Carson Poole

this was lots of fun and lots of all nighters over the past few weeks. really happy with what we achieved!

@MartinShkreli different view… have heard it’s a generator on fire but no clue if that’s true

Well, that’s not normal

@allTheYud is getting victory lap material the likes of which have heretofore never been seen. my fast takeoff has been updated from ~0 to 5% and I’m unironically having slightly whimsical thoughts of what happens if the electric grid is down in the morning

GPT-5.2T vs Opus 4.5 shows some major big model smell

someone pls make a worldle that has a daily leaderboard for whom can make the highest logprob sentence in some range of tokens

@karpathy new visit to @dwarkesh_sp’s podcast soon?

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.

I have never understood why people need these tools? What is hard about (# billion params) * (2 for 16 bit, 1 for 8bit, etc) * (fudge factor for activations, kv cache) < (vram on your GPU in GB)

@ID_AA_Carmack I find it interesting that like many things before, meaningful ML research progress manifests as a 1-2 line of code delta (without hiding complexity): Dropout, Adam(W), ResNets, BatchNorm, etc

#PaperADay 7 Cautious Weight Decay arxiv.org/abs/2510.12402 This is a 36 page paper about a very simple idea: Don’t apply weight decay when it is in opposition to the current optimizer step. If the step is moving the weight farther from zero, there is no decay. If the step is towards zero, decay moves it in faster. They spent 20,000 H100 GPU hours (about $60k!) testing this across multiple optimizers and models, and it looks like it is basically always a modest improvement, with no changes to any hyperparameters. My current models use weight norm on most of the parameters, but there are still some with traditional weight decay. A first test with this idea does seem to be a tiny improvement, but I will need to do more runs to have confidence in it. Two modifications of the idea come to mind: Use the current gradient instead of the optimizer step (which includes momentum) for masking, as in arxiv.org/abs/2411.16085. If using a cautious optimizer, explicitly masking the optimizer step before calculating cautious weight decay would do this automatically. Weight decay is an exponential effect, which mixes with a linear learning rate. It might be interesting to just have a larger learning rate when the step is heading towards zero than away from it, which would do similar things, but with different learning dynamics.

the comparison is really striking between when Google released Lion versus everybody quietly switching to Muon

@Bojangles when Manhattan?

NEW YORK WE'VE ARRIVED! Come see us at 📍 5910 Church Avenue, Brooklyn, NY 11203 🍗

@fleetwood___ Had more short term conviction on the long side of the trade, so it’s done alright in the past 6 months

@CarsonPoole Would have been a very bad bet

just gonna go ahead and lock in this prediction: diffusion LMs will eat GPT models and I’m probably long FLOPs short HBM

Hello World 👋 Welcome to Drafted — an AI tool that lets anyone design a home from scratch, tailored to your life. techcrunch.com/2025/12/23/thi…

a phenomenon I haven’t seen anybody point out is what happens when you can “few shot” a robot? with sufficient scale this ability emerged with LLMs. instead of training it to perform a specific task, can you show it 2-3 representative examples of itself doing said task?

@stochasticchasm Deepseek did this for R1; they went from 7k dim for the attention to 2k for the (non-shared) experts. I’m assuming there’s something more here they’re doing beyond that to qualify this as truly novel? (I haven’t looked at their code yet)

Just straight up doing MoE in lower dim, kinda interesting

the momentum is building

>1000 tps single batch inference. short HBM long FLOPs

my take (x.com/CarsonPoole/st…) continues to age well

why does this look like nvidia selling gloves

TIL the originator of the phrase "embarassingly parallel" is Cleve Moler, the creator of matlab (sorry if that gives you painful flashbacks)

make (positive) science fiction a reality

@gregisenberg x.com/CarsonPoole/st…

people outside the tech bubble call chatgpt just “chat” pretty cool.