Alexandre Oliva on Slippery Slop (LLMs)
I'm kind of growing tired of reading reports about how people start using LLM tools just for some specific purpose, and then end up surrendering their thoughts entirely to them.
An earlier blog post was fictional, but similar things seem to be very common.
People experiment with some questions, and eventually they're best friends with the machine, whatever that means. As thinking gets outsourced, brains figure out they can save energy, and the person becomes gradually impaired, to the point of eventually becoming unable to think, to write, even to read without the aid of an LLM.
Maybe they start using them for software development, just for code reviews. Then to generate code just for this one project that they'd otherwise never get to. Then to improve existing projects, but always reviewing changes manually. Then at some point the human steps out of the loop, and eventually it's all slop.
Communities of software development start accepting LLM-generated bug reports, then start accepting patches, and eventually their code base is taken over and rewritten by slop machines and their horribly messy coding style, to the point that humans can no longer even understand the code and LLMs with their brute force and infinite patience become indispensable to make further changes to the codebase, because humans run away in disgust.
High executives at tech and even nontech companies feel compelled to get involved with this mAnIa, despite the lack of evidence of any company but nVidia actually turning a profit out of LLM tech, and plenty of evidence of failure upon failure upon failure of projects that have embraced the hallucinating pursuit of automation. And yet, despite failure to experience benefits, execs keep on doubling down on the bet that it can't be the case that everyone else is talking nonsense and only themselves aren't succeeding in unleashing the superpowers they have been led to believe these systems would get them.
Even the companies developing these technologies can't help doubling down on their capital investment, despite the lack of perspective of recouping it, let alone turning a profit, because the moment any of them blinks and stops digging, the bubble bursts. If not sooner.
The general pattern is clear: there don't seem to be safe amounts of involvement with these tools. Whoever gets captured by the gravitational field of this tech seems to be dragged in further and further, with irresistible force, towards doom bottom.
It seems fitting to refer to this general pattern of progressively gliding towards slop machines as "slippery slop."
So blong, █
Copyright 2007-2026 Alexandre Oliva
Permission is granted to make and distribute verbatim copies of this entire document worldwide without royalty, provided the copyright notice, the document's official URL, and this permission notice are preserved.
The following licensing terms also apply to all documents and postings in this blog that don't contain a copyright notice of their own, or that contain a notice equivalent to the one above, and whose copyright can be reasonably assumed to be held by Alexandre Oliva.
This work is licensed under the Creative Commons License BY-SA (Attribution ShareAlike) 4.0 International. To see a copy of this license, visit https://creativecommons.org/licenses/by-sa/4.0/.