GradientDissenter

• GradientDissenter 14 Aug 2025 23:36 UTC

  1 point

  0

  in reply to: Caleb Biddulph’s comment on: CoT May Be Highly In­for­ma­tive De­spite “Un­faith­ful­ness” [METR]

  Oh oops! Yes I did. I’ll fix that.

CoT May Be Highly In­for­ma­tive De­spite “Un­faith­ful­ness” [METR]

METR’s Eval­u­a­tion of GPT-5

• GradientDissenter 24 Apr 2025 4:31 UTC

  10 points

  4

  in reply to: MichaelDickens’s comment on: MichaelDick­ens’s Shortform

  I don’t know what’s going on inside the heads of x-risk people such that they see new evidence on the potentially imminent demise of humanity and they find it “exciting”.

  I take your point, and it’s an important one, but I find your claim to not know what’s going on in these people’s heads to be too strong. I feel excited about some kinds of new evidence about “the potentially imminent demise of humanity” like the time horizon graph you mention because I had already priced in the risks this evidence points to and, the evidence just made it way more legible and makes it much easier to communicate my concerns (and getting the broader public and governments to understand this kind of thing seems paramount for safety).

  This is especially true for researchers getting excited about publishing their own work because they’ve known their own results for months usually before they’ve published it and so publishing it just means they’re more legible while the updates are completely priced in.

  I think there’s also a tendency I have in myself to feel much too happy when new evidence makes things I was worried about legible for the same reason I enjoy saying I-told-you-so when my friends make mistakes I warned them about even though I care about my friends and I would have preferred they didn’t make these mistakes. This is definitely a silly quirk of my brain but I don’t think it’s a big problem; it definitely doesn’t push me to cause the things I’m predicting to come to fruition in cases where that would be bad.

• GradientDissenter 4 Mar 2025 0:58 UTC

  6 points

  7

  in reply to: winstonBosan’s comment on: Daniel Koko­ta­jlo’s Shortform

  I assume “let’s think step by step” is a joke/​on purpose. The “first” and “not so fast” on their own don’t seem that egregious to me.

• GradientDissenter 9 Jan 2025 18:40 UTC

  92 points

  10

  on: Gra­di­en­tDissen­ter’s Shortform

  Humanity has only ever eradicated two diseases (and one of those, rinderpest, is only in cattle not humans). The next disease on the list is probably Guinea worm (though polio is also tantalizingly close).
  
  At its peak Guinea worm infected ~900k people a year. In 2024 we so far only know of 7 cases. The disease isn’t deadly, but it causes significant pain for 1-3 weeks (as a worm burrows out of your skin!) and in ~30% of cases that pain persists afterwards for about a year. In .5% of cases the worm burrows through important ligaments and leaves you permanently disabled. Eradication efforts have already saved about 2 million DALYs.^[1]
  
  I don’t think this outcome was overdetermined; there’s no recent medical breakthrough behind this progress. It just took a herculean act of international coordination and logistics. It took distributing millions of water filters, establishing village-based surveillance systems in thousands of villages across multiple countries, and meticulously tracking every single case of Guinea worm in humans or livestock around the world. It took brokering a six-month ceasefire in Sudan (the longest humanitarian ceasefire in history!) to allow healthcare workers to access the region. I’ve only skimmed the history, and I’m generally skeptical of historical heroes getting all the credit, but I tentatively think it took Jimmy Carter for all of this to happen.
  
  Rest in peace, Jimmy Carter.

  1. ^

     I’m compelled to caveat that top GiveWell charities are probably in the ballpark of $50/​DALY, and the Carter Center has an annual budget of ~$150 million a year, so they “should” be able to buy 2 million DALYs every single year by donating to more cost-effective charities. But c’mon this worm is super squicky and nearly eradicating it is an amazing act of agency.

Gra­di­en­tDissen­ter’s Shortform

• GradientDissenter 24 Dec 2022 3:05 UTC

  21 points

  0

  on: LessWrong read­ers are in­vited to ap­ply to the Lurkshop

  Update: we sent out decisions to everyone who applied to the upcoming program, which is now Jan 6-8!

  The fact that these decisions were so delayed and communication was poor and the program got postponed is all my fault and I would like to apologize. I understand that my being unresponsive created costs for you and makes planning harder. It also eats into a valuable resource of trust that makes people more suspicious of other events and generally makes coordination difficult.

  This happened because I didn’t set firm enough deadlines or loop my collaborators in enough on the project, and then once I fell a little bit behind because we got more applications than I expected, I grew very averse to evaluating them. That’s not to say this was excusable—this is the kind of mistake that should cause people to trust me less. This is a fairly unprecedented error for me and I wrote down some reflections on why this happened, which I’ve shared with everyone who applied. I don’t think I’ll make another mistake of this flavor any time soon.

  One thing I’d like to emphasize is this was entirely my mistake—Jonas and the rest of the people working on the Lurkshop and at Atlas and LessWrong tried to help but I didn’t communicate with them well or keep them in the loop.

  Again, thank you everyone for your patience and I’m sorry for the hassle I’ve caused.

  I’m excited to see some of you in person soon!

• GradientDissenter 24 Dec 2022 2:37 UTC

  5 points

  0

  in reply to: Lost Futures’s comment on: LessWrong read­ers are in­vited to ap­ply to the Lurkshop

  Everyone who applied to the upcoming program should have heard back now! (Decisions were sent out shortly after you posted your comment.) People who said they couldn’t make the upcoming program but wanted to be considered for future programs we might run haven’t been notified, since we don’t know what those events will look like yet or if they’ll happen.

• GradientDissenter 1 Dec 2022 10:53 UTC

  1 point

  0

  in reply to: ai dude’s comment on: LessWrong read­ers are in­vited to ap­ply to the Lurkshop

  Sure, if people prefer getting money up front I’m happy to do that.

LessWrong Lurk­shop (ap­ply by Dec 1st)

• GradientDissenter 30 Nov 2022 1:37 UTC

  1 point

  0

  in reply to: Eric Raymond’s comment on: LessWrong read­ers are in­vited to ap­ply to the Lurkshop

  Thanks Eric! I’ll be sure to reach out if/​when I run something like this in the future! I hope you recover well.

• GradientDissenter 29 Nov 2022 10:41 UTC

  4 points

  0

  in reply to: porby’s comment on: LessWrong read­ers are in­vited to ap­ply to the Lurkshop

  1. Lodging is on-site. It’s at a renovated former sorority house. Everyone should get their own bedroom, and if getting a private bathroom is a crux for doing we can make sure you get that too, though by default bathrooms will be shared.

  2. You’d submit a reimbursement request for expenses. If that poses a challenge, we can work something else out. I expect the only significant expense you’d need reimbursement for to be travel—we’ll provide food and lodging.

  3. I expect 20 people, but it wouldn’t surprise me if it turned out to be significantly more or fewer. In general, it will be small.

• GradientDissenter 26 Nov 2022 22:52 UTC

  3 points

  0

  in reply to: GradientDissenter’s comment on: LessWrong read­ers are in­vited to ap­ply to the Lurkshop

  Here’s the tentative schedule! (Also added to the post).

• GradientDissenter 25 Nov 2022 3:43 UTC

  4 points

  0

  in reply to: Yitz’s comment on: LessWrong read­ers are in­vited to ap­ply to the Lurkshop

  We’re pretty unsure what the quality of the applications will be. My 85% CI is that we accept 5-50% of people who just read this post and apply.

• GradientDissenter 23 Nov 2022 18:41 UTC

  1 point

  0

  in reply to: GradientDissenter’s comment on: LessWrong read­ers are in­vited to ap­ply to the Lurkshop

  Another point related to the above: if I had content I wanted LessWrong readers to know that was easy to transmit in written form, I’d just post it on LessWrong. I’m trying to make the most of the opportunity have interactions that can’t easily happen over the forum (and I think there’s a lot of important interactions of that class!).

• GradientDissenter 22 Nov 2022 22:46 UTC

  6 points

  4

  in reply to: Dagon’s comment on: LessWrong read­ers are in­vited to ap­ply to the Lurkshop

  Hiya! I’m the co-organizer Jonas mentioned.

  This is partially an un-conference style event. The point is not that we the staff have something we want to present to the participants; the point is if we get these people together, we think something worthwhile will happen. (My personal prediction is that at least two participants’ lives will majorly shift because of this program.)

  Most of the most valuable events I’ve been to had excellent people and ~no scheduled content. For the most valuable events I’ve been to with content, the content wasn’t the thing that provided the majority of the value for me or most other attendees.

  I’ve run ~9 other large workshop-type events and many smaller events, and all of them had a lot of content, but the content mostly functioned to set a context for good conversations. I suspect for the demographic of LessWrong readers, having lots of classes won’t be as useful for context-setting, since there’s already a large body of interesting topics that are common knowledge that people can discuss.

  All that said, if I have time, I’ll share a default/​baseline schedule later tonight.

LessWrong read­ers are in­vited to ap­ply to the Lurkshop