Replying to
@khleedril @shanselman ooh! You probably right. Never thought about that. Quite possible we achieve this Archeology quest
Remote
Andrii Kurdiumov
@kurdiumov@hachyderm.io
mastodon 4.7.30 Followers
0 Following
12 Posts
Joined November 22, 2022
RE: @fsinn@mas.to
Never thought that Waymo is like “our comfort is more important then people who for some reasons don’t have money, or god forbid like to walk”? If you cannot build AI just bust please. No competence no money
Replying to
@timnitGebru@dair-community.social Ooh, yes, I agree. There some parallels obviously. Probably I do not get fully your initial message and anwsering partly to voices in my head.
The Zig to Rust rewrite is interesting but one number does not tell whole story. So I decide to take a look at how Bun was tested based on public information. Here a bit of numbers
https://kant2002.github.io/en/llm/2026/05/16/bun-pr-analysis-en.html
Next article about #obfuscators . Now how to build control flow graph from IL and how to use that for complicate control flow graphs
https://kant2002.github.io/en/obfuscators/2026/05/02/how-to-build-obfuscator-part-ii.html
Update NixOS support for FSharpPacker! I did not expect that from myself, but if code is rotten, it should be improved.
A friend of mine wrote a book systematizing typical architectures, their names, their common properties, and the differences between related architectures.
So, while Denis is arguing with other in the Ukrainian architects' chat that he shouldn’t charge for his work, and other beg him to take money, you can read it—it's all free.
Website - https://metapatterns.io/
Book - https://leanpub.com/metapatterns
If you liked the book, please donate.
Replying to
@timnitGebru@dair-community.social Isn't this weak point? Yes, if we assume that programming should be formal, and give precise results, yes. But society have less precise rules (laws) to give precise results when interpreted by humans (judges). Moreother lack of precision, is exactly which allow the process to be more or less stable, and society to learn laws, and eventually understand how to interpret them. I do not assume that using imprecise interpretable rules is way to go, but we do as society have such structures, and they works too. I believe statement works both ways.
Replying to
@timnitGebru@dair-community.social I would speculate that for some part of computing landscape the programs, is kind of automation business processes, which are essentially interactions between people, and initially these business processes was executed by humans. Right now obviosuly lot of these processes hyper optimized, but seems to be this hyper optimiation require lot of smart people to operate, whcih is somewhat detremential to society IMO. And if whole landscape would be hyperoptimized going further on strictness have diminishing returns (possibly) and so that's why pushback on strictness really welcomed by some group of people since it essentially remove burden from them. In a sense I view "Let's be very precise all the time, and control lot of things" as ableist statement to some degree. Another observation is that we have business people and manager use human-human interface to achieve more or less reliable results. And sometime software developers are really unreliable, cheating, incompetent and at the same time these group of people achive more or less stable results. So again, I don't think naїve dismissal is that easy.
Replying to
@mathew@universeodon.com @JohnMashey@mstdn.social @timnitGebru@dair-community.social i definitely know that laws don’t ideal. And they definitely cause harm. I choose laws and management not because I like these disciplines as software engineer. Precisely because they irritates me a lot. The point I was trying to make, is that we have examples how putting humans in the interpretation loop can solve some inefficiencies(maybe large one) to our own benefits