ilyazub

> The technical side is all trivial.

:)

And iOS keyboard produces these dashes on double click on the dash (“-“) button: Single press: - Double press: —

I instructed agents observing other agents at runtime and it resulted to several documentation fixes, that multiple `/review` runs didn't catch.

Essentially, the prompt was `apply TDD or Nielsen's user testing to agents-observing-agents via tmux or subprocesses`.

Details are on the blog post and GitHub repo.

Wow, government-led mobile proxy network. Did they attempt to build a search index? :-)

Cool, congrats and wishing you good luck and success!

> monthly "who wants to be fired" thread Reminds me of Mad Fientist blog.

Feel free to just share the details in an email as suggested above.

> I have high confidence that I could accomplish this task using a lower end model with a high degree of accuracy.

Cool, cool. I'm super interested. Please share the process and the results.

I recently wrote a blog post where I take a deep look into Playwright’s locators, specifically comparing the Page#getByRole locator with CSS selectors in terms of performance.

Playwright’s getByRole is 1.5x slower than CSS selectors because Page#getByRole uses querySelectorAll('*') and matches elements by the accessible name.

The post breaks down how `getByRole` works under the hood, how it stacks up against CSS selectors performance-wise, and what this means for your testing suites.

Would love to hear your thoughts!