AlexTCH
Ir al canal en Telegram
Что-то про программирование, что-то про Computer Science и Data Science, и немного кофе. Ну и всякая чушь вместо Твиттера. :)
Mostrar más333
Suscriptores
+124 horas
+17 días
+130 días
Archivo de publicaciones
333
https://www.youtube.com/watch?v=eUjx72GApN4
Em Chu's talk at JuliaCon'26 regarding what's expected from a mature lowering pass for a (dynamic) language (with macros) and how to get there. Primarily it's provenance for error messages, of course. And macro hygiene.
333
It's curious that LLMs (or "agents"), which are "software 2.0", are writing regular code for regular apps, which are "software 1.0".
333
Кот: делает какую-то хуйню.
Ты: умиляешься и гладишь.
Кот: начинает регулярно делать эту хуйню, рассчитывая, что его будут гладить.
333
Не стоит прогибаться под изменчивый мир Пусть лучше он прогнётся под насⓒ Машина Времени
We didn't know to be cut out for this It would have to be cut into usⓒ Marilyn Manson
333
An absent-minded street thug struggles to live up to his legendary older brother's reputation and longs for the days of gang warfare.ⓒ https://www.imdb.com/title/tt0086216/ THE FUCK?! 🤣
333
That's kinda explains why the heck one needs (higher) category theory for.
https://www.youtube.com/watch?v=ZkWmKMig7Rk
333
https://scp-wiki.wikidot.com/scp-610
Suddenly it became even funnier. Or less funny depending on how you look at it and the news.
333
— Does structural integrity still hold?
— Vacuously.
— What do you mean?
— There ain't no structure no more.
333
https://www.lms.ac.uk/events/lms-bcs-facs-seminar-2026
12 November 2026, 16:00 - 18:30
"Alan Turing, The Absent Father of Computing" by Andrew Hodges (University of Oxford)
Alan Turing (1912-1954) has often been dubbed the ‘Father’ of computing, this vague word conveying the complex connection between On Computable Numbers in 1936, the digital computing plans of 1945, the first working electronic computers in 1948-51, and computer science thereafter. I will address this complexity, starting with an update on the well-known question of whether and how von Neumann’s EDVAC report of 1945 was related to Turing’s work. But my emphasis will be on what happened after 1945, regarding it as a puzzle that Turing drew so little attention to his own 1936 foundations, despite their relevance to his famous Intelligent Machinery predictions. His extraordinary reticence affected the emergent science and technology of computing, as I will illustrate with some little-known documents. My conclusion is summarised in the title: it is fair to consider him a ‘father’, but he avoided taking either praise or responsibility for his offspringOnline attendance is possible via Zoom after a registration.
333
https://www.youtube.com/watch?v=xkKeGtIalXM
Yeah... Now I feel like Rents but without any drugs. And it sucks ass big time.
333
"Super-average-human-intelligence"
Probably, doesn't sound as scary, but it does sound much more plausible.
And as they say, to escape a lion you don't need to outrun lion, you only need to outrun the next guy.
333
Julia 1.13 has been released:
https://julialang.org/blog/2026/09/julia-1.13-highlights/
No major new features, but a handful of various performance improvements including faster package downloading, precompilation, loading and the hash function replacement (new one implemented in Julia).
The most interesting improvement could be the major GC phase speed up. Which was achieved by NOT marking and traversing image objects. Essentially, the objects `mmap`ed from the system and library images which never get deallocated. Implementation is a bit more involved than that, but the principle remains: don't perform useless work. Figuring out which work is useless is the real trick.
333
NVidia officially has adopted Rust and exposes CUDA through Rust libraries (WIP):
https://developer.nvidia.com/blog/introducing-cuda-rust-two-tracks-for-writing-gpu-kernels/
333
It's funny how "AI startups" with their LLMs made pure math fancy and newsworthy. Two years ago nobody gave a flying fuck to weird Navier-Stokes solutions or The Fermat Theorem.
Very few actually do, but now thousands are talking.
