#6 Regime Change: new data shows Trump beating Biden and beating AOC & Kamala more easily - plus DeSantis matchups

Another crazy election cycle with added craziness of AI deployed (maybe by China & Russia as well as both campaigns) & crazy reporting about AI

Unda fert nec regitur [you can ride a wave but not make one]… State socialism is on the march and there is no stopping it. Whoever embraces this idea will come to power. Bismarck

The world revolves around the inventors of new values, it revolves invisibly. But the people and fame revolve around actors… Far from the marketplace and from fame happens all that is great, far from the marketplace and from fame the inventors of new values have always dwelt… [T]he democratisation of Europe is an involuntary arrangement for the breeding of tyrants. Nieztsche

After the midterms, a small team did some research to figure out some dynamics for the 2024 Presidential campaign, primaries and general.

We built a model (with many times more data than normal polls) to predict who would win the electoral college in different matchups. We explored attitudes to possible Presidential candidates. We explored crucial issues like the cost of living, health, abortion, crime and Ukraine. We tested some ideas for how a Presidential campaign could best launch — what are the most important waves to ride and avoid, how on crucial issues do you navigate between your core vote and the crucial swing voters in crucial states? We ran focus groups in crucial states including Arizona, Georgia, Michigan, Nevada, Pennsylvania, Wisconsin. These included a) Obama>Trump voters (a fascinating group), b) no vote 2016 > Biden 2020, c) no vote 2016 > Trump 2020 voters, d) hardcore MAGA voters, e) non-MAGA-GOP voters. All data was collected before the Trump-arrest story got going seriously. His position with GOP voters has improved since we collected data. (We also experimented with what we called ‘synthetic focus groups’ and ‘synthetic polling’, i.e running focus groups and polls with ‘synthetic’ voters inside Large Language Models, see below.)

We made more effort than is made in normal polls to ensure we had an accurate sample of low-education, low-trust voters who do not pay much attention to normal news (exploring factors like attitude to vaccination and trust in media). Since 2016 a failure to sample these voters properly has been an important factor in the failure of many pollsters, connected with the failure to appreciate the importance of graduates shifting left on many issues. E.g In 2015-16 in Britain and America many pollsters did not weight by education. An important reason why we and Trump won is that we and Jared Kushner built better data science operations and therefore understood crucial voters (and turnout models) better than the old parties or the old media. With both Brexit and Trump, polling errors reinforced Insiders’ other errors. Despite the stakes, public polling in 2020 in America was also badly wrong but it got less attention than 2016 because Biden won.

Below are some data and a little analysis. I’ll add to it over the next few days then make it public after feedback.

Media organisations are free to use this data provided you make clear where it came from and link to this blog. (I can’t explain why I did this nor will I publish everything or the underlying data.)

The sooner Insiders wake up to the possibilities for much worse chaos over technology and politics in 2023-24 than the aftermath of 2016 — when we had widespread misinformation about misinformation believed by Insiders to this day — the better. Better for politics and better for AI which could easily be subject to extremely misguided regulation amid panic — e.g beneficial applications (cancer, drugs etc) being slowed by rules that do nothing to stop bad actors. I am pessimistic but, ‘pessimism of the intellect, optimism of the will’…

There’s no going back. And there’s good news — the HR-pundit complex will get decimated, many of the worst voices in the old media will soon cry out then fall silent.

As Marc Andreessen says, don’t expect any sort of ‘going back to normal’ — now is as normal as it’s ever going to be, it’s going to get weirder and weirder.

Ye, voice of the AI renegades…

---

Trump wins the electoral college against Biden but it’s very close

In our model Trump loses the popular vote by about 2 points but Trump wins the electoral college 294-244.

Trump won very narrowly in 2016 because he won Michigan, Pennsylvania, Wisconsin — total margin ~78k votes. These states were close again in 2020 and Trump won more votes here than in 2016 but Biden won them narrowly. Biden won Arizona by 10k, Georgia by 12k, and Wisconsin by 21k. These three states, total margin of ~43k votes, nudged Biden over 270. This was much closer than public polls and models suggested. And Biden won ~7 million votes more than Trump overall, though Trump got more votes than Obama.

In our model Biden’s vote in Georgia, Michigan, Pennsylvania, and Wisconsin falls just enough that they flip back to Trump.

But NB. in our model these four states are all very close and a 1-2% change gives Biden over 270 again.

Trump wins by more against Kamala and AOC

Trump wins easier against Kamala:

And wins easier against AOC:

Kamala is already laughed at a lot in focus groups. And there are so many videos of her waiting to be made into ads I think she would self-destruct and Trump would win easily against her. AOC would be tougher for reasons I’ve explained before (much better at politics, communication, building a team, plus smarter and ruthless).

DeSantis beats all three easier than Trump does

Against Biden, DeSantis also flips Georgia, Michigan, Pennsylvania, and Wisconsin (a little easier than Trump does) plus he flips Arizona.

Buttigieg beats Trump and DeSantis

But it’s really difficult to become a very well known politician in America. It’s important that neither Buttigieg nor DeSantis are nearly as well known as Trump and Biden who have been national figures for decades and President.

So their numbers are noisier than Trump’s and Biden’s. In particular, for complex reasons related to name recognition and how you model voters who don’t know who Buttigieg is, we think these numbers for Buttigieg are roughly best case for him. (And remember Buttigieg could have a big problem in primaries given his trouble winning black voters in the south.)

We did not model Newsom but we did focus group him. He is better known than AOC or Buttigieg but in a negative way. Scandals have stuck to him. And California is widely regarded, from the Silicon Valley billionaires to the swing voters of Wisconsin, as a basket case. ‘Trump was right we should build a wall … around California’ is a repeated idea in focus groups particularly because of the national coverage of California’s homeless problem and the way in which San Francisco authorities have given up on crime (follow the excellent @garrytan if you want to see a lovely, smart technology professional trying to rally a network to do something about the horror of local SF politics). So although Newsom’s far left politics are in tune with DEM donors and activists, I would guess from our research (plus my assessment of him as a candidate) that he would be easier to beat than AOC.

(The candidate most core DEM voters want is Michelle Obama but all indications are she won’t run. We also looked at some non-political figures. While some may be viable, Trump has queered the pitch for many such outsider campaigns. To have a chance of winning the Presidency most would, like Reagan, have to spend time as Governor somewhere to build credibility.)

Why does Trump win?

Trump maintains a very strong grip over GOP voters. It took a small hit after the midterms but has gradually recovered — yes, improved since last November.

And for most GOP voters beating the DEMs is more important than whether they agree with Trump on everything or like him. So when GOP voters who never liked Trump or have gone off him or prefer DeSantis are faced with a choice of ‘Trump or Biden’, they mostly still rally behind him.

The two parties are close to 50-50 so any contest starts off close. Relative enthusiasm is important (and did for Hillary) and the decisions of the small number of swing voters are important.

Biden is not popular and enthusiasm is low even among DEMs and people who voted for him. (His approval is almost exactly where Trump was at this stage but Trump’s approval was uniquely stable.) This level compared to previous Presidents does not suggest he can’t win — Reagan, Clinton and Obama bounced back from similar levels. But it does mean that unless he improves, other things equal the 2024 election is likely to be very close.

Biden average approval, FiveThirtyEight

Biden average approval, Civiqs

Biden doesn’t need to slip far in the close states to lose the electoral college even if he is millions ahead in the popular vote. The trend by which a DEM candidate wins states like California by more and more but loses the electoral college can easily continue.

Why does DeSantis win by a little more?

Like Trump he consolidates the core GOP cote.

He has better net favourable ratings than Trump.

He does a little better with the small number of swing voters.

Some voters are less likely to bother voting for Biden if DeSantis is the candidate.

People want change and Trump offers it — and people are angry about the economy and think Trump did a good job on it

A fundamental aspect of why Trump won in 2016 was because he was the change candidate.

This is why he won over a crucial set of voters who voted for Obama.

For pundits, the idea of voting Obama > Trump seemed crazy.

But when you realise why the voters thought it consistent, it makes sense.

Don’t believe me?

Listen to research from Obama’s own team. They also wanted to know what had happened in 2016 so they went back to research Obama > Trump voters.

Voters who chose Obama in 2012 and Trump in 2016 were not sending a signal nor repudiating the President in any way. Rather, many felt those choices were consistent by voting for a candidate who promised to shake up the political system and bring change to Washington. Many were highly conflicted in their 2016 decision, but ultimately chose Trump because they concluded that Hillary Clinton was too flawed as a candidate and would not deliver on promises to move the country forward. (Quoted in Lucky, italics added.)

Also bear in mind that although Trump is presented by the media as an extremist, in 2016 he was seen as a more moderate GOP candidate than the normal. E.g he explicitly rejected decades of Republicans campaigning on Social Security. (There’s a podcast where David Shor says this and Noah Smith’s eyes widen and roll around in his head, so weird is it to think of Trump as the more moderate candidate.)

I think Trump missed huge opportunities to win in 2020 and alienated some of these voters.

But he still is a candidate who can credibly promise real change to a country and crucial demographics who are desperate for it. And the cost of living crunch is making people more desperate for change.

And Biden is a weak politician. Again, this is not controversial. There is a reason his campaigns in the 1980s and 1990s flamed out. There’s a reason David Plouffe, Obama’s brilliant campaign manager (his book is one of the very best on how to run a winning campaign), advised Biden not to run in 2016. And there’s a reason Obama staff such as Axlerod have said publicly they think Biden should not run again and the DEMs should embrace an open competitive primary contest starting ASAP. Biden was never a good candidate and he’s now visibly ‘frail’ at best.

After the midterms there was a concerted push by the old media to talk Biden up but our research makes clear that result was not because of enthusiasm for Biden’s ideas or achievements — it was because Trump made a big mistake in focusing on the ‘stolen’ 2020 election rather than the economy, which all demographics agree is his strongest card. Even some Democrats who hate Trump say things in focus groups like ‘I voted for Biden but gotta admit Trump did a good job on the economy and look at it now’.

Even those Republicans who agree with Trump that 2020 was stolen want him to move on. As they say in focus groups, they talk to friends and family and they hear people say the same thing over and over: I want to move on from 2020, the country wants to and needs to move on from 2020.

This is a central issue for 2024.

On crime and the border, the DEM leadership in DC & mainstream media are out of touch and this gives GOP a big opportunity

DEM donors, activists, politicians, plus the old media like the NYT are a powerful network.

This network herds powerfully on most important issues, particularly cultural issues. Twitter automates the herding, accelerates it, and helps police it.

This network has a set of views on crime and the border.

These views are in tune with a set of younger, richer people who went to the best universities. They are out of touch with most of the country.

Consider these two perspectives:

(A) Stop early release for violent and sex criminals. More police on streets. Take back control of our border. Stop illegal immigration.

(B) Stop sending so many to jail, especially for minor drug offences. Reduce sentences. We need more immigrants. Amnesty for illegals. Stop ICE abuse. 

It depends on the exact wording but a) 85-90% of GOP are (A) and b) somewhere around a third to a half of DEM voters agree with MAGA voters on (A). This adds up to a clear majority. And if you look at older voters, i.e more likely to vote, there is even stronger support for the ‘tougher’ approach. Of >55s, only about a fifth support the more ‘progressive’ position on crime.

A fundamental fact is that DEM elites (and a large section of mainstream conservative elites!) are far out of tune with the voters and do not want to face this fact.

How bad is this problem?

I’ve spoken to Obama staff who say they are shouted down if they voice these fears in private discussions of DEM strategy.

This ought to be a red light flashing danger for the DEMs but because of how Insider networks operate it is not. Wrong ideas are reinforced by the network structure.

Big questions for 2024 will therefore be:

Can the GOP make crime/the border a core focus?

Can the DEM nominee neutralise it or will they be pushed by activists and donors towards campaigning irrationally (like Hillary on immigration in 2016)?

NB. Exact wording in focus groups and polls obviously matters a lot so don’t read much into the precise numbers above, the important thing is this is a very clear finding and winning tough elections is much more about finding such big clear important things than small nuances. I also think you don’t get to the truth on these subjects just by looking at poll data, you have to spend time talking to voters and watching them respond to each other and policy ideas, ads etc. When you watch normal DEM voters discuss crime you see many are much closer to MAGA voters than you’d think looking at mainstream analysis.

The imminent meltdown of Insider-world over AI

I’ll do a separate post on AI shortly updating what I wrote in March — 2022 was to AI/LLMs as Jan-Feb 2020 was to covid in the sense that, as I wrote last year, you could see the debate about to jump from very niche to mainstream.

Here I’ll just say something about the relationship of the debate to the 2024 election.

Could AI be used to generate strategic advantage in politics and elections?

Without doubt. We used it to improve prediction of the true critical voters in 2016 (but not to improve the execution of digital marketing, per the Cadwalladr conspiracy) and the true critical voters and true marginal seats in 2019. Competent campaigns everywhere could already, pre-GPT, use AI tools to improve performance.

We did some simple experiments last year to see if you could run ‘synthetic’ focus groups and ‘synthetic’ polls inside a LLM. Yes you can. We interrogated synthetic swing voters and synthetic MAGA fans on, for example, Trump running again. Responses are indistinguishable from real people as you might expect. And polling experiments similarly produced results very close to actual polls. Some academic papers have been published showing similar ideas to what we experimented with. There is no doubt that a competent team could use these emerging tools to improve tools for politics and perform existing tasks faster and cheaper. And one can already see this starting (look at who David Shor is hiring).

It’s a sign of how fast AI is moving that this idea was new last summer (I first heard it discussed among top people roughly July), we and others tested it, and focus has moved to new ideas without ~100% of those in mainstream politics today having any idea these possibilities exist.

But that’s not the most crucial question. 

It’s already possible to use ancient wisdom to generate strategic advantage to win elections and change history. As history repeatedly shows, the nature of politics and its real incentives makes it extremely hard to absorb such wisdom then implement it without getting killed by your own side. For agents in the top 1% of ability (or 1 in 100,000) the really difficult thing is not figuring out how to defeat ostensible opponents, it’s acquiring then maintaining control of a centralised power network in the teeth of perpetual internal enemies, while also dealing with the problems thrown up by the world. (E.g beating Trump or Biden in 2024 is probably easier than getting into a position to beat Trump or Biden.) The gains from implementing the unrecognised simplicities of effective action (e.g real clarity of priorities, really fighting the entropy of talent in large organisations) come mostly over longer time horizons than the costs and problems. This is why these simplicities are, as Charlie Munger says, ‘unrecognised’ and why we live in a world where there is perpetually vast amounts of apparently low-hanging fruit in politics that are incredibly hard to pick and eat.

So the more interesting question is: which team (with what sort of skills) will create/use these new tools to gain strategic advantage in navigating this problem?

Progress here would be truly world-shaking, far more important than the usual questions asked about AI in politics.

At the moment most attention is on the mostly-silly subject of misinformation, itself a legacy of the Cadwalladr-conspiracy going mainstream 2016-17. A huge fraction of ‘misinformation’ academics and hacks who write about it are themselves sources of misinformation, either through incompetence, ludicrous trust in false things they see in the old media, or political bias. But misinformation is a relatively trivial issue relative to the likely near-term implications of this technology.

This suggests we’re headed for an ugly, chaotic irony…

The ‘AI was used in 2016 to steal wins for Trump and Brexit’ meme was powerful and influential and false. But 2024 will be the first election in which AI could be a significant factor.

What if it is used by Trump and Biden? What if it’s used by Putin?! What if it’s used by Biden, Trump, Putin, China and other players like MBS who want to see Trump win?!

Many players around the world regard Biden as senile, the Democrats as insane, and the DC system — together with European capitals — spinning out of control towards more war. The Insider-Twitter network has been effective in policing views inside western elites but its very success is generating counter-forces outside the West. In spring last year the White House and No10 were briefing ‘China will join the West’s sanctions’ and even though this seemed obviously ludicrous to me the message was amplified across Insider-Twitter. The opposite has happened. And look at how recently China popped up with a remarkable deal between MBS and Iran — an under-reported event because it undermines conventional wisdom on the West’s Ukraine policy. It’s clear that dynamics unleashed by Ukraine, sanctions, the US seizing Russian assets, Biden’s shift against the multi-decade bipartisan approach to Taiwan and so on has sparked many countries to recalibrate attitudes towards NATO countries.

Perhaps many powerful players will conclude — if Trump (or DeSantis) wins, then no wars, a deal on Ukraine and back to ‘One China’ on Taiwan, DC’s madness is more limited to DC rather than cascading across the world, and everybody makes more money. Powerful incentives!

One core problem with the 2016 Facebook story was that almost nobody in politics, media or academia understood communication, digital marketing, how Facebook actually worked and how ‘AI/data science’ could be connected to these subjects.

But a deeper problem was and is: most of the old political media and Left-academia aren’t actually interested in the truth, they just repeat whatever crazy nonsense they see on Twitter and if their social network believes it so do they! A small but telling recent example: look at how they swallowed the ‘Andrew Tate arrested because of his twitter spat with Greta’ story, classic fake news reported as fact across old media and swallowed whole by the Insider twitter network that generates Official Truth because it was simple and striking propaganda with a moral twist that the audience wanted to believe. Everywhere you looked, Harvard and Oxbridge graduates, so keen to write op-eds and tweets about how non-graduates are suckered by Putin-Trump ‘misinformation’, tweeted their delight at the story. This is normal. Another example from the last few days — London and Washington immediately lied to the media over the Pentagon leaks and these lies were immediately believed by much of Insider-world (from The Times):

So — given how crazy the Insider network went over 2016-Trump-Brexit-Putin when the core claims were false, imagine the hysteria when some core claims may be true! 

This will happen in an environment in which top political pundits are generally bad at explaining the use of TV in politics — which has been central for ~70 years — so there is almost no chance of them explaining extremely new technology. And an environment in which technology leaders in America privately describe the relationships and trust between the tech sector and Washington as ‘all time lows’.

(Ps. A BBC reporter, self described as ‘BBC North America tech reporter. I try to explain tech clearly and simply. Based in San Francisco’ goes to interview Elon and crashes incompetently — good example of how MSM ‘tech reporters’ are often charlatans and why much of the Valley has turned against the MSM.)

Some thoughts

1. These electoral college numbers are obviously not a prediction of what will happen in 2024. The point of a model like this is to figure out which demographics are leaning towards whom, where, why and so on.

2. The numbers show that the mainstream media assumption that ‘Trump is finished’, which they pushed after the January 6th riots and after the midterms, is wrong. Crucial states for the electoral college remain very close.

3. The Democrats have a very hard problem. Biden is a weak candidate. He answers questions about Ukraine saying ‘Fallujah’. His own voters are not enthusiastic about him running again. Given his age and mental impairment he could easily suddenly deteriorate. And this could give Trump another win. (Biden could campaign sedately from his basement in 2020 because of covid but won’t be able to do that again. Presidential contests are brutal tests of endurance.) But if the DEMs replace him, perhaps in panic after a sudden deterioration, someone could win the primaries who is easier for Trump to beat.

4. A huge amount in tough elections boils down to the fundamental questions people think the election is about. In 2016 Vote Leave tried to make the choice: do you want to stop sending £350 million per week of taxpayers’ money to Brussels and spend it on the NHS instead, and do you want to hit a big red button forcing politicians in London to TAKE BACK CONTROL of our border and swerve the EU letting in more countries, like Albania and Turkey, with free movement? In 2019 Vote Leave tried to make the choice: do you want to GET BREXIT DONE so the country can move on and we focus on health, crime, science, technology, higher real wages and reforming government, or do you want to spend years more on a second Brexit referendum then maybe a third — do you want to spend the next few years arguing about Brexit with Corbyn as PM or do you want to focus on other things?

5. But there are huge pressures that pull campaigns away from the objectively optimal strategy. As I have said many times, one of my basic principles of politics is: politicians generally do not, contrary to the standard theory, optimise for *winning elections*, they actually optimise for a) immediate feedback from the old media and social media, and b) signaling to in-group factions, which they think are critical for their career prospects on a very short timescale (days to months). Feedback from elections is much slower, distorted and often genuinely hard to interpret. If you assume they’re optimising for ‘win the election’ the news and campaign activity will often seem weird, if you accept my rule they will make more sense. In every political effort I’ve been involved in (the euro, education reform, Brexit, 2019, No10 generally), figuring out how to beat the ostensible opponent has been 100X easier than persuading my ostensible ‘own side’ to follow the optimal strategy.

6. This explains why, for example, Cameron/Remain could not counter Vote Leave on ‘£350 million for the NHS’ and ‘Turkey is joining the EU’. And it explains why the winning strategy for Vote Leave in 2016 was extremely hard to execute. Core aspects of the Vote Leave strategy were strongly opposed by most MPs and supportive media: they hated £350m/NHS (and largely refused to amplify it), hated ‘Turkey is joining’, and hated the slogan TAKE BACK CONTROL. They wanted the campaign to be about what they had talked to each other about over decades — free trade (irrelevant to most voters) and what they called ‘Global Britain’ (incomprehensible even to other Insiders never mind voters). Market research showing their ideas didn’t work was simply dismissed (‘not what I hear on the doorstep’). Few learned. In 2019 many MPs still wanted to campaign on ‘Global Britain’. Then, in 2020, after winning in 2019, when we didn’t need to listen to their useless ideas on communication, Boris actually adopted Global Britain as a slogan, even though he (sort of in half his brain) knew it didn’t work. Why? To signal to the Telegraph and a set of MPs! In 2016, Hillary could have chosen to focus voters’ choice on issues where crucial voters agreed with her. Instead, she agreed with Trump that support for immigration was a central question. This path made her popular with her activists and donors and much of the old media but was disastrous. This was not an isolated aberration: her campaign spent many millions on ads that really pleased donors and activists but objectively helped Trump.

7. For the old media there is another important incentive: money. ‘Trump may not be good for America, but he’s damn good for CBS’, as Leslie Moonves put it. This complicates figuring out what’s going on. a) Most political hacks at CBS, NYT etc oppose Trump. b) They increasingly see it as justified for them to act as campaigners, including spreading false stories and suppressing facts that are seen to help Trump. c) Trump is good for ratings and NYT subscriptions so they have a financial incentive to amplify Trump. d) Much of what the old media says in its attempts to harm Trump actually helps him, so the more they amplify the more they can help both their own profits and Trump. e) The fact most of the hacks are on Twitter intensifies their campaigning role, their personal incentives to be dishonest, and their delusions reinforced by the network. f) The blatant way in which Insiders and the old media shift from one false-coordinated-line to the next has generated strong counter-forces, much of which are hidden from the mainstream Insider network (hidden because the latter don’t want to see it but also because people are keeping their views and roles hidden). This seems to be leading more and more previously quite mainstream elements of elite networks to views adjacent to ‘almost anything that throws a wrench in the DC/NYC madness-spiral, even Trump, is good’. It adds to the chances of weird dynamics helping Trump or DeSantis.

8. What does research suggest about the fundamental focus of 2024? The DEMs need the focus to be: Trump’s obsessed with 2020 plus health and social security. The GOP needs the focus to be: the economy, violent crime and taking back control of the border. Most DEM voters are closer to GOP voters than they are to DEM activists and donors in their core concerns: they do not want the election to focus on race, sex and the running insanity generated by elite universities.

9. Perhaps everything will hinge on whether Trump listens to good advice or wings it on his own instincts again. If Trump focuses on the 2020 election, he is likely to lose. If he focuses on the economy, his record, Biden’s failures, and sets out a credible plan for how he’ll turn it around again, he could win. Certainly if he runs the optimal campaign I think it would be super-close and close enough for another contested election in which Trump may have reasonable arguments that he’s been cheated. Trump’s hiring is his biggest weakness. Bannon. Giuliani. His string of rubbish lawyers. The best hope for the DEMs is this run continues. 2024 will be close enough that the campaigns matter, unlike, say, 1936 or 1984 — and perhaps three state campaigns persuading 50,000-100,000 total voters will be decisive again. If Trump persuades Jared Kushner to run it and let’s Kushner pick the team and follows advice, I would bet on Trump winning. (I don’t think this will happen.) But even if you had a) Trump campaigning in character minus the worst 10% of unnecessary trolling / burying his own message focused on the economy plus b) a really good third party/PAC effort focused on subsets of voters in six states, I’d bet on Trump to beat Biden. (And, per above, imagine if foreign forces decide to run their own efforts in these six states. In 2019 the DEMs incompetently blew up their own Iowa primary. The US is not up to handling a competent foreign state intervention. Imagine the news chaos if a foreign intelligence service starts running a LLM-based campaign, with the old media trying to describe what’s happening and claiming ‘the CIA says this is Russian/Chinese misinformation’.)

10. Most Insiders don’t want to hear it but the economy remains a very strong card for Trump and the way policy on Ukraine has worsened the cost of living makes it stronger. A powerful GOP message in 2023-24 will be: ‘Spend American taxes building American schools and hospitals, and taking back control of our own border, instead of sending our money to turn Ukraine into rubble.’

11. Most attacks bounce off Trump with GOP voters. In focus groups with Trump supporters, if you propose some attack on Trump you will hear over and over again statements like ‘just more fake news’, ‘yeah, like Hunter Biden’s laptop is Russian misinformation and not a true story? Haha as if I’m going to believe this.’ The only attack that really works at all with these voters is: ‘it’s crucial to win in 2024, Trump is obsessed with 2020 and his campaign will be another crazy circus, but the country wants to move on from all the chaos and division and the media will do anything to stop him, he can’t win, so we should get behind DeSantis instead so we win’. (No other GOP figures seem to have more than a Hail Mary chance.) 

12. DeSantis has much higher name recognition and is more liked by GOP voters than a year ago. His midterm victory obviously helped nationally and with GOP primary voters. He has a path but it is hard to find and execute. A basic problem: if he persuades Trump fans that he’s good, the reaction is — ‘yeah, sounds good, OK so he’s mini-Trump, but … if he’s mini-Trump, why vote for him, why not stick with the real deal, especially since Trump’s got the experience, he knows how the swamp works and how to beat it, and we know what a great job he already did on the economy?’ The best path for DeSantis is: a) build an economic plan and a story bringing together GOP primary voters, entrepreneurs, swing voters behind productivity growth and higher real wages; b) a credible political story of how he will change Washington; c) a political strategy reaching beyond MAGA with a story to the country then using this success to tell GOP primary voters: ‘you can see I’m more likely to beat Biden than Trump, here’s how the GOP and the country move on, come with me if you want to win’. So far he has failed to capitalise on Trump’s blunder of focusing on 2020 in the midterms. If the GOP primaries were now, Trump would comfortably win. Trump retains a strong hold over the GOP base and his advantages remain underrated.

13. The New York Times, Washington Post and other bastions of the old media are no longer trusted by Republicans even to report basic facts honestly. There’s extreme polarisation on the axis of ‘do you trust the news media’: the GOP wins by a landslide with those who don’t, the DEMs win by a landslide with those who do.

14. If you’re trying to figure out who will win in 2024, the normal 1k sample national polls reported by the media will add noise not signal. Follow Nate Silver instead. What’s needed to fix the polling errors of 2016 and 2020, and do better than Nate Silver, is a lot more high quality data from the ~5-10 potential swing states in the hands of people who know how to use it. Even I am surprised by how little interest there seems to be from the old media in investing in this given the potential returns. After all, they are actually interested in polls, unlike, say, procurement and pandemics! And I’m more surprised by how little interest there seems to be even from organisations that are genuinely expert in data like hedge funds. (Perhaps it’s because ‘Trump is finished’ has been so widespread?)

15. I don’t expect either side to try to talk to the voters in the way you would if you were really optimising for ‘win the election’. I think that there’s an opportunity to exploit a) the madness of DC/NYC and b) technology to run a much more effective campaign by actually taking policy and how Washington works much more seriously than is normal in a Presidential campaign. But I must admit I’m in a tiny minority in thinking this and even if I’m right it would take an unusual team, and some risks, to execute it.

A final observation.

Listening to normal voters over months it’s striking, as it always is, how differently they see politics to the discussion between Insiders on Twitter.

The world of mainstream pundits and academics treats Trump voters as if they are extremely closed-minded and ‘trapped in narrow information bubbles’. But listening to them it’s striking how open-minded they are, and willing to discuss opposing perspectives, compared to the mainstream pundits and academics! It’s also striking to me that Trump voters are more open to discussing things from, say, CNN than most DEM voters are to discussing things from FOX.

But more importantly, poorer people who don’t watch much news are generally much more open-minded about politics than graduates living in big cities who consume a lot of news, who are much more ‘trapped in narrow information bubbles’ than the average GOP rural voter who pays little attention to politics. And pundits and academics are the most closed-minded of all while thinking of themselves as the opposite. They herd to a few acceptable opinions but think they’re the few able to step outside herding and observe objectively. Another golden rule of politics is that it’s the intelligentsia who are easiest to fool with simple moral propaganda tales. (Thanks to Twitter you can build a Twitter list and observe this phenomenon in real time!)

These dynamics repeatedly generate huge opportunities for political entrepreneurs. But they are, and must remain, mostly unseized because the short-term cost of exploiting them is, at best, serious tension/conflict with powerful networks. If this were not the case, if it were easy to exploit herd consensus, then it would happen more often and the world would look very different.

Having watched Washington and New York since 2016 the most reasonable prediction is that most Insiders will continue to tell themselves fairy tales and peddle misinformation to each other while thinking it’s the MAGA plebs who are the victims of misinformation. There will be incredible dislocations between Insider debates over 2024 election and what’s really happening, what normal voters actually hear and prioritise. If you pour the petrol of the war and ‘AI/KGB/PRC/MBS’ stories onto the fire, it could easily be even crazier than 2020.

Further reading

Ian Hogarth wrote this recently in the FT to explain some of the core developments to Insider-world. Ian is an investor in some of the key companies and knows most of the key players so it’s an actual AI-Insider lifting the lid on actual AI-Insider discussions for political-Insiders.

David Deutsch podcast on AI/AGI.

Conor Leahy podcast on AI/AGI.

I wrote this in 2017 about Allison’s book on China/America/Thucydides and how Taiwan could be the 1914/Belgium. It’s relevant as Biden disastrously ditches the old ‘One China’ policy now denounced by many Insiders as ‘Chamberlain appeasement’ but supported by that President described by the old media as ‘extreme’ and ‘mad’, Ronald Reagan.

Some AI stuff:

• March 2023 re politics scrambling to catch up after Chat-GPT

• Jan 2020 ‘weirdos and misfits’ appeal

• 2018 Judea Pearl and causation

• 2018 re AI, Brexit, regulation

Previously on US:

• November 2022

• September 2021

I strongly recommend Jared Kushner’s book, Breaking History. It’s an unusual DC memoir. It makes a strong case for Trump’s successes, such as progress with Middle East peace by doing things very strongly against the conventional wisdom of the foreign policy bureaucracies. It’s also interesting how similar his experiences were to No10’s on covid, deciding to hugely accelerate vaccine and other technology development etc.