Thanks everyone who joined the Racket meetup!
We started by talking about Racket meetups in general, especially offline/local ones. There is already a Bay Area meetup coming up: Bay Area Racket Meetup · Luma, and an experimental pub meetup in Scotland: UK Racket meet-up · Luma. Everyone is encouraged to try organizing an offline Racket meetup in their own area.
We also mentioned the RacketCon 2026 call for participation:
Some projects we discussed:
- Racket on WebAssembly: https://racket-wasm.netlify.app/
uxnsh, a project related to 100 Rabbits / uxn / Varvara, with a memory implementation in pure shell: Dominik Joe Pantůček / uxnsh · GitLab- Viridithas, a strong chess engine written in Rust by a UK developer based in Edinburgh: GitHub - cosmobobak/viridithas: A superhuman chess engine. · GitHub
Naturally, we also talked about AI: how people are using it, what helps it produce decent results, and how important fast feedback loops are. Error messages came up as a particularly important part of this, with Rust as one example. This recent paper on verbose errors was mentioned:
https://arxiv.org/pdf/2606.01522
The chess engine comparison led into a broader AI discussion: chess engines used to be weak, but with better algorithms and more hardware, even a phone can now beat the world champion. We discussed whether AI might follow a similar path, including the possibility that in five years useful models could be locally hosted on ordinary machines.
We also talked about the current economics of AI: hardware scarcity, subsidies, token costs, and whether the promised returns on AI investment are realistic. This little site was mentioned:
Related links from that discussion:
- Former Google CEO Eric Schmidt getting booed while talking to graduates about AI: https://youtu.be/tNH43a1EI7s
- Collapse OS and concerns about hardware availability by 2030: Collapse OS — Bootstrap post-collapse technology
On the programming languages side, we mentioned that guaranteed Rust tail-call optimization has been merged into nightly:
We also briefly discussed the take that “every language is either a C or a Lisp”, with JavaScript described as more Lisp-like and Python as more C-like, mostly in terms of how expressions and statements work.
Finally, Justin Slepak showed a demo of implementing reduction rules inspired by the WebAssembly abstract syntax paper:
https://dl.acm.org/doi/pdf/10.1145/3062341.3062363
He followed the small-step reduction rules from the paper and asked for ideas on improving the solution.
Thanks again to everyone who came along and contributed!