Snippets 9: Tyler Cowen interview on talent; the coming Kami-kwasi collapse; what UK politics looked like in 2010

'Push what is falling'

Tyler Cowen interview

TC’s blog is more influential among influential people outside politics than any newspaper.

People in the overlapping worlds of tech/VC/EA/rationalists/YC etc are interested in talent and high performance.

People in politics are not, they’re anti-interested. This was one of many shocks for me 20 years ago when I first got involved in politics. It’s central to why government is so bad. It’s no accident the civil service now treats its own people so dreadfully, while spinning about ‘diversity, excellence, flexitime, investing in people’ etc. It’s a natural product of the culture of elite politics.

Some highlights from a recent TC interview [my comments in square brackets]:

• Travel beats history for learning. [Agree. I spent a few years reading in detail accounts of the outbreak of wars in 1864, 1866, 1870, 1914, 1939 and comparing them — leading history books routinely make simple chronological errors that undermine their own central arguments. My trust diminished a lot. (Often older books seem more reliable than newer ones.) Then I read first draft of history books about events I’ve participated in. Much is invented. These inventions will be used as foundations by future history.]

• Biographies are inefficiently long. [Agree. A great service would be someone mining biographies for examples that fit some very abstract principles of extreme performance. E.g for most people reading Pflanze’s >1000 pages on Bismarck would be a big time commitment and the details of Prussia/European history not interesting to them, but an abstraction of ‘here is how Bismarck illustrates the unrecognised simplicities of effective action’ would be super-valuable. I’m going to do this for Bismarck. When will it be possible to train a GPT Large Language Model to trawl biographies for examples of ‘why speed was critical’, ‘why picking the right person was critical’, ‘how they faced reality when most were fooling themselves’ etc?]

• Talented 15-17 year olds are neglected. They should have less homework, more freedom, some elite tutoring. [Agree. I’ve become more and more anti-normal school for at least the top 2% of IQ, >~130 and even more so for 1:1,000, ~>145.]

• Context is that which is scarce.

• The best talent scouts in general (very rare) are not necessarily themselves super-talented. E.g most great NBA general managers were not great players. The super-scouts have a weird magic-intuition that can probably be trained a bit but is mostly natural. [Alan Kay says great science funders like General Groves (Manhattan Project) or Bob Taylor (ARPA/PARC) are scarcer than Nobel winners. I think this is true more generally about Steve Jobs level CEOs. This talent is enormously underrated in politics/government. If we had such people in crucial political roles the world would look incredibly different but far less of this talent pool goes into public service than in the past, hence the feeling of growing crisis that our ‘leaders’ can’t cope with.]

• Picking talent is inevitably hard, look at VCs who are super-smart, highly incentivised, good feedback etc — they make constant errors. [Few people in government know much about VC-world and its relevant lessons for how the world works including talent scouting. Also the way HR incentives work make it incredibly hard for officials to deal with risk well.]

• Small investments are underrated. [Seems likely. I don’t understand investing including why, given the stats on VC returns (very long-tailed so it’s vital to try to avoid missing the huge hits), they don’t make more smaller bets.]

• ‘Remarkably bimodal distribution’ in Emerging Ventures — most are a clear yes/no. [Interesting. In hiring over the past decade it seems to me that ~95-99% are clear No. The ones I want to hire for politics/government then have to be filtered for whether they’re the type who will try fanatically hard, which you can only be sure about by putting them under pressure and watching. I noticed this week that in an interview by TC with Marc Andreessen, one of the best talent scouts/VCs in the world, MA was asked ‘what data would you most like to have on founders that you don’t have?’ and he replied: video of them under pressure. ]

• Women are better at spotting deceit, presumably for biological reasons (cost of deceit viz pregnancy). [Seems true to me. Like many things I didn’t notice this in my 20s but do a lot now. My wife is like a witch — she clearly sees conversations with senses I don’t have, like infrared vision, that means she constantly infers motives, lies, plots that I don’t see — then some time later what she saw becomes clear. I saw this in No10 too though that environment is unlike a normal good company because the civil service and political parties select so strongly for people who can discern chimp politics over people who can execute effectively.]

• Intelligence has convex returns. But for most jobs IQ isn’t the most important thing. [Seems obviously true. Want to compete with Grothendieck and von Neumann? IQ vital. Want to win an election? More complex. Yes more IQ is going to keep helping if you have a certain personality but won’t otherwise. Look at Kami-kwasi, new UK Chancellor. He’s much smarter than me, probably has the highest IQ in Parliament, he’s also super-lazy, maybe also the laziest MP — the combination of highest IQ and laziest is a fiasco in terms of being good at politics, though being rubbish at politics is also no barrier to climbing to the top of political hierarchies in the modern west.]

• Athletes are much smarter than is realised, this is underrated… Whoever is the best gardener, I’d be super-impressed by. [Seems true to me. I had a very close friend who was a brilliant boxer (won British amateur superheavyweight championship, was going to the Olympics but failed the eye exam). Even though he had terrible schooling and couldn’t talk like an academic, he was fascinating to talk to about the world, more so than almost all graduates I know. He had super-senses I don’t have. 20 years ago I got into a boxing ring with him for 3 x 3 minute rounds. It was fascinating. It’s the only time I’ve thrown up from exhaustion. I obviously expected not to be able to hit him. But what shocked me was the immediate realisation that he knew what I was thinking in real time, literally mindreading. The gap in brain speed was extraordinary. Although you know this must be true, experiencing it is very different from knowing it. As I thought ‘jab’ and started to move, he’d already moved and I could feel the total pointlessness of my action. Then I thought ‘fake then hit’ but he knew it was a fake. After each failure he’d tap me lightly on the forehead with the back of his glove. I never saw anything coming. Highly recommended experience if you can trust them not to injure you! I remembered this years later when I read Colonel Boyd on OODA loops — I remembered that boxing ring and what it was like to have an OODA loop so much slower than my opponent. I tried to apply this in politics. Highly underrated fact about politics: if you can move much faster than your opponent, they start to implode. I saw this with the CBI over the euro 1999-2000, in the 2004 referendum, in the Department for Education, in the EU referendum, in No10.]

• Internet has transformed talent spotting, so much easier to find brilliant 15-20 year olds.

• Not so worried about w*ke [I refuse to use the word and if you think it’s rubbish you should also not use it!], in some ways it’s good — e.g it’s part of why America is open to talent and good at innovation. Europe is very chauvinist, has ‘bad cultural norms for innovation’, too few dimensions of status competition. [Obviously I agree on Europe. I’ve said many times this should be a much much bigger focus for policy people and it was a big part of the reason why I did Brexit. But this is a vicious circle. Policy people in Europe/UK are just generally not interested in innovation and people don’t want to face the reality of why America has done so much better. It’s connected to their lack of interest in high performance.]

• Maybe YC has got a bit stale recently, you hear that. EV unlikely to go stale. Just 2 evaluators. TC not paid so limits how it scales. No staff to be captured by. Low costs of exit.

• Few startups succeed so the competition selects for self-belief/overconfidence. [Will power is a supersense like deceit-spotting. It can change reality, make things possible that seem impossible. It can also cause disasters. Generally for Alexander the Great it was positive but even he stumbled into the Makran desert fiasco. I don’t think there’s any general way to predict reliably whether talented person X is a) ‘again bending reality to fit their will and will win’, or b) ‘they’re refusing to see reality and are heading for disaster’. If it were possible the world would be totally different, the distribution of VC returns would be different, politics would be totally different etc!? I do think you should be cautious when people have a track record. E.g now everybody is predicting Elon will fail with Twitter. But making Twitter work seems to me much easier than egetting either SpaceX or Tesla going. The trickiest part with Twitter is having the will power to push through the political/media shitshow, which Elon seems to have developed, unlike most of his similarly talented elite-CEO cohort.]

• Although I’m good at talent spotting, I don’t think I could do it with novelists. Until you’ve written the novel, it’s hard to see if you could do it. It’s not the same as non-fiction where you can spot able people writing and think ‘they could write a great non-fiction book’. Startups have lots of other early tests. What could those be with novels? [Seems likely he’s right but I don’t know why. Judging whether a novel will be judged brilliant decades in the future is surely very hard/impossible. Tolstoy had supersenses but probably only another brilliant artist has the supersenses to see such supersenses in someone before they write War and Peace? One of the reasons in Jan 2020 I sent out a call for ‘weirdos and misfits’ at home in a William Gibson novel was the desire to have some people with unconventional supersenses near crucial political decisions. My experience in No10 made me more confident that this is a good idea. E.g now you have a set of very similar people trying to predict what Putin will do. How many of them have given orders to kill named people? Almost none, most spend their time at a desk in elite universities etc. How well do they understand mafia like Putin? I think poorly.]

• EA are right that x-risk is neglected and superimportant but are too epistemically confident, they go into all sorts of super-specific scenarios that aren’t relevant to figuring out useful action. They’re under-invested in history and culture. [Feels true. I think also a lot of EA is a) scared of being actually effective, overthink action, and aren’t comfortable with the constant ‘mmm, 70-30, ok do X’ decisions you have to take many times a day in politics, b) are far too worried about their reputations to get involved in the inevitable blood and mud of politics. They think they can keep their reputations intact and have people agree ‘cos reasoning’. This will not work for x-risk as covid showed and UKR shows. If they want to be more effective, they have to take more risks including with their own reputations. Hence I don’t think EA will be very influential or get the best return on the cash. I’ve had academics researching x-risk ask me for advice and flinch from conflict with the leadership of Harvard — how effective are they going to be in persuading D.C?!]

• Chance of civilisation being wiped out is low in any year but hard to see how we avoid it long-term. [Sadly agree. Buffett put it in one of his letters. People intuitively don’t understand that a 1% chance of wipeout per year means the probability of wipeout is about two-thirds within a century. A 1:30 chance per year means a ~97% chance of wipeout in a century.]

• Government has a lot of people thinking about things like UKR/nukes, EAs can’t add much value here. [My biggest disagreement. I think Tyler understands the limitations of politicians much better than almost all academics but still underrates normal incompetence and how much low hanging fruit there is! A priori he would assume that the secret handling of nuclear weapons is much more thoughtful and competent than it is. If it were legal for me to explain the full horror I saw (it isn’t, I won’t) he would be shocked. Also this has got much worse over the past 30 years as part of the general decay of public service and the trend of great people doing other things. Getting the odd super-talented person to do nuclear issues would be close to the best investment you could make, hence, again, my call for weirdos, but this is clearly extremely hard because the old bureaucracies fight to the death against this.]

• After nuclear/whatever-tech war, we probably won’t build back. Crazy wrong to think we can just read the US Constitution then rebuild. We made no progress for thousands of years, we don’t have good theories for why things took off after ~1500 that could reliably be used to rebuild. So then other things like asteroid protection won’t get built. Very daunting! History of mammalian species is pesimistic. We have sex because if you stand still you’re destroyed, sex is the clearly winning model, and it’s a sign that it’s very tough for complex things to survive long. [Seems reasonable!]

• If he only had 5 years to live he’d do more Emergent Ventures. Talent spotting and institution building is more important for him now than writing books.

• Worldviews developed by successful people like Thiel help defend them against meme-idiocy bombarding them from the world. Maybe also help with their work, e.g Thiel spotting Facebook.

• A lot of public intellectuals place big bets on one big idea. Probably not a good strategy. [I suspect brief fame then long decline to near irrelevance is probably the best most of them can hope for but not a good strategy for TC and he’s been right to avoid it.]

• There’s a trade-off of ‘focus on X you really care about and get it done’ and ‘be generally interesting’. Mill was interesting, Cobden and Bright got the Corn Laws done but it’s not very interesting to read them. [Seems right but some people are generally interesting and highly influential — e.g Nietzsche whose influence is everywhere and who also understood our culture better than anybody I’ve read. But I think Nietzsche is better seen as an artist who does what they must do than someone who thinks about optimising strategies!]

• Selection filters for academia strongly select for conformists and they’re not trained to take risks. [Seems clearly true. Cf. the famous Schwinger line: . Innovation in funding could help a bit but more important is to get more of the smartest people out of universities and into different sorts of institutions.]

• Decline in research productivity [e.g physics Nobel now is not as significant as one in 1930]? He’s more inclined to think the ‘low hanging fruit is picked’ argument is more important than ‘regulation is killing research’. He’s probably a bit different on this than Collison. [I’m more with Collison. Because I spent os much time in government watching this and trying to act. I noticed how hard it was for me to remove even things that are so stupid literally not one person would defend it. And spent time talking to scientists crushed by the grant process etc. I suspect TC underrates how much appalling nonsense a top biologist in Cambridge has to suffer from bureaucracy. And by the way, I think TC appreciates this problem better than ~100% of economists.]

• Why do some things in government work well, like Operation Warp Speed? A few people in the system care and they make it better! [Agree. Things like the Vaccine Taskforce happened because some odd people happened to be in No10 and could create a critical mass for it instead of doing the default EU vaccine procurement. This is unusual. The ‘shielding’ program for a million people in covid was another example, an official (Jen Allum) cared unusually about doing the right thing and tried hard to do it. Her colleagues thought she was mad. She has left the civil service and is now at Google in California. Some dreadful colleagues got promoted. This says a lot!]

• Simple question to cut through a lot of babble: you long or short the market? He’s long! In a recent blog, TC added to this question with another to EAs: are you long volatility?

  As many of you know, when I confront mega-pessimists I like to ask them “But are you short the market?” Not once in my life have I heard a satisfactory rejoinder to this query. (The last answer I heard was “I will be!”.)

  I now have a new question for those who see a reasonably high likelihood of AGI. Of course AGI could wreck the world, make the world a whole lot better, or simply overturn multiple sectors of the economy, in both good and bad ways. If you believe in AGI, typically you believe it will matter a lot, though there is considerable disagreement on the cost-benefit ratio.

  So basically you should go long volatility, both of the aggregate market and sector-by-sector.

  Can I call them AGIers?  (“Aggies” is already taken.)  AGIers ought to be long volatility.

  Are they?  I have started asking this question, and I plan to continue the practice.

This is another question. My answer: I’d like to be long volatility but I don’t know how to do it and am reluctant to spend a lot of time learning about investing, I’d like some people like TC to offer a retail product allowing me to be long volatility.

Example: in 2020 I watched Trump talking about the election. I was concerned about it enough that I asked the Joint Intelligence Committee to look at it and talk to the PM. I can’t say what they said but it did not make me less long volatility. (Boris waved away warnings.) Sure enough the election happened and Trump incompetently tried to fight the results, Congress was occupied etc. The political world was mostly surprised. I was not. BUT my financial expectations were completely wrong. I expected stock markets to sink but they kept going up right through the political chaos! In Feb/March 2020 I had been right the stock market would fall and I think what I wrote on here autumn 2021 was also right re the impending debacle on financial markets. But this 11/20 judgement was totally wrong. What is the right product if you want to be long volatility, if for example you think ‘another shitshow coming like Feb 2020, November 2020, Putin/UKR etc’…?

Listen to it all, self-recommending!

Also recommend from his blog this short thing on single payer healthcare.

There is a natural tendency on the internet to think that all universal coverage systems are single payer, but they are not.  There is also a natural tendency to contrast single payer systems with freer market alternatives, but that is also an option not a necessity.  You also can contrast single payer systems with mixed systems where both the government and the private sector have a major role, such as in Switzerland.

I’ll say it again: single payer systems just don’t have the resources or the capitalization to do well in the future, or for that matter the present. Populations are aging, Covid-related costs (including burdens on labor supply) have been a problem, income inequality pulls away medical personnel from government jobs, and health care costs have been rising around the world. Citizens will tolerate only so much taxation, plus mobility issues may bite. So the single payer systems just don’t have enough money to get the job done. That stance is conceptually distinct from thinking health care should be put on a much bigger market footing. But at the very least it will require a larger private sector role for the financing.

Obviously in the UK no party has anything useful to say about the economics of the NHS. The Tories are so lacking in moral credibility anything they might say is a disaster but they’re so braindead they have nothing to say. Labour doesn’t have to have answers to win and is too braindead to develop answers even if they wanted them.

All around me people who can afford it are shifting out of the NHS.

If the Truss team were not imploding so spectacularly and the war, economy etc were not huge stories, the collapse of the NHS would be the main political story. MPs still seem to underrate how bad it is.

---

Truss & Kami-Kwasi: ‘push what is falling’

It’s pointless commenting on all the ins-sand-outs of the meltdown and pointless to try to analyse No10 a any sort of serious agent. All talk of ‘strategy’ is more laughable than ever.

The core issues:

1. Her and her team are so bad it’s close to impossible (i.e 99% confidence) that she can recover and be seen by the public as a decent PM. Because they’ve oriented by the Telegraph opinion pages their model of reality is hopeless. They aren’t even ‘bad at communication’. I know teenagers who would do a better job.

2. Will she be replaced? Conventional wisdom: no. My guess: yes. No10 will continue to screw up amid the global chaos. They have the war and possible nuclear crisis, an economic crises, perhaps a new financial crisis, an NHS crisis, and a political crisis. They will babble incoherently. Plotting will grow. Whenever the tectonic plates shift, a set of MPs and donors suddenly pop up on my phone trying to figure out what I’m up to. They are here again. ‘Hi Dom, hope you’re well, was always a big fan, what you up to, fancy a drink?’ Peston said a few days ago that it’s impossible to ditch her because it looks too mad. I think this gets the logic wrong. It sure would look mad but at some point enough will conclude that it’s even more mad to let someone so clearly unable to do the job and so clearly driving the party into oblivion that the letters will go in.

3. When? 50-50 before Christmas, you can imagine overwhelming events shortly (banks blow up, Putin blows stuff up) mean MPs keep delaying.

4. Another members vote? My guess: No. If they get to the stage of removing her, it will be in a panic. There will be a strong argument for ‘we already look mad, another 3 month contest amid NHS collpase and economic carnage all over the news is unsustainable, no choice but to axe the member vote, if it works we’ll be forgiven, if not we’re dead anyway’. They’ll probably try to get it all done in days.

5. Yes, I know generally I say ‘assume Tories will do the most stupid thing’ and this works well — e.g earlier this year predicting they’d send the letters in then keep him then panic and do it all again, prolonging the nightmare and making their situation worse. Why go against this heuristic? She’s so bad plus there are such awful multiple crises plus the polls are so bad plus the more they see of her the more they’ll know ‘she can’t do this job’ plus the election is no longer ‘years’ away, the long campaign starts next September and the civil service will mostly shut down on anything controversial then (at the latest). So I suspect that panic will overwhelm.

6. Bailey is a disaster too. As I said a week ago, it’s sort of insane the MPs would not have demanded explanations over the systemic risk posed by pension funds’ leverage — except it’s exactly what you’d expect given SW1’s handling of Iraq, 2008, Brexit, covid, Ukraine, China’s infiltration of British critical infrastructure etc. It looks increasingly like the government could be forced into a total u-turn on their tax cuts. Obviously if so, it makes it more likely Truss goes.

7. Given the debacle over the budget, Truss and Kami-kwasi have lost all moral credibility with the country. They will not be able to push through hard spending cuts. They will do what the Tories have tried to do since 2010 except when VL was in No10: cut R&D, vandalise all sorts of valuable work to make their bullshit numbers ‘add up’ (i.e not add up). Of course, this would be anti-growth. But it would be an entirely fitting end to the Tory Party that a set of morons who arrive in No10 babbling ‘growth growth growth’ actually vandalise growth as they collapse.

8. Caveat: I almost never talk to any MPs. It’s possible that a) economic collapse, b) another 2008-like crisis, c) war are so big that MPs feel for a while ‘not now’. Perhaps they’ll hang fire until it seems the worst of the awful winter coming is over. But I wouldn’t bet on it.

I’m a little conflicted on what I want to happen.

On one hand, it’s a great chance to replace them.

On the other hand, the nuclear situation combined with Truss is appalling. Part of me wants rid of her now even though it lowers the chane of Carthage Scenario for this reason. But part of me thinks ‘they could replace her with someone even more likely to make appalling nuclear decisions’ so…

If Tory MPs are watching her on TV explaining blackouts for the first time since the 1970s, and listen to normal families enraged by this and the collapse of the NHS, yes they might say ‘not this week’ but it won’t be long before they say ‘my god enough, though the heavens fall’.

A few other dynamics:

1. Billionaires I know who look at these problems professionally think it’s plausible that most of Europe’s energy companies will be nationalised over this winter, so awful is the energy crisis.

2. They also say: unwinding zero real rates is going to reveal hidden leverage and hidden insolvency all over the European economy, things will blow up.

3. The Eurozone institutions will be under fire again.

4. The proportion of working-age people not in work/available for it in UK is 22 — to 21.7 per cent. There are now 9 million economically inactive people of working age,

I tweeted about this the other day and got a revealing answer from a strongly Remain pundit:

I think a large fraction of political pundits, who as a set are strongly Remain, have fooled themselves about this. They are so desperate to think that ‘obviously Brexit was insane, obviously it will be a disaster’ that they do not want to face the weaknesses of European financial, economic, and political institutions.

Perhaps they’re right and I’m wrong and this winter will see Britain in crises while the Eurozone is fine. We’ll see…

What covid showed is how determined the pundit class is to cling to its world view and simply memoery-hole inconvenient truths.

So I assume that even if I’m right and they’re wrong and the Eurozone does suffer crises this winter, they will not start writing ‘mmm yes Vote Leave did predict this and argue that it’s a reason to lower our exposure to these problems’.

---

Reminder on AGI and expert predictions

Many expert predictions on neural networks have been wrong.

But someone back in 2009 predicted it pretty well. Who? A British guy now working near King’s Cross.

After OpenAI published the original GPT much of academia expressed scepticism that ‘just scaling it up’ would work. Wrong again. And even after GPT3 the same people said that scaling would have to break down. Maybe. But so far the lines have stayed straight. Having spent time in San Francisco recently talking to people building the next versions of these models, I’m very confident that the political and government world will get further big shocks over the next year.

The guy who was so prescient, and who helped build DeepMind, also predicted 50% that we will achieve human level AI by 2028, 90% by 2050, and ‘human extinction will probably occur’.

Pessimism of the intellect, optimism of the will…

---

How SW1 looked in 2010

I came across this which I haven’t looked at in years.

It’s a summary I jotted after a wargame on the dynamics of UK politics in 2010.

It’s a reminder of how day-to-day politics feels very much like one ‘world’ — there’s a set of people and issues and dynamics that seem quite stable — but even a few years later it can seem entirely different, it’s psychologically hard to recover the feelings of the time etc.

Generally almost everything the Cameroons hoped about the next few years did not happen but a) the LibDems collapsed and b) Ed Miliband was so rubbish he couldn’t exploit the Cameroons’ inability to grip power, so c) they clung onto power in 2015 despite having failed in almost every important way to change the trajectory of the country from procurement to MoD to R&D to productivity. School reform was an exception, though having worked at the DfE 2011-14 I would say that in most ways this was despite No10 not because of it (highly underrated fact about how we got so much done: ~15-20 of the top people were replaced in a department that was appallingly managed). They got what they wanted on AV. The Olympics were an unexpected success (I assumed it would be a nightmare and went to Mexico for my honeymoon for a month). As you can see most of the big problems remained big problems.

CHARTS

VC down a lot

We don’t spend much on energy R&D given how vital it is

Historic interest rates since 3000b.c

Non-bank share of credit, 2008 vs 2022