Milan Weibel

• Milan Weibel 10 Jul 2022 15:14 UTC

  4 points

  0 ∶ 0

  on: A list of EA-re­lated podcasts

  Un equilibrio inadecuado (Spotify—Apple Podcasts—Google Podcasts)

  Interviews in Spanish on EA topics. I particularly enjoyed the episode with Andrés Gómez Emilsson from Qualia Research Institute. Sadly, no new content since October 2021.

• Milan Weibel 20 Dec 2022 0:00 UTC

  3 points

  1 ∶ 0

  on: MORE Effec­tive Altruism

  Interesting. I agree that second or third-order effects such that as the good done later by people you have helped are an important consideration. Maximising such effects could be an underexplored effective giving strategy, and this organization you refer to looks like a group of people trying to do that. However, to really assess an organization’s effectiveness, epecially if it focuses in educational or social interventions, some empirical evidence is needed.

  • Does SENG follow-up on the outcomes of aid recipients?

    • How do they compare with those of similar people in similar situations, but who didn’t recieve help?

  • What programs does SENG run?

    • How much does each cost per recipient helped?

ChatGPT un­der­stands, but largely does not gen­er­ate Span­glish (and other code-mixed) text

An as­pira­tionally com­pre­hen­sive ty­pol­ogy of fu­ture locked-in scenarios

• Milan Weibel 11 May 2023 2:34 UTC

  4 points

  0 ∶ 0

  on: RIP Med­i­cal Debt

  EDIT: Other commenters have pointed out reasons why the elimination of debt sold really cheap is unlikely to affect much the lives of recipients. Still, if the debt relieved did in fact significantly help the beneficiaries, it could turn out to be very effective. However, we won’t know until RIP releases recipient outcomes data.

  TL;DR: About as cost-effective as GiveWells’s top charities, IF my assumption about outcomes is broadly right. $14.16 to provide debt relief to one person. If one assumes a lifespan increase of 0.2% (less than two months) as the effect (by preventing healthcare avoidance), it comes out to $7080 per death-equivalent-in-lifespan averted. I recommend looking further into it, particularly with respect to outcomes.

  Hi Layla, welcome to the Forum! Thanks for posting!

  This looks like an interesting opportunity. Within the cause area of health in the US, RIP seems to have chosen a big and tractable problem, and to be triaging their beneficiaries according to the relevant metrics.

  Here is my attempt to have a rough idea about RIP’s cost-effectiveness.

  RIP claims that it has “helped 5,492,948 individuals and families” and has relieved $8,520,147,644 of medical debt. The average debt relieved per recipient is thus $8,520,147,644 /​ 5,492,948 = $1551. If, as you say, “every $100 donated clears $10,000 in medical debt”, then the cost per recipient is $15.51 (!!!).

  I was initially skeptical of this calculation, but it checks out. In its 2021 year end report, RIP says that it relieved debt to 1,312,697 people during the year, and in its 2021 financial statement declares total expenses of $18,587,272. So the cost per recipient is $18,587,272 /​ 1,312,697 = $14.16.

  It’s hard to estimate the benefit from medical debt reduction. Let’s say, for the sake of simplicity, that the avoidance of medical treatment and mental health problems derived from struggling with medical debt make people live 0.2% shorter lives (1.92 months if starting out with an 80-year lifespan), and that the debt relief provided eliminates that effect. It follows that preventing 0.002 death-equivalents costs $14.16, and thus preventing one death-equivalent unit of lifespan reduction costs $7080. This is about as cost-effective as GiveWell’s most recommended charities.

  This would be huge if true. However, my priors advise me against getting too hopeful. It should be hard to find a charity about as cost-effective as GiveWell’s top charities. RIP has been assessed by Charity Navigator, and does a fair bit of marketing. It would be weird if no EA had picked this up before. I have reason to believe that I am overestimating the positive effects of debt relief.

  To find out whether RIP is really so effective, it would be great to have numbers on the welfare outcomes of debt relief. I found this report on RIP’s site, which while a potentially useful qualitative source, makes no effort to quantify outcomes.

• Milan Weibel 15 May 2023 3:06 UTC

  2 points

  1 ∶ 0

  on: EA and AI Safety Schism: AGI, the last tech hu­mans will (soon*) build

  A more pessimistic counterargument: Safely developing AGI is so hard as to be practically impossible. I do not believe this one, but some pessimistic sectors within AIS do. It combines well with the last counterargument you list (that the timelines where things turn out OK are all ones where we stop /​ radically slow down the development of AI capabilities). If you are confident that aligning AGI is for all practical purposes impossible, then you focus on preventing the creation of AGI and on improving the future of the timelines where AGI has been successfully avoided.

• Milan Weibel 15 May 2023 3:09 UTC

  9 points

  3 ∶ 0

  on: EA and AI Safety Schism: AGI, the last tech hu­mans will (soon*) build

  Side note: calling a world modelling disagreement implied by differences in cause prioritisation a “schism” is in my opinion unwarranted and (low-probability, very negative value) risks becoming a self-fulfilling prophecy.

• Milan Weibel 15 May 2023 15:13 UTC

  1 point

  0 ∶ 0

  in reply to: Milan Weibel’s comment on: EA and AI Safety Schism: AGI, the last tech hu­mans will (soon*) build

  Just for the sake of clarity: I think the word “schism” is inaccurate here because it carries false connotations of conflict.

• Milan Weibel 15 May 2023 15:16 UTC

  2 points

  1 ∶ 0

  on: AI safety logo de­sign con­test, due mid-May

  PSA: The form accepts a maximum of 10 files, that is, 5 design proposals maximum (because each proposal requires uploading both a .png and a .svg file).

• Milan Weibel 23 May 2023 13:50 UTC

  1 point

  0 ∶ 0

  on: How I learned to stop wor­ry­ing and love skill trees

  Intriguing. Looking forward to the live demo.

• Milan Weibel 16 Nov 2023 22:05 UTC

  1 point

  0 ∶ 0

  on: Milan Weibel’s Quick takes

  In a certain sense, an LLM’s token embedding matrix is a machine ontology. Semantically similar tokens have similar embeddings in the latent space. However, different models may have learned different associations when their embedding matrix was trained. Every forward pass starts colored by ontological assumptions, an these may have alignment implications.
  
  For instance, we would presumably not want a model to operate within an ontology that associates the concept of AI with the concept of evil, particularly if it is then prompted to instantiate a simulacrum that believes it is an AI.
  
  Has someone looked into this? That is, the alignment implications of different token embedding matrices? I feel like it would involve calculating a lot of cosine similarities and doing some evals.