Viewing profile — Leynos

Is this using the server side compaction of the responses API?

1st line of defence, use something like ponytail to enforce brevity. Use property testing and behavioural testing on top of unit testing. Enforce readability standards so you will …

If you ask Fable or 5.6 Sol to improve performance, it will generally know to build a benchmark and create a test corpus. I'm not sure where the contrary suggestion is coming from.…

The problem with Empire of AI is that it takes such a scattergun approach and doesn't really build a coherent thesis. It is also very difficult to draw clear directional informatio…

Zerover for life

[flagged]

I'd prefer a tag to the mounds of "this looks like it is AI generated, I can tell from the pixels and from having seen quite a few AIs in my time" comments. That way the people who…

You'd need the whole edit tree along with all the prompts used along the way, which most people are not yet set up to capture.

For normal building work I use Opus to plan and GPT 5.6 Terra to build. The point is, these are not normal constrained building tasks. Perhaps I should have had more faith in Opus'…

The sort of thing Fable and Sol excel at are long horizon tasks. The sort of thing I have been using them for is migrating large numbers of repositories to new tooling simultaneous…

Things that I reckon will become a lot more important from a developer's perspective over the next year: - Shaping work so it is more decomposable, legible, verifiable and understa…

Yeah, I went through a period after 4.7 launched of not using Claude for code work at all because of the condescending refusals. (Kept using it for planning and design work). Still…

While there are some coding focused models (composer, for example), the majority of frontier models are pitched as general purpose. The coding harnesses for Claude and GPT are even…

From a purely utilitarian standpoint, direct to cell feels like a good thing to me. Large swathes of Scotland don't even have sufficient mobile connection to send a text message (s…

The prompted response is far from the finished piece of writing. You'd probably want to share the full edit tree and include subsequent refinement prompts in the commit messages.

Yes, that was the point. It made unsafe behaviour visible in a way that could be addressed. I hadn't heard any reports of it being dysfunctional.

As a reader (not as someone who is posting the articles), the AI prose generally doesn't bother me. I'm usually more concerned about what the article says than how it says it.

Samsung are back up 5% today on news of a planned buyback.

Alerts on test fixtures, so suspect it is doing nothing new.

You can see the validation approach they used here: https://github.com/adamraudonis/prylint/blob/main/harness/ch...

Currently, there are things pylint does that ruff doesn't. To use these, I was running pylint on pypy to get it running at a reasonable speed. Having pylint reimplemented in Rust s…

I said "outside of situations where it is required by contract", which I believe would include a CLA.

Which model was used for the benchmark results shown on your GitHub README.md?

Context: https://www.businessinsider.com/what-is-le-chaton-fat-mistra...

I quite like my mechanical spider from Wild Wild West and the coffee it makes with a 50% success rate