Wait, what!? Human slop?
@lech @salvadorbs The thread was about access to models for cybersecurity work. We replied there to criticize the way OpenAI and Anthropic handle access to those. The thread was a call for people to give constructive criticism about that for OpenAI and we did.
Someone responded to us saying projects using AI models to find bugs are AI slop. We disagreed and explained how bad the status quo is for core infrastructure including the Linux kernel and these tools are heavily helping to address it.
@lech @salvadorbs Our post jokingly referred to the millions of lines of bug riddled code in the Linux kernel as being human slop in the context of someone repeatedly calling it AI slop. It was humans who wrote the vast majority of the code including nearly all the bugs in it. Humans and AI models both make lots of mistakes and software cannot rely on that not being the case. Linux uses an architecture, language and many other design choices heavily depending on mistakes not happening.
@salvadorbs We were responding to someone who was already calling Linux and other projects slop for using AI models to explain it was already riddled with bugs before LLMs. It was very fitting in the context.
You should read what was written yourself instead of believing people on Reddit clearly misrepresenting what was said. Even with the posts taken out-of-context and most of what was said omitted, it's still clearly being heavily misrepresented and warped into something we didn't say or do.