Adrian Roselli, pH0Mar 6

RE: https://mastodon.social/@mgifford/116106634552309135

(I asked Mike on LinkedIn.)

Has anyone tried this, measured outcomes?

Asking because:
https://arxiv.org/abs/2602.11988

It found AGENTS‌.md “files tend to reduce task success rates compared to providing no repository context, while also increasing inference cost by over 20%.”

#accessibility #a11y

Show threadzSchön 
@aardrian There’s an whitepaper which confirms this for expert level guidance for accessibility: https://conesible.de/wab/whitepaper_webaccessbench.pdf
And here the relating posts by @kc:
- https://chaos.social/@kc/116100159237726006
- https://chaos.social/@kc/116120023354340978
Mar 7 at 3:03pmWeb
Show threadCaseyMar 7
@TheAlgorythm @aardrian Yes, thanks for sharing! I conducted tests specifically for production-level accessibility implementations and results are similar, without knowing about the other study before. Results and a paper on the study design are on https://conesible.de/wab and will be on arXiv as soon as I get around to it.
Accessibility Is Civil Rights. AI Must Stop Shipping Barriers.

Show threadAdrian Roselli, pH0Mar 7
@kc @TheAlgorythm
This is a case where the subject is outside my domain. Hence my questions. So I appreciate you citing more information.