Posts by artificialmind@fosstodon.org
 (DIR) Post #AhtfW8RUdhBpP3LPrU by artificialmind@fosstodon.org
       2024-04-24T12:42:18Z
       
       0 likes, 0 repeats
       
       @martin Yeah it really depends on where you are. I have to do some part time web dev in addition to my core dev tasks. The most time consuming part is vaguely describing what I want to do and hoping Google can find the arbitrarily named API of the js lib of the month I'm trying to use. That is a task that LLMs excel at, so I'm getting a lot of value there.
       
 (DIR) Post #AhvrXpZ5NrkzCVOHNw by artificialmind@fosstodon.org
       2024-05-06T09:14:47Z
       
       0 likes, 0 repeats
       
       hot(?) take: React and ImGUI are conceptually almost the same.(they might have started out vastly different but the set of strengths and weaknesses is quite close. both let control flow drive the UI, which is 'immediate' from the user perspective. React has more memoization/caching builtin and is assumed pure-by-default, while that is explicit for imgui. I imagine ImGUI and React would converge in Haskell.)
       
 (DIR) Post #Ak3kq1oGhR3amt2EOO by artificialmind@fosstodon.org
       2024-07-18T12:28:03Z
       
       0 likes, 0 repeats
       
       @robin Thanks, this gave me a lot to think about.Maybe you can help me with the following paradox:You make a good case that non-binding values & reserve control are counterproductive. Commitments must be concrete and measurable enough to allow accountability.However, this also immediately tingled my Goodhart instincts: Selection pressure for easy-of-measurability in ethics is a fast-track to dystopian results.
       
 (DIR) Post #Ak3kq31QC1AkXyCFX6 by artificialmind@fosstodon.org
       2024-07-18T12:28:13Z
       
       0 likes, 0 repeats
       
       @robin If my python bot becomes the most moral entity around by simply min-maxxing mechanical rules, then that's not a moral system I'd like to be part of.A quick search tells me that Goodhart's Law's relation to accountability is kinda known but I have no idea how you reconcile them in non-hand-wavy ways.
       
 (DIR) Post #AypI5PyGVeSSEhAOOm by artificialmind@fosstodon.org
       2025-10-03T07:59:29Z
       
       0 likes, 0 repeats
       
       @aras exponential volumetric dog
       
 (DIR) Post #AzXC5ZxDLHiwQbbajY by artificialmind@fosstodon.org
       2025-10-24T12:41:48Z
       
       0 likes, 0 repeats
       
       @musicmatze treat data like cattle, not pets
       
 (DIR) Post #AzYsLUZdVKE1bjo7Ki by artificialmind@fosstodon.org
       2025-10-25T08:10:01Z
       
       0 likes, 0 repeats
       
       @musicmatze but how do I rotate the secrets in my repository otherwise?(/s for those who need it)
       
 (DIR) Post #B1JUna6DVHQS4IJ5iy by artificialmind@fosstodon.org
       2025-12-16T17:49:42Z
       
       0 likes, 0 repeats
       
       @musicmatze I wonder if you can easily do a latex -> PDF -> embed into .docx pipeline. You know, as a form of malicious compliance.
       
 (DIR) Post #B1ZTnqFUjgIhCNJwSu by artificialmind@fosstodon.org
       2025-12-24T10:53:55Z
       
       0 likes, 0 repeats
       
       @musicmatze just configure double sized spaces in your editor.
       
 (DIR) Post #B1m71aQca4pgINaaI4 by artificialmind@fosstodon.org
       2025-12-30T11:47:53Z
       
       0 likes, 0 repeats
       
       @wolfpld @amonakov @wolf480pl Imho a valid criticism of printf is lack of extensibility. Having to drop 10 lines just to debug print a vector<int> is grating. I've had the equivalent of std::print (the c++23 one) in my foundational lib for years and got great mileage out of that.