I think it's worth noting that 5/8 of the problem profiles you call bizarre are bundled under a single "Emerging Priorities" problem. I also think these five are substantially more bizarre-sounding than the other three and that their being bundled in this way reflects a reasonable self-awareness on 80k's part. 

Separately, this does not follow from 80k listing the priorities that they do:

it implicitly endorses the idea that the more strange an idea is, the more likely it is to be impactful. This turns EA as a project into a search for increasingly bizarre ideas by setting bizarreness as a goal rather than a happenstance.

Neglectedness is an independent desiderata. Because of this, EA priorities will often appear more bizarre than more familiar priorities (because neglectedness correlates with bizarreness), but that does not mean EAs (or 80k in this case) are trying to maximize bizarreness unconstrained by strong arguments for overall importance and tractability.

I agree that *EA* has lost something here, but 80k has a strong argument that *the world* has gained enough from the change to offset that loss. They think the impact of the users attracted by- and directed-to AI safety and these more niche areas will outweigh the impact of users turned off by their unusual priorities. It's a conscious decision on their part. I've talked to them about it.

The broader point is that EA is not self recommending. Sometimes the most EA thing to do is not to dwell on cause prioritization, but to act under greater-than-ideal amounts of uncertainty or public appeal.

This post elides a distinction between reputation as process and reputation as an ~instrumental goal. 

I think reputations are (properly) earned by *how* you do what you do, not *what* you choose to do. As (I think) Claude says: emails not returned and invitations not extended. It matters to have a reputation for being honest, attentive, forthright, transparent, diligent, fair, etc. If your reputation consciously consists more in "doing things that are popular," that is one of several roads to losing yourself. 

People mock EAs for caring about shrimp and insects and viciously attack them for prioritizing future generations. Hiring a much-maligned ex-con because they are the best fit for a role is far lower stakes than those things, but it is still a substantive decision; the kind of thing where if you let "what will others think?" dictate your decision (independent of whether their objections have object-level merit), there's no principled reason you would refuse to abandon some important moral constituency on the same grounds if you thought the ends justified the means. Another important reputation to maintain is "will not be cowed out of their beliefs by mere social pressure."

More importantly, if people knew that you were taking popularity into account for its own sake in deciding what to do, this would and should *hurt* your reputation. Honesty about what you would do on the object level is a virtue.

There should be unpopular EAs and unpopular EA orgs. We should not try to be a monolith managing each other to manage other people's impressions of EA as a whole.

If someone on the Forum team has an automated way of making these each their own event or otherwise advertising them as such, that'd be very helpful!

As a sometimes-helper and sometimes-hater of you videos, I can certify that this one was real good. Gripping from start to finish and did a nice job illustrating the kinds of headwinds people trying to put goodness first will face when trying to intervene in AI.

And just as a piece of art, the way you changed frames from the board drama play-by-play to the flashbacks was masterful.

Very excited to see you two in particular doing this! I'm happy to facilitate a cohort or help with whatever else you need.

I think your suggestion about talented people's ability to triage advice is at least partially true. A problem with it is that a smart response to bad or inapplicable advice is to stop listening to the advice-giver entirely. So you only get a few chances to make contact and you shouldn't spend them on the lower-value content.

Thanks Vasco. Conceptually was very helpful in generating my original list. I think the other lists have value, but are more more observational heuristics that might explain many micro issues, but not really help you make sense of the world and how things work overall. 

Thank you so much Raymond these are great. All Tier 1 and 2 concepts I failed to include. And thank you for catching a typo in the sheet!

Thank you Jen! And we might just take you up on that...