I am so thoroughly convinced that anyone who thinks AI 100x's their output is a liar or a lunatic.
You are telling me you can make 1 years worth of decisions in 3.65 days? Let alone describing those accurately and coaxing the result from the AI... (1.8 days european time)?
I’ve mentioned this before: this is one of the oncoming trains for corp-security. We’ve long failed at least-privilege, but weren’t often punished for it.
Helen in HR (or Bob in accounts) didn’t know what to do with the extra perms they didn’t know they had.
Their agents will.
Future people will look back on current discussions about LLMs being conscious in the same way that we look back on Victorians discussing whether the telephone could be used to contact the spirit world.
I've got an agent in a loop optimizing a renderer with the goal to minimize frame times (and tests to measure). It got times down from 88ms to 2ms and allocations down from ~150K to 500. Sounds good, right? Wrong. This is exactly why agent psychosis is a big fucking problem.
As an experiment, I rewrote the Ghostty core render state in Go, with access to identically laid out data structures as Ghostty and the exact same validation tests. I made a purposely naive renderer (simple, correct, but slow). 88ms per frame with 150,000 allocations (horrendous, lol)!
I then kickstarted a Ralph loop to bring the frame times down. I told it it can't modify input data structures or the public API or tests (they're correct), but it can do anything else it wants. It got to work.
It has worked for about 4 hours. I've spent around $350 on this experiment so far. The results?
88ms => 1.5ms
150K allocs => ~500 allocs
Incredible right? Nope.
My hand-written renderer I ported has frame times (same benchmark) of ~20us (0.020ms) and 0 allocations in the update path.
This is the problem with psychosis and lacking systems understanding. If you don't understand the system, you're going to accept that this is an incredible result. If you understand the system, you'll see better solutions immediately and can do roughly 75x better on throughput.
The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity.
Standard disclaimer: I use AI all the time. I like AI. The point I'm making is to not blindly accept results. Think. Analyze. Learn.
No one could have known that telling programmers "spend as much money as possible on this new agent technology; whoever spends the most money wins!" would result in companies spending too much money on this new technology.
@HSVSphere@CompareAndSwap@kerckhove_ts It is a way better reference for obscure stuff. But for some important topics I think people will wind up with the kind of understanding you get from reading all of a math textbook but ignoring all of the exercises.
Clankers are now good enough at writing Haskell to not be terminally annoying for me to use. They spend a lot of time fixing their type errors, which is great. Total Haskell vindication in the age of clankers.
@kerckhove_ts I feel bad for the kids these days though. The clanker-do-it-for-me button is even more tempting for these things with steeper learning curves. This means it takes crazy willpower to actually learn.
Feels like I was on one of the last choppers out of Saigon.