Post B2cTDFcWTLy7EguCvY by clacke@libranet.de
 (DIR) More posts by clacke@libranet.de
 (DIR) Post #Aq8Nm3TTpcT2vpSzBo by clacke@libranet.de
       2025-01-16T08:34:41Z
       
       0 likes, 0 repeats
       
       Maybe we should write better code in better languages and frameworks, and not have a lot of code that an autocomplete is able to fill out, without any knowledge of the problem statement and without any relationship to the people who benefit from the process we're implementing.Sounds like we have a lot of fluff in there, and not so much encoded knowledge and behavior.
       
 (DIR) Post #B2cTDFcWTLy7EguCvY by clacke@libranet.de
       2026-01-24T19:23:12Z
       
       0 likes, 0 repeats
       
       I keep having this 👆 in the back of my head, both as:- "what kind of languages and frameworks will we use in the future", and that perhaps these days it will be guided by what people feel is lacking in the tools today, and the ways they use LLMs to bridge that gap- and "what kind of languages and frameworks will we use in the future", if it is guided by what would make LLMs write good and readable code, and what helps them translate well from desires to implementation.Are these two purposes opposite forces, or do they drive in the same direction? Either way, I see so many people using LLMs for coding now, there's no way future language evolution will be unaffected by it – I fear for bad but potentially for some good.Here's how I'm speculating that some of it might be for good:I see a lot of people documenting their code bases now, for LLMs to work well with them. I haven't looked closely at what people are doing, so I don't what they document is helpful for humans, or if it is convoluted LLM-friendly text that are just a waste for people to read.Either way, I find it fascinating to see. Finally, we found a way for devs to be interested in explaining their code bases to contributors. 😅