Notices by Else, Someone (nobody@mastodon.acm.org), page 2

Sorting pdfs from the laptop-before-the-last-laptop, most of them still waiting to be read... exhilarating

@theorytoe anything particular inspiring this thought?

@p

@civodul Nah `nix run` (`man nix3-run`) is just a less inconsistent version of `nix-shell --run`, there's no containers there

@nemobis @evan One could start the argument by just asserting that the whole point of E2EE is that trusting the clients on both ends is "enough", in the sense that it makes no difference whether the channel and the server are compromised.

With Signal one tends to trust the client because it's open-source and the releases are (claimed to be) reproducible. With Whatsapp... well maybe you could inspect its traffic with mitmproxy idk, but ultimately there's no reason to trust the client

Aalto, and Finland in general, feature very active student communities, especially in the MSc and BSc programs. There are notorious groups raiding Helsinki-Stockholm Viking and Silja Line ferries, there are clubs teaching sailing, and there are groups promoting Lean or PL research.

OtaNix Ry, a non-profit run by inspired students, is organizing a Nix Workshop this Wednesday at the Aalto University. Join! https://discourse.nixos.org/t/nix-meetup-helsinki/48211/10

Meantime, I'm in Helsinki and packing my stuff for moving from Jatkasaari to Herttoniemi

@civodul I think (or hope) authoritarianism in communities and organizations is just a teenage sickness, manifested by people's irrational desire for the community as a "virtual entity" or a "thing in itself" to "exist", to be "constituted", and for its members to "align" and "share values", to be homogeneous, all of which requires setting up "standards" and "filters", and institutionalizing "rules"

@civodul I was actually thinking that if I were to partake in the LLM rush I'd try to use Hydra's build logs train a predictor that'd take a .drv graph (with access to e.g. cmake flags and potentially to inputSources) and predict build errors. If one were to achieve decent accuracy and inference cheaper than the actual builds (plausible for e.g. tensorflow), it could make for better merge trains and rolling release channels