Posts

I always find myself getting lost in the Cloudflare dashboard. Every time I need to set up a redirect rule, it seems to have moved. I ask an agent where to find it, and it confidently tells me to click through menus that don’t exist anymore. Then I’m the one clicking around again, and the agent can’t help because it has no idea what I’m looking at. The fix was pretty simple. It’s what you would do with a friend who’s a patient expert. I took a screenshot and sent it over, and the agent saw the menu as it actually was that day and told me it had moved to a different submenu. ...

The best part of pair programming is pretty much just having someone else look at what you wrote. They notice the stuff you stopped seeing an hour ago. I wouldn’t merge my own work without a review, so why would I let the model that wrote the code be the only one to review it? A model reviewing its own work has the same blind spots it had while writing it. It made its choices for reasons that still seem good to it, so asking it again kind of gets you the same thing back. ...

REA, short for Reverse Engineer Anything, went from nothing in April to about 370,000 GitHub stars this month. It’s an MCP server that hands your coding agent the reverse-engineering tools people have used for years, like Ghidra, IDA, and Hopper, and lets it work out how a program behaves without the source. It should be noted that these tools typically are only used by experts and they take a lot of skill to use. ...

Agents have started causing problems on our clusters. Heavy work kicked off on shared login nodes, Slurm getting polled in tight loops, that sort of thing. I’d been using a small CLAUDE.md for myself on Nibi for a while, and I was asked to turn it into something for everyone. The Alliance clusters get their environment from ComputeCanada/software-stack-config, so I opened PR #111 to add an agents/AGENTS.md there on request. Claude Code reads /etc/claude-code/CLAUDE.md system wide, and most other harnesses look for an AGENTS.md, so one file and a symlink covers a lot of ground. ...

Today I had to rate limit a user on SHARCNET’s internal LLM service for the first time. In the last 24 hours they had sent about 800,000 requests, and every single one was under 500 tokens. To be clear, they didn’t do anything wrong. The service is set up to let them do it. They had a pile of small work and a working endpoint, so they wrote some sort of loop. I probably would have done the same. But a loop that sends hundreds of thousands of tiny prompts is a batch job, and this was a really good reminder of why batch clusters are still super important. ...

I hate when you sit down to work at 9:30 after clearing emails, send your first prompt, and by 1pm you’re out of usage. The window doesn’t come back until 2:30, so you lose an hour and a half of the afternoon waiting on a timer. To be clear, there are actually two limits here, and they do pretty different jobs. The five-hour window controls when you can use your plan. The weekly cap controls how much you get in total. Most people worry about the five-hour one because it’s the wall they keep hitting, but in my opinion the weekly cap is the real budget. ...

I recently used Tectonic to rerender my résumé and make changes with a coding agent. This made the back-and-forth a lot easier. I could ask for a change, have the agent edit the underlying LaTeX, and get another PDF to check. In my opinion, Tectonic and a coding agent are now a pretty practical alternative to Overleaf if you’re working on a document by yourself. The agent can work directly on local files and run the compiler. This way you don’t need it clicking around a browser through computer-use tools trying to diagnose how your tables are formatted. ...

After I updated my model recommendations, I got some feedback from people still reaching for Sonnet or models like Luna. I get it. If I’m breaking a job down and checking every step, Sonnet can get me pretty much the same result as Opus most of the time. But why am I doing all that planning for it? That defeats the purpose of a fully agentic approach. What I typically try to do with Opus is give it a wider task: give it the goal, let it work through the branches, and come back to review what it did. ...

When I tried to write tests for whether EEG got cleaner, several of my metrics rewarded rejecting everything. This is an absolutely classic case of badly defined test metrics. Consider an ERP study and two cleaning policies. One rejects nearly every blink, along with a lot of trials. The other leaves a little ocular activity but keeps more trials. If I only count blinks, the first one wins. If I only count retained trials, the second one wins. Neither one is really better than the other. ...

Astra was the first model that made me stop and go “wow.” With other releases, I could usually predict what they’d try next when a task got hard. Astra spun up multiple Playwright browsers to test and compare things, then recorded videos to show what it had done. I’d seen people talk about doing that online. I hadn’t asked Astra to do it. It also overengineered some open-ended problems I gave it. And I can’t use it every day without upgrading my plan. That matters more to my recommendation than how impressive it was. ...