Stefan Schubert

Our book on the psychology of effective altruism is now out. The digital version is available for free: academic.oup.com/book/56384?l... There's also an audiobook. podcasts.apple.com/us/podcast/e... And you can save 30% on a physical copy if you use the code AUFLY30. global.oup.com/academic

We intuitively underestimate how intertwined common sense and social science are. Social science is shot through with common sense, and conversely, it has gradually changed common sense in ways we often fail to see. www.update.news/p/social-sci...

Many people are optimistic about the potential for further progress in psychology, but I think they underestimate the obstacles, as well as how much of the low-hanging fruit has already been picked. www.update.news/p/the-limits...

Why are our explanations of other people’s actions often so cynical? In today's post I lay out a series of causes and argue that this cynical bias is akin to our neglect of human agency. www.update.news/p/the-roots-...

Existential risk estimates depend on both the difficulty of the problem (what it'd take to prevent disaster) and our likely response, but the first factor gets most attention. Here I argue that's a mistake. www.update.news/p/the-respon...

Dan Williams recently argued that pessimism about human motivation – especially about overcoming selfishness – is more warranted than epistemic pessimism. I agree, and here discuss the psychological and evolutionary roots of this asymmetry. www.update.news/p/why-were-e...

My friends and I made this path thirty years ago by larping in the ditch. I'm glad to see it has survived.

Genuinely interesting and important take

Chesterton's gate

I wrote a post about cultural differences between psychology and economics -- things were I specifically believe econ is doing better. Surprisingly, I still stand by those points 8 years later. Somebody from econ should write a counterpart though about things that psychology does better 🤭

Danish paper finds big economic gains from GLP-1 drugs. “We find that GLP-1 treatment reduces long-term sickness leave by 17.3%. We estimate total fiscal benefits of GLP-1 initiation of approximately 1.3–1.5% of annual labor income per employed individual.” www.nber.org/papers/w35475

The UK used to be much more unequal than Sweden, but the Gini coefficients have almost converged. Perceptions are lagging.

I gave a talk on the case for epistemic discipline at EAGx Nordics. When you encounter people who don't even try to give arguments, it's better to calmly explain that this is the case than to respond sarcastically or try to reinterpret them charitably. www.youtube.com/watch?v=mTml...

“We’re good at solving problems, but not as good at seeing how other people solve theirs” @stefanschubert.bsky.social explores the Rationality Gap. It may be more than merely “not seeing” others’ capability, eg a WYSIATI-like, ignorance-flavoured underestimation: buff.ly/7qE6E7i

We're often much more rational as actors than as outside observers of other people's actions. It distorts our worldview, making us think we sleepwalk into disasters and bad outcomes. www.update.news/p/the-ration...

Shootings have fallen more than 2/3 in Sweden the last three years

You can’t dismiss AI revenue growth as hype the way you might dismiss investments or high valuations. www.update.news/p/the-tables...

AI sceptics have long dismissed short or medium timelines as speculation with no basis in data. But recent revenue growth undercuts this argument. It's now the AI sceptics who have to show that current trends will break. www.update.news/p/the-tables...

AI sceptics have long dismissed short or medium timelines as speculation with no basis in data. But recent revenue growth undercuts this argument. It's now the AI sceptics who have to show that current trends will break. www.update.news/p/the-tables...

People love sharing the OkCupid chart where women find most men unattractive, and the rumour that one in ten children has a different father than they think. But some stories are too juicy to be true. www.update.news/p/women-dont...

Enjoyed the linked post by Alex Imas: "If the model is right, the durable jobs of the future won’t be about monitoring AI systems or prompt engineering. ... . The durable jobs will be in the relational sector, where the human element is the product itself." aleximas.substack.com/p/what-will-...

AI propaganda has so far not been as effective as people feared a few years ago www.update.news/p/ai-economi...

AI propaganda has so far not been as effective as people feared a few years ago www.update.news/p/ai-economi...

There's been another good debate among economists about the impact of AI on the labour market, this time centred on a nice essay by Alex Imas. But I think some economists still underestimate how weird the world of advanced AI is going to be. www.update.news/p/ai-economi...

Why clinical trials are broken & how to fix them: a reading list forum.effectivealtruism.org/posts/wsMxed...

Sweden and German intelligence agencies report that Russian official economic data is fake www.ft.com/content/04a9...

Good news everyone, at least it's a draw!

When ChatGPT was launched, almost all users were men – but now, women are the majority www.update.news/p/will-ai-dr...

Anthropic just passed $1 trillion in valuation on Ventuals. OpenAI still has the most users but have fallen behind in terms of valuation. www.update.news/p/nature-stu...

The debate over AI and jobs is too nearsighted, and should consider where we're eventually heading. AI leaders talking about widespread automation isn't a mere marketing strategy. www.update.news/p/growth-won...

The list of Swiss constitutional amendments triggered by citizen-initiated referendums is strikingly populist. www.update.news/p/american-c...

Incredible turnout in Hungary, especially in Magyar-voting Budapest

Magyar now 80% likely to replace Orban as PM of Hungary according to Polymarket

Anthropic overtakes OpenAI on Ventuals, a (small) market for private company valuations

American students are moving toward higher-paying college majors www.update.news/p/american-s...

Clinical trials today are highly bureaucratic, expensive, and time-consuming. Most trials fail to recruit enough participants to answer their questions at all. This great post by Adam Kroetsch explains how things got this way. www.clinicaltrialsabundance.blog/p/clinical-t...

Might AI gradually become less politically salient, like the internet? I don’t think so, since unlike the internet, AI will change drastically with time. www.update.news/p/driverless...

I think revenue extrapolation may be a bit underrated as a method for predicting transformative impact from AI. While it certainly has its issues, I think it comes with fewer contentious assumptions than some rival methods. Recent revenue growth speaks against very long timelines.

Why is public discourse so much more AGI-pilled in the US than in Europe? I discuss five reasons: 1. Higher AI adoption rates 2. The leading labs are in the US 3. Cultural openness to speculation 4. A thinner safety net 5. A larger public discourse www.update.news/p/why-does-t...

Why is public discourse so much more AGI-pilled in the US than in Europe? I discuss five reasons: 1. Higher AI adoption rates 2. The leading labs are in the US 3. Cultural openness to speculation 4. A thinner safety net 5. A larger public discourse www.update.news/p/why-does-t...

The British government tries to get Anthropic to expand in the UK www.ft.com/content/6bfd...

Philosophers cite more empirical sources than they used to www.update.news/i/192954532/...

The reactions to this does not reflect well on this site. Some people need to update fast.

There's been a debate on how “intelligent” AIs could become, but however you define that concept, they could hugely outmatch us at manipulating the world. www.update.news/p/how-smart-...

There's been a debate on how “intelligent” AIs could become, but however you define that concept, they could hugely outmatch us at manipulating the world. www.update.news/p/how-smart-...

The AI safety community would benefit from more epistemic modesty. www.update.news/p/how-to-dis...

While social media is polarising, evidence suggests AI may nudge people towards the centre. www.ft.com/content/3880...

The AI safety community would benefit from more epistemic modesty. www.update.news/p/how-to-dis...

Some argue that Trump helps the European left – a "reverse Midas" effect – but this is overstated. His demands for Greenland didn't stop the Danish right from making gains in yesterday's election. www.update.news/p/right-wing...