Deniz Akyildiz

assistant professor of statistics, @imperialcollege

[Michael Choi @michaelchchoi]

Card shuffling, one of the earliest example of MCMC imo

[Deniz Akyildiz @odakyildiz]

Universities are worried that AI will expose how fragile many grade inflated light assessment degrees have become. The solution is not that complicated:

Bring the rigour back

https://t.co/siDNOXasUm

[David Bessis @davidbessis]

My physics oral exam at the ENS: "Have you seen videos of supraconductors levitating above magnets?
1. Prove that they do levitate.
2. If a supraconductor levitates for 30 min, estimate an upper bound for its resistivity."

1 hour exam at the blackboard, no preparation, no documents, no computer or calculator, no pen or paper... just you standing alone with a piece a chalk and a top physicist staring at you.

Supraconductivity wasn't even part of the program. The whole point was to figure out a decent model and derive everything from first principles (ie Maxwell equations, which you were supposed to know by heart, alongside approximate values of the universal constants).

Note: my specialty was math, not physics. I can't even imagine what the questions were Iike for the physics students.

[Deniz Akyildiz @odakyildiz]

Globally 6.3 bags per 1000 passengers are mishandled

@SunExpress mishandled my luggage on my last three flights 3/3

That is roughly a 1 in 4 million streak if you are average

So either I am a statistical miracle or your baggage operation is impressively incompetent @SunExpress

[Anarch97 @anarch97]

Edits are probably the most important art form right now

[signüll @signulll]

your only goal should be to believe in something before many other ppl do, one time for each parts of you life. that’s the only skill that matters if you want to escape orbit.

e.g. you have to believe in a person before others do (your wife), you have to believe in whatever you're working on or even where you're working before the rest of zietgiest does (like everyone wants to work at the labs right now, that will make it difficult), & you have to find a city/neighborhood that you can afford but still has upside/potential for a house etc.

developing asymmetric skillsets like these is more important now than ever before cuz the market liquidity in every area is limitless & the competition is global.

[Deniz Akyildiz @odakyildiz]

Great way to put it: "the permanent underclass" *is* about agency, not assets.

[Robert Parham @kn_owled_ge]

Teaching in the age of LLMs:

I failed 4 students, for the first time ever. I also gave more A+'s than ever before.

In previous years, students realized after the first or second HW that they weren't in Kansas anymore and needed to work hard.

No more. Just solve it with LLMs.

But then the midterm arrives, and they can answer 0 of 40 questions. Do they reform their ways? Nah, they just decide to "give up" on class, assuming they'll get a B, or a C, or whatever, because they submitted HW and got decent grades on those. And never before have they encountered a professor who will dare fail them.

The flip side is that the most "agentic" students now have the world's best tutor at their disposal. They deeply understand the material and aced my (intentionally very difficult) exams. As if we live in "The Diamond Age".

Inequality galore.

From my vantage point, "the permanent underclass" appears to be about agency, not assets.

[Statistics (Machine Learning) Papers @StatsPapers]

Consistency Regularised Gradient Flows for Inverse Problems

Alessio Spagnoletti, Tim Y. J. Wang, Marcelo Pereyra, O. Deniz Akyildiz
https://t.co/VaJRTpi8Ev [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙲𝚅 𝚌𝚜.𝙻𝙶]

[Kimon Fountoulakis @kfountou]

Verification of math currently takes multiple days, if it succeeds at all. It needs to drop to O(generation time).

[Sander Dieleman @sedielem]

My first blog post in over a year is a deep dive on flow maps🗺️, or how to learn the integral of a diffusion model to enable faster sampling and several other cool tricks.

It's the longest one yet👀 Let me know what you think!

https://t.co/O8bBGZ9qjC

[Dwarkesh Patel @dwarkesh_sp]

Terence Tao spent a year at the Institute for Advanced Study - no teaching, no random events of committees, just unlimited time to think. But after a few months, he ran out of ideas.

Terence thinks that mathematicians and scientists need a certain level of randomness and inefficiency to come up with new ideas.

[Kpaxs @Kpaxs]

There is a specific kind of intelligence that is almost never celebrated but is consistently effective: the intelligence that recognizes when the game being played is not the game worth playing.

[Lev Reyzin @lreyzin]

Now is basically the last chance to finish up your theorems. A few more months, and everyone will assume the real work was done by ChatGPT.

[Haider. @haider1]

OpenAI Sebastian Bubeck says deep expertise is more important than ever in the AI age

to get maximum value from AI, you need enough real understanding to describe the problem clearly

"this creates the gap between people who keep studying and those who rely too much on AI"

[Aaron Haspel @ahaspel]

The pleasure in having people admire your work lasts until you see what else they like.

[vivian @jazzloaf]

Getting too into cellular automata is LM psychosis for physicists pre-2023.

[Deniz Akyildiz @odakyildiz]

Today in things on arXiv I wish I'd come up with

[Deniz Akyildiz @odakyildiz]

We are recruiting for two roles at Imperial College London (Department of Mathematics) to join the PRISM (Probabilistic Rare-event Inference for Safety of Models) project: https://t.co/FZGUkyBXrC

[Stat.CO Papers @StatCOupdates]

James Cuin, Davide Carbone, Yanbo Tang, O. Deniz Akyildiz. [statML]. Efficient Stochastic Optimisation via Sequential Monte Carlo. https://t.co/nwtU3X8sRz

[Stat.CO Papers @StatCOupdates]

James Matthew Young, Paula Cordero-Encinar, Sebastian Reich, Andrew Duncan, O. Deniz Akyildiz. [statML]. Diffusion Path Samplers via Sequential Monte Carlo. https://t.co/VKUWj64R66

[Andrej Karpathy @karpathy]

A few random notes from claude coding quite a bit last few weeks.

Coding workflow. Given the latest lift in LLM coding capability, like many others I rapidly went from about 80% manual+autocomplete coding and 20% agents in November to 80% agent coding and 20% edits+touchups in December. i.e. I really am mostly programming in English now, a bit sheepishly telling the LLM what code to write... in words. It hurts the ego a bit but the power to operate over software in large "code actions" is just too net useful, especially once you adapt to it, configure it, learn to use it, and wrap your head around what it can and cannot do. This is easily the biggest change to my basic coding workflow in ~2 decades of programming and it happened over the course of a few weeks. I'd expect something similar to be happening to well into double digit percent of engineers out there, while the awareness of it in the general population feels well into low single digit percent.

IDEs/agent swarms/fallability. Both the "no need for IDE anymore" hype and the "agent swarm" hype is imo too much for right now. The models definitely still make mistakes and if you have any code you actually care about I would watch them like a hawk, in a nice large IDE on the side. The mistakes have changed a lot - they are not simple syntax errors anymore, they are subtle conceptual errors that a slightly sloppy, hasty junior dev might do. The most common category is that the models make wrong assumptions on your behalf and just run along with them without checking. They also don't manage their confusion, they don't seek clarifications, they don't surface inconsistencies, they don't present tradeoffs, they don't push back when they should, and they are still a little too sycophantic. Things get better in plan mode, but there is some need for a lightweight inline plan mode. They also really like to overcomplicate code and APIs, they bloat abstractions, they don't clean up dead code after themselves, etc. They will implement an inefficient, bloated, brittle construction over 1000 lines of code and it's up to you to be like "umm couldn't you just do this instead?" and they will be like "of course!" and immediately cut it down to 100 lines. They still sometimes change/remove comments and code they don't like or don't sufficiently understand as side effects, even if it is orthogonal to the task at hand. All of this happens despite a few simple attempts to fix it via instructions in CLAUDE . md. Despite all these issues, it is still a net huge improvement and it's very difficult to imagine going back to manual coding. TLDR everyone has their developing flow, my current is a small few CC sessions on the left in ghostty windows/tabs and an IDE on the right for viewing the code + manual edits.

Tenacity. It's so interesting to watch an agent relentlessly work at something. They never get tired, they never get demoralized, they just keep going and trying things where a person would have given up long ago to fight another day. It's a "feel the AGI" moment to watch it struggle with something for a long time just to come out victorious 30 minutes later. You realize that stamina is a core bottleneck to work and that with LLMs in hand it has been dramatically increased.

Speedups. It's not clear how to measure the "speedup" of LLM assistance. Certainly I feel net way faster at what I was going to do, but the main effect is that I do a lot more than I was going to do because 1) I can code up all kinds of things that just wouldn't have been worth coding before and 2) I can approach code that I couldn't work on before because of knowledge/skill issue. So certainly it's speedup, but it's possibly a lot more an expansion.

Leverage. LLMs are exceptionally good at looping until they meet specific goals and this is where most of the "feel the AGI" magic is to be found. Don't tell it what to do, give it success criteria and watch it go. Get it to write tests first and then pass them. Put it in the loop with a browser MCP. Write the naive algorithm that is very likely correct first, then ask it to optimize it while preserving correctness. Change your approach from imperative to declarative to get the agents looping longer and gain leverage.

Fun. I didn't anticipate that with agents programming feels *more* fun because a lot of the fill in the blanks drudgery is removed and what remains is the creative part. I also feel less blocked/stuck (which is not fun) and I experience a lot more courage because there's almost always a way to work hand in hand with it to make some positive progress. I have seen the opposite sentiment from other people too; LLM coding will split up engineers based on those who primarily liked coding and those who primarily liked building.

Atrophy. I've already noticed that I am slowly starting to atrophy my ability to write code manually. Generation (writing code) and discrimination (reading code) are different capabilities in the brain. Largely due to all the little mostly syntactic details involved in programming, you can review code just fine even if you struggle to write it.

Slopacolypse. I am bracing for 2026 as the year of the slopacolypse across all of github, substack, arxiv, X/instagram, and generally all digital media. We're also going to see a lot more AI hype productivity theater (is that even possible?), on the side of actual, real improvements.

Questions. A few of the questions on my mind:
- What happens to the "10X engineer" - the ratio of productivity between the mean and the max engineer? It's quite possible that this grows *a lot*.
- Armed with LLMs, do generalists increasingly outperform specialists? LLMs are a lot better at fill in the blanks (the micro) than grand strategy (the macro).
- What does LLM coding feel like in the future? Is it like playing StarCraft? Playing Factorio? Playing music?
- How much of society is bottlenecked by digital knowledge work?

TLDR Where does this leave us? LLM agent capabilities (Claude & Codex especially) have crossed some kind of threshold of coherence around December 2025 and caused a phase shift in software engineering and closely related. The intelligence part suddenly feels quite a bit ahead of all the rest of it - integrations (tools, knowledge), the necessity for new organizational workflows, processes, diffusion more generally. 2026 is going to be a high energy year as the industry metabolizes the new capability.