All of NunoSempere's Comments + Replies
deliberately, consciously
I think the source you mention is talking about people deceiving themselves
stomach-churning thing of all is that CFAR organized a summer camp for kids where, according to one person who was involved, things were even worse than at CFAR itself
Idk man I think this summary is a few shades more alarming than the post you are taking as evidence.
I think doing it successfully would take too large a chunk of my time, but I was considering it in case Sentinel fails
Some questions which feel alive for me:
- Should we expect risks to come from categories and places we expect to see coming, or from places we would not have anticipated afterwards? What's the proportion of black swans for the largest risks?
- How do we incorporate a term for doing good in a way that helps us do more good in the future? Companies can sell stock, whereas nonprofits can't; the more Elon Musk does the more money he makes, even while pursuing some notion of the good, whereas the more a philanthropist gives away the less he has. This seems like a strategic disadvantage. This is more like an operational decision than like a research question though.
Here is a reading list I made for a new hire at Sentinel. I think it does a good job at capturing the promise, need, yet limitations of forecasting which I've found over the last few years. Suggestions of items to add welcome.
From our writtings:
- How Likely Are Various Precursors of Existential Risk?
- Global Risks Weekly Roundup, a few weekly editions
- Making sense of Israel’s strikes on Iran
- Our WIP fundraiser
Fundaments of Probability Theory
- Probability Theory: The Logic of Science, by E.T. Jaynes, particularly the first two or three chapters on probability as a
If 3H is false but we act urgently, this false positive is far less bad, as we will have many years (maybe millions) later in which to invest resources for the real hinge of history.
But you lose the compounding, particularly if later generations make the same calculus, and so you can't implement something like a Patient Philanthropy Fund. https://www.founderspledge.com/programs/patient-philanthropy-fund
Distribution rules everything around me
First time founders are obsessed with product. Second time founders are obsessed with distribution.
I see people in and around EA building tooling for forecasting, epistemics, starting projects, etc. They often neglect distribution. This means that they will probably fail, because they will not get enough users to justify the effort that went into their existence.
Some solutions for EAs:
- Build a distribution pipeline for your work. Have a mailing list on substack. Have a twitter account. This means that
I had reason to come back to this comment. Rereading it, I don't think I'm exactly wrong, but I'm not paying enough face, enough respect to the challenges of running an organization, and so the bar that I am setting is in some sense inhuman. These days if I wanted to give similar feedback I would do so in private, and I would make sure it is understood to come from a place of appreciation.
Executive summary for this week's global risks roundup:
...Top items:
- Geopolitics: Russian jets entered Estonia’s airspace. An agreement between Pakistan and Saudi Arabia could bring Saudi Arabia under Pakistan’s nuclear umbrella.
- US Politics: A US comedian’s show was suspended after the FCC chair exerted pressure on his network and on companies that own local TV stations. The Trump administration plans to announce actions targeting left-wing groups.
- Tech and AI: Open-source AIs were used to design variants of a simple virus genome, eliminating the need for
I followed up with classic 2019 EA forum post Aligning Recommender Systems as Cause Area with a podcast with Ivan Vendrov here: https://x.com/NunoSempere/status/1965820448123629986
Here is an endpoint that takes a google doc and turns into a markdown file, including the comments. https://docs.nunosempere.com. Useful for automation, e.g., I downloaded my browser history, extracted all google docs, summarized them, and asked for a summary & blindspots.
Here is the executive summary and few sections for this week's brief on global risks, by my team @ Sentinel.
- Geopolitics: Trump and Putin met in Alaska to discuss the Ukraine war. Forecasters’ estimate of the chance of a ceasefire by October dropped from 27% pre-summit to 9%.
- Biorisks: The chikungunya virus continues to spread, including in France and the UK.
- Tech and AI: Meta’s policies explicitly allowed its AI chatbots to “engage a child in conversations that are romantic or sensual.”
- And more: Three soldiers were killed and four others injured in a d
Instead of letting the social fabric collapse, everyone suddenly turns their ire on one person, the victim. Maybe this person is a foreigner, or a contrarian, or just ugly. The transition from individuals to a mob reaches a crescendo. The mob, with one will, murders the victim (or maybe just exiles them).
Maybe this person is a contrarian, but Girard also argues that the scapegoat effect is greater if the person is like any other member of the public, because then this scares the participants more because "it could have been me"
I think if you focus on conflicts it's just smaller than other conflicts:
- Russia/Ukraine war: ~150-250K killed on the Russian side, 50K to 100K killed on the Ukrainian side; source
- Sudanese civil war: "Likely significantly more than 150,000 total killed[44] More than 700,000 children with acute malnutrition[45] 8,856,313 internally displaced 3,506,383 refugees[46]" per Wikipedia. Also at risk of famine
Here is a related tweet.
I agree that with the risk of famine losses rise to, potentially, 2.1M. But the whole population of Gaza being killed seems very unlike...
(I might delete this post later if it derails the thread, I’m not sure how useful or constructive it is—please let me know!)
I think this inadvertently highlights why an ‘EA’ (utilitarian) framing might downplay the badness of the conflict. I think the badness falls into five buckets (opinions follow; and not claiming that you do or don’t agree with these):
- The direct number of deaths and ongoing suffering is high.
- The systematic bombing and killing of civilians, coupled with the eradicationist rhetoric of the Israeli government, almost certainly constitut
I think the Sudanese civil war is a relevant comparison. I'd take the typical EA point to be something like:
"If Western diplomats spent as much time as they've (ineffectively) spent trying to avert famine / improve aid in Gaza as ending the war in Sudan, it seems like there would have been much more progress -- fewer dead, starving, maimed, irreparably emotionally harmed."
Or one level deeper, there are probably conflicts that are not yet happening that we could decrease the likelihood of and that are probably even more neglected. I'm thinking of Ethiopia a...
FWIW the latest estimate I heard from Gaza was 100,000 dead (many of which haven't been reported by Hamas) (sorry for the paywall) which is on the same order of magnitude - and as opposed to the Ukraine war, most of them aren't combatants. It's up to you what to make of that.
My team at Sentinel produces a weekly brief on global risks. Here is the executive summary and forecasts for this weeks:
...Key items this week are:
- Economy and trade: Trump announced new tariffs. US GDP growth is driven by AI capex spending while labor slumps. In response to an unfavorable labor report, Trump fired the commissioner of the US Bureau of Labor Statistics.
- Geopolitics: UK, France and Canada announced their intention to recognize a Palestinian state. Hunger is widespread in Gaza and might meet the technical definition of famine later this year, whic
We're looking into this as part of Sentinel. I agree it looks unlikely. Manifest at 3%, though I think it's even lower than that because of the low return over ~2 years https://manifold.markets/embed/NuñoSempere/will-anthropic-go-bankrupt-or-be-di
There is an additional set of beliefs which EAs share, which is something like: EA institutions are a good mean to pursue those goals, their leaders worth deferring to, their forum worth using, the brand worth expanding, the community worth recruiting people into. I think this is important, otherwise it risks stolen valor and e.g., counting Bill Gates as "an EA", and it risks eliding the ways in which the community is very dysfunctional.
I personally had a small crisis when I came to believe that the Center for Effective Altruism wasn't cost-effective. I wr...
people reporting very low P(doom) numbers just don't understand the AI alignment problem
... or disagree with the frame.
Another selection effect is that people who are more interested in xrisk will tend to differentially participate in these forecasts.
https://forum.effectivealtruism.org/posts/EG9xDM8YRz4JN4wMN/samotsvety-s-ai-risk-forecasts might be of interest.
Some related thoughts: https://nunosempere.com/blog/2024/12/04/grain-of-truth-memo/
Also related: VC-backed companies pivoting being fairly common (but much less common with nonprofits, maybe since the relationship with funders is capacity constrained and predicated on a particular theory of impact)
My team has published estimates for precursors of xrisk and has weekly updates that usually contain forecasts. Could be of interest, not sure if that's the kind of thing you are asking about.
Phil Trammell has some good work on this topic, here and here.
...Therefore, a patient philanthropist can typically do more good (from her perspective) by investing for the sake of future spending than by spending immediately. She should only begin spending under two circumstances. First, she should spend once she, and any other patient funders in an area she wishes to support, have grown wealthy enough relative to the area’s impatient funders that even the impatient are spending at less than the patient-optimal rate for the collective (patient plus impatien
It's a great question. I looked into it a few weeks ago:
...In the current political praxis of the United States, there are various workarounds that the government can pursue in order to take illegal or unconstitutional actions. Examples of such actions that previous administrations have taken include keeping prisoners in Guantanamo Bay in order to delay habeas corpus petitions, conditioning highway funding to force state compliance, and torturing the interstate commerce clause of the Constitution to regulate local matters. The point is that the US has some co
As a personal anecdote I was spending a few hundred bucks per month last year and now I'm spending a few thousand so the cost I'm paying is definitely rising :)