On 24/07/26 at 11:40 +0200, Pierre-Elliott Bécue wrote: > Lucas Nussbaum <[email protected]> wrote on 24/07/2026 at 10:01:43+0200: > > No, I don't always read the source code produced by AI. It depends on > > what I am trying to achieve. For example, sometimes I'm just trying to > > build a throw-away tool to solve a specific problem, and there's no > > reason to care about code quality. > > Fair point if you assume the tool can't have destructive actions.
.. or if I assume that it's contained in an environment where the impact of destructive actions would be manageable. Frankly I'm much more worried about sophisticated supply chain attacks made possible by the fact that the whole software ecosystem looks a bit fragile currently, with all the vulnerabilities in human-produced code found with AI-assisted tools; than about the scenario of an LLM generating buggy code that would have destructive actions. > I'm also really terrified by the cognitive decline that seems to arise > with these tools. Maybe, but at the same time, LLMs are good at doing most of the busywork of software engineering, leaving mostly more mentally-challenging activities to humans. So I fear for the future of those who use LLM to decrease their cognitive workload without focusing on those more mentally-challenging activities instead. (And that's also terrifying). Lucas
