Post #2795998
2026-03-12 04:15 UTC
@soph@grrl.me Re theft, I'm just at the beginning of learning about models developed with a care to the inputs, like this one: "Compliant Apertus is trained while respecting opt-out consent of data owners (even retrospectivey), and avoiding memorization of training data" https://huggingface.co/swiss-ai/Apertus-8B-Instruct-2509
Re higher new kinds of development, this helped me accept a future with both development as we've known it, and the newer kinds of development AIs (LLMS + these new 'harnesses') enable. https://www.caimito.net/en/blog/2025/12/07/the-recurring-dream-of-replacing-developers.html
Replies (1)
-
@slowenough@mastodon.social 2026-03-12 17:04
@soph@grrl.me "Conversational approaches to coding could be beneficial." Haven't coded in forever, but it was instant & ongoing my awe at how central it is to any reasonably complex project, to talk and talk with people about what they actually want, because we are so often unable to specify what we want. First getting clear how to start, and then accepting it keeps changing. It's wild to watch some coders love LLMs so much, and others hate it so much. Maybe the most religious programmers' war, ever.