@Anil Madhavapeddy
Professor of Planetary Computing, University of Cambridge
Relays (1)
- wss://relay.ditto.pub – read & write
Recent Notes
I was in Chennai today for the launch of FP Launchpad, a new centre at IIT Madras dedicated to crafting efficient, reliable and trustworthy software with mathematical guardrails. Having a well-funded, long-horizon centre of excellence in India is a big step forward! https://anil.recoil.org/notes/fpl-launch
The Internet needs an antibotty immune system, stat! My musings on the implications of the Mythos preview, and that we need to shake up how we approach building software beyond formal specification https://anil.recoil.org/notes/internet-immune-system
1
nostr:nprofile1qy2hwumn8ghj7un9d3shjtnyd968gmewwp6kyqpq0ujnx96rzj2m7kl9yqhyplqhxuf55el43gdqw8xxpm7pdk3ph6gqwfaz86 FP Launchpad looks amazing! KC is such a badass.
@nprofile1q... there are adorable puppies wandering around the campus! Why don't we have adorable puppies wandering around Cambridge, I ask
.plan-26-14: Tracking AI screen time and escaping to pen and paper https://anil.recoil.org/notes/2026w14
2
Tons of work today on Pterodactyl.
I haven't written weeknotes for a few weeks... To be honest, I have felt a bit blocked in the writing department.
I feel we need a weeknote just for that kind of rest thing. Maybe just post a poignant meme gif
nostr:nprofile1qy2hwumn8ghj7un9d3shjtnyd968gmewwp6kyqpq0ujnx96rzj2m7kl9yqhyplqhxuf55el43gdqw8xxpm7pdk3ph6gqwfaz86 perhaps a `lint` tool could check the consistency between information in the repositor...
@nprofile1q... yes that may be useful, but I’m skeptical that VCS based methods will actually work. Agents do a ton of branch management and it’s only really useful for license provenance. Markers in the code are just as good as commit message info and more explicit
nostr:nprofile1qy2hwumn8ghj7un9d3shjtnyd968gmewwp6kyqpq0ujnx96rzj2m7kl9yqhyplqhxuf55el43gdqw8xxpm7pdk3ph6gqwfaz86 the Linux kernel requires the use of Assisted-By tags in git commits https://docs.kern...
@nprofile1q... this proposal deliberately doesn’t depend on VCS info: it’s a marker in the code itself that’s easy to keep across rebase etc and also for the human to edit
1
Here's my proposal for voluntary disclosure of AI tool usage in published OCaml code. Are any other language ecosystems doing this? I'd love to hear about them if so! https://anil.recoil.org/notes/opam-ai-disclosure
2
This is such an ingenious Apr 1st PR from Stephen Dolan I feel like it's the exact opposite of AI slop https://github.com/ocaml/ocaml/pull/14701
.plan-26-13: Oxidised hackery, standardised embeddings, and simmering on syndication https://anil.recoil.org/notes/2026w13
nostr:nprofile1qy2hwumn8ghj7un9d3shjtnyd968gmewwp6kyqpq0ujnx96rzj2m7kl9yqhyplqhxuf55el43gdqw8xxpm7pdk3ph6gqwfaz86
In that case, why not make your own CSS class, or arbitrary HTML tag ID.
@nprofile1q... just to try align with other standards; it helps search engines like Kagi as well if we have a markup others agree on