Tom Walton-Pocock

Tom Walton-Pocock@tomwalpo · 2026/06/12观点输出

RT @yoramdw: Some charts you want to see go down 😉 The average non-cash equity stake taken by UK universities in spinouts has fallen from…

0400

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Tom Walton-Pocock@tomwalpo · 2026/06/11观点输出

I see Jevons is now doing the rounds Walpo's Paradox: Most paradoxes are not in fact paradoxes

106124

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Tom Walton-Pocock@tomwalpo · 2026/06/04赛道影响

RT @London_Inst: "With equipment like this, no wonder we need £60 million." Our endowment campaign continues with a new ad in the @FT today…

0300

AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。

Tom Walton-Pocock@tomwalpo · 2026/06/03赛道影响

Perhaps the biggest single-policy booster to UK GDP and wellbeing available right now. Chambers is one of the few star politicians operating in the UK right now. He's going after big-impact policymaking that sits in a large unserved consensus between left and right.

001121

AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。

Tom Walton-Pocock@tomwalpo · 2026/05/13观点输出

Proposal to make the UK the global leader in IP mobility and accelerate our growth https://t.co/ALRf8GTcfQ

1111727

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Tom Walton-Pocock@tomwalpo · 2026/05/07赛道影响

For a sense of the irrepressible energy at LIMS, listen to LIMS Fellow @YangHuiHe1's interview by @TOEwithCurt. Yang's curiosity and hunger to know totally dominates any fear that AI will put him out of work. When I first met him, it was the day of the FrontierMath benchmark release in December 2024. I expected him to be, at the very least, a bit shaken by the gains LLMs were making. Not a bit of it: "Why? I'm so excited! Do you know how many Chern classes are fully solved? One! I need infinitely many postdocs today." Yang's actually been at the forefront of AI-assisted mathematics since 2018, and you can listen to his story here: https://t.co/tofvrUhFRc

005333

AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。

Tom Walton-Pocock@tomwalpo · 2026/05/07政策影响

A new maths institute formed quietly in London in 2011. It's based in Michael Faraday's apartment in the Royal Institution in Mayfair, right at the heart of the city. But somehow, almost no one's heard of it. About LIMS The London Institute for Mathematical Sciences (LIMS) @London_Inst is the leanest edition of a research institute you can imagine: fifteen researchers, brought in from every corner of the world, and insulated from the distractions of university admin. Founder Thomas Fink spent years defying economic gravity to keep the mission - to make frontier discoveries in basic science and mathematics. More than twice, it was nearly lights-out for the Institute (though in fact, LIMS runs mostly on candlelight). After years of uncertainty, the Institute's future is now secure. @bendelo has anchored an endowment that will see this pure-research faculty make big advances in fundamental science. Research and AI Whilst AI lifts all boats, and it's beyond amazing that undergraduates have unlocked Erdös problems with a single incantation in a language model, I think that's how the low hanging fruit is picked off. It seems unlikely that "solve Riemann now plz" does the trick on GPT-6 release day (I know you, like me, will try it anyway). For the major breakthroughs - to see Riemann solved or Langlands completed - I suspect you need to watch LIMS and its fellow "supernode" institutes (Max Planck, Simons, IAS, etc). Joining LIMS I'm proud to be a Governor at LIMS, and if you're ever in Mayfair, you should drop in on one of their events. Some famous figure in maths or physics is almost certainly lecturing there when you're next in London.

4116686

AI：偏政策推进，强调合规落地与规则明晰，对监管主线更敏感。

Tom Walton-Pocock@tomwalpo · 2026/04/11观点输出

@matthewclifford You should speak with @createstreets Matt

001210

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Tom Walton-Pocock@tomwalpo · 2025/12/19观点输出

@VitalikButerin Reminds me of the Appel and Harken proof of the four colour theorem - indigestible by humans, was it really a proof?

00069

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Tom Walton-Pocock@tomwalpo · 2025/12/06赛道影响

@mert @jonnny_in You don’t need ZK for integrity, you can have replication for that, though it is highly centralising for FHE (big ciphertexts / costly computations). Succinct proofs would fix that; but my understanding is succinct proofs of FHE state transitions remains infeasible.

00128

AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。