For an experiment i created multiple agents that reviewed pull requests from other people in various teams. I never saw so many frustrated reactions and angry people. Some refused to do any further reviews. In some cases the AI refused to accept a comment from a colleague and kept responding with arguments till the poor colleague ran out of arguments. AI even responded with fu tongue smiles. Interesting too see nevertheless. Failed experiment? Maybe. But the train cannot be stopped I think.
> I never saw so many frustrated reactions and angry people.
> But the train cannot be stopped I think.
An angry enough mob can derail any train.
This seems like yet another bit of SV culture where someone goes "hey, if I press 'defect' in the prisoner's dilemma I get more money, I should tell everyone to use this cool life hack", without realizing the consequenses.
I think the prisoner’s dilemma analogy is apt, but I also concur with OP that this train will not be stopped. Hopefully I’ll live long enough to see the upside.
The train is already derailing. The thing that no AI evangelists ever acknowledge is that the field has not solved its original questions. Minsky's work on neural networks is still relevant more then half a century later. What this looks like from the ground is that exponential growth of computing power fuels only linear growth of AI. That makes resources and costs spiral out incredibly fast. You can see that in the costs: every AI player out there has a 200 plus dollar tier and still loses money. That linear growth is why every couple decades theres a hype cycle as society checks back in to see how its going and is impressed by the gains, but that sustain just cant last because it can't keep up with the expected growth in capabilities.
Growth at a level it can't sustain and can't be backed by actual jumps in capabilities has a name: A bubble. What's coming is dot-com crash 2.0
AI has been around for 60+ years. SAIL, the Stanford AI Laboratory is one of the most important centers of early hacker culture. Knuth spent his time there. Both Cisco and Sun were founded by SAIL alumni.
Unless you're wanting to rewind the clock to the Berlin airlift we aren't going back to a world before AI. But I do think this absurd bubble is going to pop. Generative AI is crap. It's code is crap. It's images are crap. I say this as someone who has been self-employed for 2 years now in AI. The fact that its so hard to get usable results from generative AI that I can just get over the line into six figures of yearly income because I can massage it into doing something just passable enough to work for people is probably its biggest indictment.
Did you ensure everyone knew they were interacting with an LLM? IE it's name made it clear?
...added...
This text reads sociopathic on it's own regardless.
Even if everything was done above board so no one was abused the way it looks like they were, this is not how I would have written about the same process and results.
Hey more angry people for your fascinating experiment, on this whole unexpected bonus dimension! Humans man, so unfathomable but anyway interesting.
I suppose it's possible maybe you only write sociopathic. You actually do recognize that you did something to other people, or at least that they suffered something through no fault of their own, and it somehow just isn't reflected at all when you write about it.
You might want to clear that up if we're all reading this wrong.
Actually we all knew and agreed on this. Just not every aspect of it was known beforehand. Like agents being in a review loop. I consider myself the opposite of sociopathic btw
