Yonatan Cale

• Yonatan Cale 17 Aug 2022 7:19 UTC

  4 points

  on: [AMA] Open Philan­thropy is still seek­ing pro­pos­als for out­reach and com­mu­nity-build­ing projects

  I appreciate the AMA style post!

  Are you interested in projects that do outreach and community building to mid-career professionals, especially in tech, where some of the projects help these professionals hear about and/​or transition into longtermism roles?

  (If some of these parts sound like a good fit and some don’t—could you tell me which are which? If any of this sounds like a good fit then I’ll share some projects I’m working on or considering working on)

• Yonatan Cale 16 Aug 2022 7:22 UTC

  2 points

  in reply to: Emrik’s comment on: An­nounc­ing the EA Gather Town Event Hall as a global hub for on­line events

  Yeah, that’s great (all I wanted is the ability to subscribe to it)

  Thanks, added :)

• Yonatan Cale 16 Aug 2022 7:20 UTC

  2 points

  in reply to: Karthik Tadepalli’s comment on: In­ter­na­tion­al­ism is a key value in EA

  I see

  I agree

• Yonatan Cale 14 Aug 2022 19:00 UTC

  2 points

  on: EA Or­ga­ni­za­tions Hiring Frus­tra­tion

  1. You’re not alone. I talk to many people applying to EA jobs and hear things like this sometimes

  2. I’m personally trying to improve EA hiring if I can

  3. My main challenges are things around “EA orgs not getting back to my emails”. This is improving, but slowly

• Yonatan Cale 14 Aug 2022 18:57 UTC

  3 points

  on: An­nounc­ing the EA Gather Town Event Hall as a global hub for on­line events

  Hey!

  May I suggest opening a Google Calendar that people can import and see when events are coming up? I’d totally use that personally

• Yonatan Cale 13 Aug 2022 10:55 UTC

  2 points

  on: In­ter­na­tion­al­ism is a key value in EA

  “internationalism” is only a tool towards a higher value of “doing more good [per dollar]”, at least to me

• Yonatan Cale 12 Aug 2022 10:36 UTC

  2 points

  in reply to: GG’s comment on: [Cause Ex­plo­ra­tion Prizes] Soft­ware Sys­tems for Col­lec­tive Intelligence

  The idea of weighting a decision-making process toward more qualified decision makers (however that is determined)

  My opinion:

  1. “however that is determined” is where the magic happens, not a side point

     1. As a naive example to explain what I’m talking about: if we give more weight to experts, there is the question of “how do we decide who are the experts”. Do we use expert experts? Who decides who those are?

  2. The “Software systems” won’t be the challenging part of this project

     1. Beyond, maybe, trying to give every person one vote (one “identity”) that can be used through the internet , which turns out to be really hard to do without huge problems, and the blockchain community is working really hard on already

• Yonatan Cale 12 Aug 2022 7:29 UTC

  2 points

  in reply to: GG’s comment on: [Cause Ex­plo­ra­tion Prizes] Soft­ware Sys­tems for Col­lec­tive Intelligence

  Thank you, updated.

  Btw I personally really like how the Ethereum community experiments with things like this, seems super healthy to me

• Yonatan Cale 11 Aug 2022 14:20 UTC

  2 points

  in reply to: goodgravy’s comment on: AMA: Ought

  Thank you, this was super informative! My understanding of Ought just improved a lot

  1. Once you’re able to answer questions like that, what do you build next?

  2. Is “Was this a double-blind study?” an actual question that your users/​customers are very interested in?

     1. If not, could you give me some other example that is?

• Yonatan Cale 11 Aug 2022 14:13 UTC

  2 points

  in reply to: Yonatan Cale’s comment on: AMA: Ought

  And my second worry is that the “big AI” (the collection of sub models) will be so good that you could ask it to perform a task and it will be exceedingly effective at it, in some misaligned-to-our-values (misaligned-to-what-we-actually-meant) way

• Yonatan Cale 11 Aug 2022 14:12 UTC

  4 points

  in reply to: stuhlmueller’s comment on: AMA: Ought

  it’s plausible that we’ll have such capable but unaligned AI systems

  I simply agree, no need to convince me there 👍

  Ought’s approach:

  • Instead of giving a training signal after the entire AI gives an output,

  • Do give a signal after each sub-module gives an output.

  Yes?

  My worry: The sub-modules will themselves be misaligned.

  Is your suggestion: Limit compute and neural memory of sub-models in order to lower the risk

  ?

• Yonatan Cale 11 Aug 2022 14:06 UTC

  2 points

  in reply to: stuhlmueller’s comment on: AMA: Ought

  [debugging further]

  Do you think misuse is a concern—to the point that if you couldn’t monitor and restrict usage—you’d think twice about this product direction?

  Or is this more “this is a small issue, and we can even monitor and restrict usage, but even if we couldn’t then we wouldn’t really mind”?

• Yonatan Cale 11 Aug 2022 13:34 UTC

  3 points

  on: [Cause Ex­plo­ra­tion Prizes] Soft­ware Sys­tems for Col­lec­tive Intelligence

  Attempted TL;DR [please comment if I missed something and I’ll fix it] :

  Create software systems that will replace parts of government and perform better than them, such as interesting ways of voting which are both accessible and give citizens a different weighted voted based on what fits the situation.

  Edit: GG (who is the author, I think) says: It’s more about experimenting towards better solutions. Probably better to read that comment in full.

• Yonatan Cale 10 Aug 2022 15:32 UTC

  11 points

  on: I’m Offer­ing Free Eng­ineer­ing and Con­sul­ta­tion for EA Aligned Projects

  I personally vouch for Soof and hope to work with him on important software projects myself

• Yonatan Cale 10 Aug 2022 13:34 UTC

  2 points

  on: Earn­ing to give should have fo­cused more on “en­trepreneur­ship to give”

  I think that being an employee and opening a startup are such very different career paths that the decision should be made very much based on personal fit.

  (Also, if someone is a good fit for founding a startup, I’d suggest they consider opening or being a leader at an EA aligned org. Not for everyone, but at least consider)

• Yonatan Cale 9 Aug 2022 21:51 UTC

  7 points

  in reply to: stuhlmueller’s comment on: AMA: Ought

  Every tool has some non-zero usefulness for non-central use cases, but seems unlikely that it will be as useful as tools that were made for those use cases.

  I agree!

  supporting high-quality reasoning

  This sounds to me like almost the most generic-problem-solving thing someone could aim for, capable of doing many things without going outside the general use case.

  As a naive example, couldn’t someone use “high quality reasoning” to plan how to make military robotics? (though the examples I’m actually worried about are more like “use high quality reasoning to create paperclips”, but I’m happy to use your one)

  ---

  In other words, I’m not really worried about a chess robot being used for other things [update: wait, Alpha Zero seems to be more general purpose than expected], but I wouldn’t feel as safe with something intentionally meant for “high quality reasoning”

  [again, just sharing my concern, feel free to point out all the ways I’m totally missing it!]

• Yonatan Cale 9 Aug 2022 21:47 UTC

  2 points

  in reply to: stuhlmueller’s comment on: AMA: Ought

  Would you share with me some typical example tasks that you’d give a submodel and typical good responses it might give back? (as a vision, so I’ll know what you’re talking about when you’re saying things like “well specified tasks”—I’m not sure if we’re imagining the same thing there. It doesn’t need to be something that already works today)

• Yonatan Cale 9 Aug 2022 21:42 UTC

  2 points

  in reply to: stuhlmueller’s comment on: AMA: Ought

  Do I understand correctly: “safety in the long run” is unrelated to what you’re currently doing in any negative way—you don’t think you’re advancing AGI-relevant capabilities (and so there is no need to try to align-or-whatever your forever-well-below-AGI system), do I understand correctly?

  Please feel free to correct me!

• Yonatan Cale 9 Aug 2022 21:07 UTC

  8 points

  on: I’m Offer­ing Free Coach­ing for Soft­ware Devel­op­ers in the EA community

  Sharing a few messages from Dan Pandori [6 years of experience at Google and Waymo] with permission.

  Context: I’m trying to figure out the impact of Effective Developers. Dan apparently got value from the CEA hiring post. I’m putting it here for reference.

  P.S

  Dan’s also offering help with preparing for FAANG technical interviews, which I think is awesome

  

• Yonatan Cale 6 Aug 2022 2:47 UTC

  4 points

  in reply to: Guy Raveh’s comment on: arxiv.org—I might work there soon

  I don’t know, I assume it’s done by humans.

  My priors are:

  1. Do something manually before automating it

  2. Talk to the users (medXiv, bioXiv) about their situation before picking a solution

  3. At some point this will be at least somewhat automated, reducing at least most of the human work