Jonathan Berman

@RodneyTickle I meant to clarify that this is restricted to Asian cuisines.

@JZBerman I’m shocked Italian isn’t higher than Vietnam, but maybe in Pittsburgh I get a distorted view

Thai cuisine is an outlier when it comes to popularity in the US. There are approx 4k Thai restaurants in the US per 100k Thai people living in the US. Next closest Asian cuisine is Japanese at 1.5k / 100k Japanese people. 1/2

@StefanFSchubert What gives rise to this shoddy reasoning? And how essentially is knowing this to combating it?

Many people overrate overestimate analysis of the type “they don’t want more growth/houses” and underestimate of the type “they’re intellectually dishonest about trade-offs”. Shoddy reasoning about trade-offs is a huge problem (and often blameworthy).

-Harvard limiting to 20% As -Princeton going to in-person testing -Arxiv penalizing fake references 1 Implement a trust-based system 2 Observe it works fine at first 3 Enjoy blissful period of denial 4 Confront evidence 5 Make ruies

If you pay them, they will come. Paying talent will attract talent who will then put this money to great use.

Journals shouldn’t care about formatting until acceptance. They should care about openness, transparency, hallucinated references, and other measures of underlying rigor that actually matter—and whose benefits carry over to every submission.

New blog post: The third wave of American philanthropy Hundreds of billions of dollars in new philanthropic capital will soon become liquid. The OpenAI Foundation holds 26% of OpenAI, worth about $220B at today’s valuation. Anthropic’s seven co-founders have pledged to give away 80% of their wealth and have instituted the most aggressive donor matching program for employees in tech history. How much does this all add up to? And how meaningful is that in the context of philanthropy today? I was doing some simple napkin math to wrap my head around the scale of what’s coming, and radicalized myself in the process. I had dramatically underappreciated the scale of the philanthropic capital that’s about to become available and the corresponding gap in talent and organizations that will be needed to make the most of it. This piece aims to directionally sketch the scale of what’s coming, the gap in operational capacity needed to absorb it, and what we can do to fill it. (Link to full post in reply)

new (spicy) post from me: "Economists, mobilize" economics ideas are extremely helpful for understanding AI, but academia is dropping the ball. now is the time for economists to work on the most important problems in AI and to loudly encourage colleagues to do the same.

School holidays across Europe should be staggered, so that Northern countries take holidays in May and June, when the South is at its best and Southern countries takes holiday in July August, when the North is at its best. Crowding is also reduced if tourism is spread out.

I suspect these records (those of people aged 65+) will continue to be smashed over the coming years. As people increasing focus on health and fitness throughout their lives, there will be many many more fit elderly people.

@StefanFSchubert I’d be curious if there are strong feelings about Demis vs Dario

Most reasonable high salience AI leader?

The Shiller P/E ratio is now within 2 points of its all-time high from the dot-com bubble.

AI strategies everywhere hinge on widely available American frontier AI. Post-Mythos, amid compute crunches, security concerns and distillation crackdowns, that paradigm is under threat. Today, I argue the era of widespread access to frontier AI is almost over.

We started out trying to benchmark the AIs... We had experts create the benchmark... we had experts validate the benchmark... ...Then AIs starting doing well on the benchmark ..Now AIs found critical errors in the benchmark itself the humans did not Who is benchmarking who?

@StefanFSchubert Makes me wonder how those people who don't want to believe these things will reject AI's conclusions (e.g., "I think for myself").

Good use case. People underuse AI for these kinds of things. Smart people can put them to a lot of epistemically good uses.

I got my Claude to do a UK version. … Fun prompt. Below is my honest swing at it for the UK. Caveat upfront: "fix everything switch" overstates things — these are policies where I think the case is unusually strong and the standard objections look weak relative to the upside, not magic wands. Confidence ratings are about "this would meaningfully improve UK welfare/growth if well-implemented," not "this is politically easy." 1. Radical planning liberalisation (housing + infrastructure) Confidence: Very high. Reasoning: The UK's discretionary planning system is the single biggest unforced error in British policy. It drives the housing crisis, suppresses productivity (workers can't move to high-wage cities), worsens regional inequality, slows clean energy buildout, kneecaps infrastructure, and arguably depresses fertility. Most rich-country peers build more housing per capita. Moving to a zoning/rules-based system with by-right approval, looser greenbelt around stations, street votes, and curtailed judicial review on planning would be transformative. Why not done: Homeowners are an enormous, high-turnout voting bloc whose paper wealth depends on scarcity. Greenbelt is politically sacred. The losses are visible and concentrated (a new block next door); the gains are diffuse and slow. Both main parties have flirted with reform and flinched. 2. Council tax revaluation and property tax reform Confidence: Very high it's needed; high it would be beneficial. Reasoning: Council tax bands are still based on 1991 property values. A flat in Hartlepool can pay more council tax than a mansion in Westminster. This is regressive, distorts the housing market, and starves councils. Revaluing — or better, replacing council tax and stamp duty with a proportional property tax or land value tax — is one of the most-endorsed-by-economists reforms going. Why not done: Revaluation creates visible losers (London/South East homeowners) and quieter winners. Treasury also fears the transition. The Mirrlees Review laid this out in 2011 and basically nothing has happened since. 3. Abolish stamp duty on primary residences Confidence: High. Reasoning: Stamp duty has among the worst deadweight losses of any UK tax — it taxes transactions, so it discourages downsizing, labour mobility, and matching people to suitable homes. The empirical literature on this is unusually unified. Why not done: It raises ~£12bn and is invisible to most voters most of the time. Replacing it (ideally with #2 above) needs political capital nobody wants to spend. 4. Build clean energy and grid infrastructure fast Confidence: High. Reasoning: UK industrial electricity prices are among the highest in the OECD, which is quietly hollowing out manufacturing. Cheap, abundant electricity is the most undervalued input to almost every modern industry. Streamlining onshore wind, solar, nuclear (SMRs and large), and especially grid connections and transmission would compound for decades. Connection queues currently stretch into the 2030s. Why not done: Planning (see #1), Ofgem incentives that historically rewarded keeping grid investment low, judicial review, NIMBYism around pylons, and a regulatory culture that treats new build as risky and the status quo as safe. 5. Devolve real fiscal and policy power to cities and regions Confidence: High. Reasoning: The UK is one of the most fiscally centralised democracies in the OECD. Manchester, Birmingham, Leeds, Glasgow are dramatically underpowered relative to peers like Munich or Lyon. Mayors with proper tax-raising power, transport authority, and skills budgets would let cities tailor solutions and compete. Most evidence on agglomeration economics suggests this matters a lot. Why not done: The Treasury genuinely hates losing control of money; this is a deep institutional preference that survives changes of government. Westminster MPs don't like rival power centres either. 6. A genuinely competitive high-skilled immigration regime Confidence: Medium-high. Reasoning: Top-talent immigration is close to a free lunch for receiving countries — fiscally positive, innovation-boosting, and concentrated in tradeable sectors. The UK has the language and the universities; it's leaving money on the table with slow visas, high fees (the NHS surcharge plus visa costs for a family is brutal), and a Global Talent visa that's narrower than it looks. Why not done: Immigration politics is toxic and the public doesn't reliably distinguish high- from low-skilled flows. Total numbers have been high recently, which makes it hard to argue for more of any category even when the composition matters more than the total. 7. Childcare cost reform Confidence: Medium-high. Reasoning: The UK has among the highest childcare costs in the OECD as a share of household income. This depresses female labour force participation, suppresses fertility, and acts as a giant marginal tax on the second earner. The recent expansion of "free hours" helps but the supply-side problem (provider economics, ratios, planning for nurseries) isn't fixed. Why not done: It's expensive, the sector is politically diffuse, and ratios/quality regulations have genuine trade-offs that aren't easy to legislate around. 8. NHS productivity reform plus sustained capital investment Confidence: Medium. Reasoning: NHS productivity has fallen meaningfully since 2019 even as funding rose. The UK has unusually few CT/MRI scanners per capita, old buildings, and weak digital infrastructure. Capital investment with productivity-linked reform (better IT, more diagnostic capacity, GP/secondary care interface) is probably higher-ROI than another marginal headcount increase. Why not done: The NHS is genuinely a political third rail; structural reform gets framed as privatisation regardless of content. Capital is the easiest budget to cut in any given year, so it gets cut repeatedly. 9. Curtail judicial review of major infrastructure Confidence: Medium. Reasoning: HS2, Heathrow, nuclear, transmission lines — UK infrastructure costs are 2–4x peer countries partly because of endless legal challenges and the consultations that anticipate them. A narrower "did the decision-maker follow the process" standard, with tight time limits, would lower costs without sacrificing core rule-of-law protections. Why not done: Constitutionally fraught, lawyers and the legal establishment push back hard, and any government doing this is accused of authoritarianism — sometimes fairly. 10. Reform the pensions triple lock and the wider intergenerational settlement Confidence: Medium-high on the need; medium on whether reform alone "fixes" much. Reasoning: The triple lock ratchets state pension spending up faster than wages or prices and is a major driver of long-run fiscal pressure. Combined with NI exemption for pensioners and protected pensioner benefits, the UK has built a fiscal regime that systematically transfers from young workers to older asset-holders. Untangling this would free fiscal space for investment in #4, #7, #8. Why not done: Pensioners vote at roughly 75% turnout; under-30s at roughly 50%. The maths is unforgiving. A few honourable mentions I cut: tax simplification (genuinely good but rarely "fix everything"), proportional representation (large effects but unclear sign), drug policy reform (good evidence base, narrower scope), regulator growth duties / a serious bonfire of Quangos rationalisation, and reforming the apprenticeship levy. The pattern. Almost every item above is blocked by the same underlying thing: a system that makes losses visible and concentrated while gains are diffuse and slow, in a country where homeowners and pensioners vote heavily. The "switch" in the meme exists; it's just wired to a circuit that the median voter has a strong short-run reason to leave alone. That's not a reason not to flip it — it's a reason the case has to be made repeatedly and honestly. Standard disclaimer: reasonable people disagree on all of this, especially #6, #9, and #10, and "well-implemented" is doing a lot of work in every line above.

It's an unimpressive-sounding word, but one of the most powerful motivations is the motivation of the hobbyist. That's what keeps successful founders working on their companies long past the point when they've made enough to quit. It's their beloved project.

If highly profitable, disruptive technologies create negative externalities (e.g., job displacement, environmental damage) it is reasonable to require the winners to contribute, via taxation, to mitigating those harms.

British leadership contests are baffling. From a citizen’s perspective, all the twists feel entirely due to vibe shifts—disconnected from any tangible reasoning or well-thought goals.