Controlling Reasoning Effort in LLMs
How LLMs Learn Low-, Medium-, and High-Effort Reasoning Modes
The latest — Jul 18
Simon Willison · Director & Django co-creator
Tool: SQLite Query Explainer Julia Evan's, in Learning a few things about running SQLite : Maybe one day I’ll learn to read a query plan. Big same.... which inspired me to have Fable build this interactive explain tool, which runs SQLite in Python in Pyodide in Web Assembly in…
How LLMs Learn Low-, Medium-, and High-Effort Reasoning Modes
Claude make Fable 5 permanent An update from the @claudeai account on Twitter: Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits. Pro and Team Standard users will continue to have access to Fable via usage credits, and will…
nascheme/quixote A certain vintage if Python web nerd might be delighted to learn that the most recent commit to the Quixote web framework was six hours ago . The oldest commit in that repo is from 21 years ago, and that was the initial import of Quixote 2.4 from Subversion into…
Hello! I’ve been working on a Django site recently, and I decided to use SQLite as the database. When I was getting started with using SQLite as database for a website I read a bunch of blog posts about how it is totally fine to use SQLite in production for a small site and I…
See also: day one, day two. There is only one thing that is better than two days of HTTP workshop, and that is of course three days of HTTP workshop. The final day of this edition of the series started out with us again shuffling around where we parked ourselves around the big…
The decline of Europe is not inevitable, despite how much Americans love to joke/proclaim that the continent is doomed to become an open-air museum. Sure, it's possible that things have to get worse before they get better, but believing that "it's over" is just loser talk. It's…
To a sceptic, spending $165K to migrate Bun from Zig to Rust sounds very expensive. But to a realist, shortening a 1-2 year migration down to 11 days opens amazing new opportunities for devs. However, a thoroughly-tested project is required to pull it off.
When people think of legacy modernization, most folks aren't imagining the target environment will be Java 8. But this was the challenge facing Nik Malykhin when he needed to run a Java 1.5 codebase on today's hardware. His early use of LLMs gave plausible answers that did not…
Short note on Thinking Machines Lab's 975B Inkling open-weight model, its benchmark profile, sparse MoE design, short convolutions, embedding RMSNorm, and relative-position bias.
A lot of people simply don't understand the degree to which competition matters in B2B software, and what the experience of competing effectively will and should feel like. Competition is simply the art of increasing your win-rate when you and a competitor vie for the same…
If you missed it. I already described day one. Caffeinated and ready, we all gathered in the same spacious room as yesterday, but seated in new places as “suggested” by our captain. Some of us even remembered to move over the name tags we wrote yesterday to our new seats. No…
At the beginning of the year, I made a 26 for 26, which is a list of 26 mini-goals for 2026. Here they are with their status. I use ✅ for done, 👌 for easy, ⚠️ for hard but possible, 🛑 for probably will not get done. So 11/26 done, which is less than 13 […]
Outline
"Build a smart home", they said. "It'll make life so much better", they said. Well, life wasn't very bloody good at 23:00 the other night after travelling 33 hours from Paris only to find the IoT doorlock batteries dead and the
Engineers need to own the outer loop, the accountability for these systems. A written version of my AI Engineer World's Fair 2026 closing keynote: quality, verdict, and answerability, the three hidden costs of delegation, and building software factories you can stand behind.
New here? Hi, I’m Michael. I’m a software developer and founder of small, indie tech businesses. I’m currently working on a book called Refactoring English: Effective Writing for Software Developers . Every month, I publish a retrospective like this one to share how things are…
We've been buying servers from Dell since the 2000s at 37signals , but I was never too impressed with their personal computers. They either felt cheap or enterprisey to me. Like they were made exclusively for people who are handed standard-issue laptops by corporate, and not…
On this hot summer’s day in Basel, Switzerland, the seventh HTTP workshop started. These events tend to work roughly the same way and the people in the room are also to large extent familiar and known since previous editions. Forty people in a meeting room, where we take turns…
AI made our pull requests 3.5x bigger, but our defect detection dropped. PR data from two projects shows what happens to code quality when AI writes the code. The post We 3.5x’d Our Pull Requests with AI: Now We Catch Fewer Bugs appeared first on Pawel Brodzinski on Leadership…
Towards the end of 2025, it seemed that we couldn’t get through a week at Provet without some kind of infrastructure degradation. Outages were piling up, they were hard to diagnose, and customers were losing patience. When a major AWS outage hit in October, our customers didn’t…
What does this mean?
LLMs generate code incredibly fast, but to ensure they generate exactly what is intended, they need clear boundaries. Abstractions and Domain-Specific Languages (DSLs) provide a strong harness that guides LLMs right from the start. Unmesh Joshi describes how the example of…
To be brutally honest, this (making Claude my social media manager) all started because I felt inadequate about self-promotion. And being an engineer, my response to feeling bad about something is to ask what I can build and what data would make the feeling go away. So at the…
TL;DR: is your refrigerator running malware? If so, you better catch it!
Your “Bluesky account” is not just a Bluesky account: it is an account that can be used with a variety of other applications. This post is going to be an exploration of part of what that means from a technical perspective, so if you’re not a software developer, this post isn’t…
A presigned URL is a replay attack you did on purpose. Replayable auth tokens are the textbook way to create vulnerable systems, but Tigris ships them as a first-class feature with presigned URLs and so does every other object storage system on the planet. However this isn't an…
Taste used to be a byproduct of the reps. Agents took the reps. So if you're junior you now have to go get the taste and judgment on purpose. A builder's look at the weakening entry level, the two debts we now owe, and concrete ways to build judgment when the machine writes the…
One hundred episodes. For this occasion, years in the coming, Lyle and I talk about the thing itself — the nuts and bolts of the podcast, including how the robots now do an alarming amount of the work, what a hundred episodes taught us, and why we use the “The One About…” naming…
Some more of my notes from Thoughtworks Future of Software Development Retreat . When we had our first retreat in Utah early this year, nobody had heard of Harness Engineering . This time we had a whole session on it. When comes to the guide side of harnesses, most of the…
Look at the past history of this blog. There are many blog posts about programming with AI, a few of them date back to January 2024 (like this: https://antirez.com/news/140). I’m a relatively well regarded programmer, after all. I don’t have the need to still be in the “loop” as…
I feel that some vibecoded software changes somewhat randomly and unexpectedly. That made me think about Bruegel’s “The Tower of Babel” which shows an already quite chaotic depiction of the Tower of Babel. The story is usually told as one about pride and ambition and ultimately…
(Full Canada holiday writeup to come) Got ~1 hour sleep on the plane, and ~1 hour in the cab home, so didn't get going to pick up Cookie right away from Anna's parents - but after some caffeine I was ready to drive us to pick up Cookie Then Cookie decided to be rather alert…
(Full Canada holiday writeup to come) Reading: Tiamat's Wrath Leviathan Falls Watched: The Drama (2026) Fantastic Beasts: The Secrets of Dumbledore (2022) Switzerland vs Algeria (World Cup 2026) Canada vs Morocco (World Cup 2026)
A busy week before our trip to Canada 🎉 (Full Canada holiday writeup to come) It was very toasty this week As a way to reduce the overall heat, I ended up leaving my PC off while working, and found that worked quite well - but does mean I need to listen to music through my…
In the 1980s, France started 43 nuclear reactors across 14 sites . On average, each reactor took just seven years to build. Forty years later, all but one of these reactors are still running, and they continue to produce nearly half of France's electricity . Can you imagine…
Short note celebrating Ahead of AI reaching 200,000 subscribers.
The path to promotion is not competence but a learnable set of parasitic "rat" behaviors, and here is the concrete skill-set that produces the climb.
Eight years ago, I wrote about my theory of restoring struggling teams , which came down to four steps: A team is falling behind if each week their backlog is longer than the week before. Solve by hiring more. A team is treading water if they’re able to get their critical work…
I’ve recently been thinking a lot about the concept of “soil horizons”, which is the idea that there are many distinct layers of soil, from topsoil all the way down to bedrock, which all combine into a soil horizon. Translating this idea into software, the ideal codebase would…
In the last few months every engineer I hired stumbled in exactly the same way: Project management. They learned quickly, but that first big project has become a rite of passage.
This week’s discussion of the ChatGPT app and its move to Electron merits a link to my evergreen article The Persistent Gravity of Cross Platform : At the highest level, cross-platform UI technologies prioritize coordinated featurefulness over polished simplicity. I’ve added a…
Hint--it's in Explore & Expand
The QuadRF (pictured above) a phased-array radio built around a Raspberry Pi 5 and an FPGA board with picosecond-level timing. It does advanced signal processing and beamforming. It can see WiFi through walls and track drones in flight. If the open source community can come up…
I haven’t had time to blog in a while, I have a lot to say, but have been doing a lot of things and so haven’t gotten words down in a while. One of those things in the past few weeks has been picking back up my Rue project. I’m really excited for what it’s turning into, even…
Power users generate 10x as many lines of code vs the median, most of the AI spend is coming from input tokens not output ones, and almost half of AI changes are accepted without manual review by devs (!!)
Valve wins by doing absolutely nothing
Years ago I predicted that columnar storage would remake observability. What I didn't see coming: vendors would build it, nerf it, and sell a worse version back to you as "Datadog, but cheaper"
In our 99th episode, Rands is back on the Pro Leisure Circuit — infinite time, a suspicious amount of Arc Raiders, and no wind in the sails — and we wrestle with what actually motivates a person when the engine of work goes quiet. The usual suspects: Arc Raiders, absorbing…
The 'Special Value' Pi 4 pictured above is probably the rarest Raspberry Pi I own—even rarer than my blue special edition Pi . A Raspberry Pi reseller briefly listed a special 'value edition' Pi 4 . But the product page 404's now. While it was up, my curiosity got the better of…
How's this for a location?! I mean, last week was nice with Scott in Mallorca, but Marrakech is, well, wow 😮 Anyway, about those data breaches... This week I'm talking about the futility of attempting to remove piss from a pool , yet here we are, with
Today, we're releasing the second season of the Now Go Build documentary series. Five episodes featuring technology leaders from around the world solving the hardest problems in healthcare and education.
After the first cohort of DRI Your Career I wrote that the first time you do something, you can believe in it but you can only really hope. But we just wrapped up our second cohort of DRI Your Career and have two other courses we’ve run full cohorts of. I’m going to call it […]
Review season. You start your career thinking the review is the most important milestone of the year, but as an individual, you only see part of the picture. Everyone cares about the compensation (base salary, bonus, and stock). That’s the data they care about, but where did…
Feedback from an expert is how you learn taste. Your code doesn't matter until it's shipped.
I got the Xteink X4 device, and it's really fun to hack around with!
Understand the tradeoffs of team size and span of control for managers
AI gets good at anything with an answer key. Your career is everything that doesn't have one.
Every time I have given up on America, it has been a mistake. And yet, America has never, ever given up on me.
Happy July 4th! For those of us around the world contemplating independence, it's a good day to think about how we came to rely on expensive cloud infrastructure for our fundamental computing needs. With that in mind, here is my latest toy project: an open source tool that makes…