Ananyo Bhattacharya
@Ananyo · London · joined 18 Feb 2009
Chief science writer @London_Inst. 'The Man from the Future', on the unparalleled influence of John von Neumann, available everywhere. For my substack see link.
16 961Followers
6 645Following
28 322Posts total
38.6KViews on collected posts
Latest posts
Looking for new posts…
This is going to be awesome! Come if you're in London and do maths!
869 views · 9 likes
· 0 reposts · 1 replies
Open on X →
A special seminar 2pm next Thursday 24 Sep!
In the spirit of deepening human understanding in an age of AI maths, our Fellow Nikon Kurnosov, will unpack the recent proposed solution to the six-dimensional sphere puzzle by Claude, guided by Levent Alpöge.
https://t.co/6RPjH07eNE
1.6K views · 4 likes
· 1 reposts · 0 replies
Open on X →
James Phillips, former special adviser to the UK Prime Minister: “We know that there is now not much time, we considered this before we started writing this. Have we just left it too late? Honestly, maybe, but as we say above, we have to try.” https://t.co/MaPniseDLn
683 views · 5 likes
· 2 reposts · 0 replies
Open on X →
Leading mathematicians have written to the President of the Royal Society arguing the rapid progress in maths by AI may be mirrored in more dangerous areas like cybersecurity.
Whatever you think of that, Britain seems ill-prepared and has under-invested. https://t.co/uqL5H1etOC
512 views · 3 likes
· 1 reposts · 1 replies
Open on X →
In a new paper, Prof. Yang-Hui He and colleagues show how computer vision can distinguish elliptic curves from random data and predict analytic rank by reading arithmetic patterns in digital images. https://t.co/SE6odYImcI
697 views · 5 likes
· 1 reposts · 0 replies
Open on X →
@Ananyo @Nature @London_Inst @GoogleDeepMind interesting but you only discuss the positive aspects of AI
277 views · 2 likes
· 0 reposts · 1 replies
Open on X →
Co-authors: @MikhailBurtsev, @YangHuiHe1, Evgeny Sobko at @London_Inst and @ThoreG -- our governor and Distinguished Research Scientist at @GoogleDeepMind. Full article free to read at this link: https://t.co/8CKIuInZq4 10/
806 views · 11 likes
· 2 reposts · 3 replies
Open on X →
Finally: Solving and verifying results. Hat tips here to AlphaEvolve and @OpenAI for the unit distance problem. Accelerating progress here but "For now, the decisive creative leaps are still made by humans. The real promise lies in partnership." 9/
770 views · 6 likes
· 0 reposts · 1 replies
Open on X →
Next: Proposing conjectures. "AI generates many conjectures, most of which are trivial, previ- ously known results or false. Human experts still decide which conjectures are worth pursuing." 7/
529 views · 6 likes
· 0 reposts · 1 replies
Open on X →
The next step--linking AI-generation of conjectures with agenda-setting. Rather than working blindly, AI systems could first map the existing body of maths knowledge to identify bottlenecks, gaps and unexpected parallels, and then generate conjectures to bridge them. 8/
533 views · 6 likes
· 0 reposts · 1 replies
Open on X →
We also give the Feynman path integral as an example of unformalised but massively useful device. This sort of very human, intuitive and transformative leap forward is also something we've yet to see from AI.
AI will not "solve" physics but could accelerate physics. 6/
553 views · 7 likes
· 0 reposts · 1 replies
Open on X →
Next: Formalizing ideas. "Even the most accomplished mathematicians can benefit from a system that insists every inference be made explicit." Reducing human labour here would lead to larger bodies of verified maths, which in turn could be used to train better AI models. 5/
581 views · 7 likes
· 0 reposts · 1 replies
Open on X →
First: Setting the agenda. "One of the most distinctly human acts in research is deciding which questions are worth asking in the first place." But AI can help! Future tools might scan OEIS or ArXiv, for example, finding new connections and structural parallels between fields. 4/
653 views · 7 likes
· 0 reposts · 1 replies
Open on X →
We argue that theorists have nothing to fear and much to gain from AI. "The task now is to build these systems with care and ambition. If they can make the frontier more navigable — and more deeply interconnected — they will accelerate discovery, not replace discoverers." 2/
858 views · 8 likes
· 0 reposts · 1 replies
Open on X →
We split the research pipeline into four overlapping phases: setting the agenda, formalizing ideas, proposing conjectures and solving and verifying results. This is not perfect or definitive but it's a useful way to assess progress and future challenges. 3/
746 views · 7 likes
· 0 reposts · 1 replies
Open on X →
Our @Nature comment this week on the use of AI in maths and theoretical physics - and why the community should embrace it!
Authors @London_Inst & @GoogleDeepMind.
First draft 8 months ago but edited many times as the field steamed ahead!
Free-to-read link at the end of 🧵1/ https
27.4K views · 278 likes
· 64 reposts · 14 replies
Open on X →
Against accounts of the same size
6 posts from the last 90 days, next to the 10K–100K follower range. ordinary reach for its size, weaker reaction than most.
Median views690this account924median for 10K–100K
Reach, %4.07%this account3.62%median for 10K–100K
Engagement, %0.99%this account1.52%median for 10K–100K
| Metric | This account | Median for 10K–100K | Ratio |
|---|---|---|---|
| Median views per post | 690 | 924 | 0.75× |
| Reach (views ÷ followers) | 4.07% | 3.62% | 1.12× |
| Engagement rate | 0.99% | 1.52% | 0.65× |
Others in this range → Compare with another account → How these benchmarks are built →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
6538 Jun
581
553
533
529
770
806
2779 Jun
69716 Sep
53817 Sep
512
683
1.6K18 Sep
869
Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
1.23%8 Jun
1.38%
1.45%
1.31%
1.32%
0.91%
1.99%
1.08%9 Jun
1.00%16 Sep
0.56%17 Sep
0.98%
1.02%
0.37%18 Sep
1.27%
Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.
What the audience does
Likes52.6%373 in total
Reposts10.0%71 in total
Replies4.1%29 in total
Quotes0.4%3 in total
Bookmarks32.9%233 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Similar accounts
National Security Division, U.S. Dept of Justice ✓
@DOJNatSec
33.9K followers
· 934 posts
Michael Chopra
@MichaelChopra
33.9K followers
· 2.2K posts
黒魔 / Chroma
@c_h_r_o_m_a
33.9K followers
· 3K posts
Motorsportive ✓
@MotorsportiveHQ
33.9K followers
· 10.6K posts
DaveAY ✓
@Daveay7
33.9K followers
· 3.2K posts
Angel Figueroa Jaramillo ✓
@jaramilloutier
33.9K followers
· 15.5K posts
CelticsUnite ✓
@CelticsUnite18
33.9K followers
· 7.4K posts
Kingcallz ✓
@KickoStocks
33.9K followers
· 46.5K posts