Benjamin Barrell 🪶

🇺🇸🇬🇧🇨🇦🇦🇺🇮🇪🇳🇿🇿🇦🇮🇳🇵🇭🇳🇬🇵🇰🇰🇪🇹🇿🇬🇭🇺🇬🇸🇬🇲🇾🇷🇼🇨🇲🇿🇲🇳🇦🇧🇼🇵🇬🇯🇲🇹🇹🇬🇾🇱🇷🇸🇿🇧🇿🇻🇨🇦🇬🇧🇸🇧🇧🇩🇲🇬🇩🇫🇯🇰🇾🇲🇭🇲🇺🇳🇷🇰🇳🇱🇨🇻🇬🇼🇸🇲🇽🇪🇸🇨🇴🇦🇷🇵🇪🇻🇪🇨🇱🇪🇨🇬🇹🇨🇺🇩🇴🇭🇳🇧🇴🇸🇻🇳🇮🇨🇷🇵🇦🇺🇾🇵🇾🇫🇷🇧🇪🇨🇭🇱🇺🇲🇨🇨🇮🇩🇯🇬🇦🇬🇳🇭🇹🇲🇱🇳🇪🇷🇪🇸🇳🇹🇩🇹🇬🇻🇺🇧🇫🇧🇯🇨🇫🇨🇬🇰🇲🇲🇬🇲🇷🇸🇨🇾🇹🇯🇵🇨🇳🇵🇹🇩🇪🇦🇹🇱🇮🇮🇹🇸🇲🇻🇦🇷🇺🇰🇿🇧🇾🇰🇬🇺🇦🇲🇩🇺🇿🇹🇯🇹🇲🇦🇿🇱🇹🇱🇻🇪🇪🇬🇪🇮🇱🇧🇷🇰🇷🇰🇵🇱🇧🇸🇦🇪🇬🇩🇿🇮🇶🇸🇩🇲🇦🇾🇪🇸🇾🇹🇳🇱🇾🇯🇴🇵🇸🇦🇪🇰🇼🇶🇦🇧🇭🇴🇲🇸🇴🇳🇱🇸🇷🇦🇼🇨🇼🇬🇷🇨🇾🇨🇿🇷🇴🇮🇩🇹🇷🇻🇳🇸🇪🇫🇮🇦🇩🇩🇰🇬🇱🇫🇴🇮🇸🇭🇺🇸🇰🇸🇮🇵🇱🇲🇰🇭🇷🇧🇦🇧🇬🇱🇰🇮🇷🇦🇫🏴󠁧󠁢󠁷󠁬󠁳󠁿🇲🇹🇬🇲🇷🇸🏴󠁧󠁢󠁥󠁮󠁧󠁿🇹🇼🇭🇰🇲🇴🇳🇴🇲🇳🇦🇲🏴󠁧󠁢󠁳󠁣󠁴󠁿🇮🇲🇵🇷🇬🇶🇳🇵🇧🇩🇦🇱🇽🇰🇲🇪🇱🇦🇰🇭🇪🇹🇹🇭

Phrasing is live on product hunt today 🎉 producthunt.com/products/phras…

@MistralAI I use mistral a bunch in my application, but I’m really struggling to know whether I should care about mistral medium 3.5? It’s expensive and you only discuss vibe coding. My honest takeaway as a mistral fanboy: it is just a programming model and can be ignored

Yep I try to get 8 hours of sleep every night. I routinely get at most 6. Doesn’t matter how early I go to bed, how dark and quiet my room is, how relaxed I am, fed I am, or trained I am: I wake up wide awake about 5-6 hours later. It’s not about working harder or being cool and not needing sleep. It’s just listening to my body. If I could sleep in I would, but if my body is really telling me no. Plus 8hrs of sleep per night is a truly modern invention, and I don’t know of many modern inventions that are half as healthy as they claim ¯\_(ツ)_/¯

Well I cannot speak to the specifics of what he says as I don't measure those things, just my personal experience. What I do know is that we're not a one size fits all species, medicine has been going that way for over a century because there was value to go after thet fat part of the curve. But there are those on the margins.

all the CEOs sacrificing sleep at the alter of productivity, is most certainly going to back fire. So much of decision making degrades as you get less sleep and no, you don't get "use to it." You just remain in a perpetual bad state.

Most people learn a language in two steps. Learning from books and courses teaches you to 'speak' a language in a manner of speaking, but then you have to relearn how to actually _speak_ the language. Phrasing helps you skip this detour and work towards the final destination from day one

How short can content be and still be considered "comprehensible input"? Can it be combined with Spaced Repetition? Here's my theory:

Whenever you feel like you’re done learning Devanagari, you’re wrong क + ष = क्ष Σ(-᷅_-᷄๑)

First non language log related podcast! Stop trying to learn language from short phrases and find out I recommend and encourage such long expressions on Phrasing

Progress update on learning 18 language: why learn so many, and how all my lower tier languages are 'progressing'

Language Log 4: Checking in on French, Italian, Dutch, and "Portuguese". The languages I have some ability in and am working to improve while learning my main languages

Checking in on my progress learning 18 languages at once. I cover my 4 tier 1 languages (Arabic, Croatian, Turkish, and Japanese) and a bit Chinese in the end!

> LLM coding will split up engineers based on those who primarily liked coding and those who primarily liked building. This isn’t just objectively wrong, but wow, one of the most pretentious lines I’ve read in a long time. This is like saying AI art will split artists based on who likes scribbling, and who likes producing.

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.

@visualizevalue Easy to do with phrasing.app

"Go to bed smarter than when you woke up." – Charlie Munger

@ThePrimeagen Everyone knows line width is what really matters

hey, asking for a friend can we stop using lines of code for a measurement in productivity? at one point we all agreed on this

If you send me a new message notification email with the message body, I might read it in the email and not open the app If you send me a new message notification email without the message body, I am definitely **not** going to open the app

@ThePrimeagen Did they rename it to copilot?

RIP GitHub

pretty cool to take my actual review history and best fit different curves to it to build predictive models. years of guessing and now I can actually test all my hypotheses

I signed up for tailwindUI many years ago. I've been using headlessUI for years (although recently removed it). I've purchased multiple copies of refactoringUI for the designers in my life. These are my thoughts in the hopes that should @adamwathan ever see them, they are somewhat helpful. However, I don't use any of these anymore. TailwindUI components are very specific, very detailed, and once you learn the tricks of the trade they're easier to write than to understand. HeadlessUI I stopped using because I always hit a wall - modals closing or forward ref breaking or something in configurable that needs configuration. I stopped buying designers refactoringUI because the UI trends it's discussing are out of date. The things I do struggle with a lot in front end are things like color, design systems, and consistency. The amount of time I've spent managing a color palette is ridiculous. I wish there was a tool to use color theory and oklch to fine tune and customize my color palette. Especially something that integrated natively into my project, and integrate natively with Figma and designers. Having a marginally unique color palette is valuable to a business. Animation is still pretty lacking in tailwind. I've started using animejs and I'm loving it. The spring based animations are glorious, and while I do a lot of GPU accelerated work in Phrasing, there is a time and place for CPU animations as well. Even with my own wrapper around animejs, it feels largely parallel to writing scss back in the day. And lastly, I find myself always typing out the same series of classes. A big part of tailwind to me is how in a few keystrokes you can add styles. That breaks down as the interface grows larger. While I've introduced some abstractions, I feel like a lot of them aren't as good as the tailwind core abstractions. It would be really cool, in the age of AI, if abstractions could be recommended. I assume the tailwind team has a unique knowledge to provide the best abstraction recommendations. Perhaps that's knowledge that could be transferred to an LLM to provide custom recommendations to codebases. I'm not going to pretend like I know how to run tailwind labs, but if colorUI, springUI, and abstractUI were tailwind products, I would use them a lot more than the current offerings.