My coworkers by and large dislike using LLMs because they can feel their skills atrophying from disuse but at the same time the pressure from above to comply and use “AI” as much as possible is overwhelming. Layoffs target those who speak against uncritical adoption of LLMs as the future, and those who are insufficiently accelerated by AI. That metric is judged by how many pull requests you deploy to prod per day, so everyone is now using LLMs to write hundred-liner PRs with a subtask for each one and slamming the reviewers to get them merged so they look good in the (easily gamed) metrics.
My coworkers by and large dislike using LLMs because they can feel their skills atrophying from disuse but at the same time the pressure from above to comply and use “AI” as much as possible is overwhelming. Layoffs target those who speak against uncritical adoption of LLMs as the future, and those who are insufficiently accelerated by AI. That metric is judged by how many pull requests you deploy to prod per day, so everyone is now using LLMs to write hundred-liner PRs with a subtask for each one and slamming the reviewers to get them merged so they look good in the (easily gamed) metrics.