterboarding each other for the lulz, they're going to great personal sacrifice - even risking jail time - to make people talk more about climate change. They aren't paid by shady companies, they have to ask ordinary people to donate to them. If you talk to any one of them it's pretty obvious that they care deeply about doing the right thing and aren't motivated by desire to acquire status or whatever. (Besides, some of their actions are deeply unpopular, so that would be a poor idea.)
Which is entirely different to how this crowd operates. You would think that if AI was actually an existential risk they'd behave a little more seriously rather than making so much noise about how ✨ cool ✨ and ✨ weird ✨ they all are. No, instead they admit to running a psyop!
This is nothing more than a month-long networking event for trend-hopping losers all united by the fact that they want attention from people on the internet to make up for their utter lack of integrity, morals, or any other once-lauded personality trait that has been made irrelevant in the meaningless hell that is late stage capitalism. God is dead.
Great piece that shows while PDKU tries not to indoctrinate its fellows, the same x-risk narrative largely reproduces itself anyway. Glad there were some non-doomers there.