Tootfinder

It only takes a very rudimentary understanding of economics to see that, in the long run, the market price of running "agentic AI" (or whatever the latest buzz becomes) and the salary of any kind of knowledge worker can only be separated by a gap that's dependent on *management's* ability to tell their output apart.
There is a cosmic-scale gap between the long-term value to a company in paying a salary to a human vs. getting an LLM to "do" the "same" work, but it's a gap that many people are investing heavily in hiding right now, and CEOs and managers become incredibly gullible when you promise them they won't have to deal with humans who might ask for raises, even though it should be obvious you plan to jack up prices once they're locked in. They also aren't well-known for a deep understanding of how their workers actually generate value, and tend to be easily distracted by shiny but shallow demos, which is exactly what LLMs can produce well.
Even in the most optimistic scenario I can imagine for these developments, the myth of upward mobility that's been a lynchpin for defusing class solidarity among wage laborers is going to take a significant beating. There's a completely obvious game plan underway to try to use fascism to shore up the power of the wealthy against any kind of mass movement, and to play the ancient divide-and-conquer game along race and gender lines just as usual, so the range of outcomes is huge, but every time you use an LLM you're pushing this closer to a sharper conflict.
#AI #LLMs

Trying out a new license clause and disclaimer:
4. This software is NOT used as training data for a model of code or text structure or construction, such as a large language model, transformer model, reinforcement learning agent, etc. If such a model is generative and predicts or outputs the code of this software in whole or in part, that shall be evidence of infringement of this clause of this license.
THIS SOFTWARE SHOULD NOT BE USED AS TRAINING DATA FOR A STATISTICAL MODEL OF CODE OR TEXT, NOR SHOULD STATISTICAL CODE OR TEXT MODELS BE ASKED TO READ, INTERPRET, OR OTHERWISE PROCESS ITS SOURCE CODE OR OUTPUTS. THIS SOFTWARE MAY CONTAIN CODE OR TEXT OR MAY PRODUCE OUTPUTS WHICH CAUSE MODEL ROLE CONFUSION OR WHICH OTHERWISE CAUSE UNANTICIPATED CONSEQUENCES WHEN PROCESSED.
Would genuinely welcome any thoughts about wording.
#NoAI #LLMs #AI

An excellent exemplar of objectively terrible research that can only be explained by completely irrational levels of LLM hype:
#AI #LLMs

I keep beating this drum, but...
It's.... Interesting that we're having an entire conversation about "AI alignment" when Sam Altman and Darius Amodei have flat out said that their goal is to make paperclips out of everything even if the human race goes extinct in the process.
Oh I'm sorry, dollars. They're making dollars, not paperclips. That changes everything, clearly.
...
What's more insidious is that even if you Luigi'd Altman and Amodei tomorrow, OpenAI and Anthropic would happily elect new leaders and continue with business as usual. There is no shortage of humans willing to step up and run the paperclip machines and who have the same misalignment issue (they're not actually a large percentage of humans, mind, but there are plenty of them). So the thing you have you defeat in order to stop the paperclipization is in fact a corporation (and the idea that wealth makes right behind it), not a human.
It's almost as if we're *already* locked in an existential battle with an inhuman entity whose effective goals are dangerously misaligned by nature... 🤔
...
I wonder if the same techniques used to defeat the entity already threatening us would work against the hypothetical threat it's waving around to distract us? Only one way to find out!
#AI #LLMs