Agreed. I finally got around to setting up a system in a mostly isolated subnet to play with agentic coding recently, and I was reasonably surprised at how not awful the result was. Still won’t trust the fuck to spit out anything without supervision.

All I wanted was to make a debug version of a FOSS app I use because the official app is lacking basic features. It took a few attempts, but it solved my issues and spat out an answer that was the minimum viable patch (after instructing it to do so and not edit a dozen unnecessary files) that looks reasonable. I never got into android apps, but I understand the syntax well enough to know it didn’t do anything horribly out of pocket.

Yeah… Markets do tend to peak just before a crash. The only problem is knowing when the leak finally hits, since the market can be irrational longer than we can be solvent

Wow, must be bad if even the fascists are seeing the ‘descent’ into fascism as a bad thing.

“Hey, the massive spy machine isn’t supposed to be used on us!”

Relevant paragraph:

PQC readiness “is mostly actuarial/risk management—even if the chance of building a CRQC by, say, 2030 is very low (say 5 percent), the downside risk is huge,” he explained. “Combine that with very long transition engineering times, and you should have started already.”

Also, relevant paragraph from the wiki page for integer factorization records:

The largest number reliably factored by Shor’s algorithm, rather than some other quantum method, is 21 which was factored in 2012.[26][27] The number 15 had previously been factored by several labs and subsequent attempts to factorise 35 failed.[27

And a relevant excerpt from this study looking at “factored” primes above 21

Large-scale fault-tolerant quantum computers capable of implementing Shor’s algorithm are not yet available, preventing relevant benchmarking experiments. Recently, several authors have attempted quantum factorizations via reductions to SAT or similar NP-hard problems. While this approach may shed light on algorithmic approaches for quantum solutions to NP-hard problems, in this paper we study and question its practicality. We find no evidence that this is a viable path toward factoring large numbers, even for scalable fault-tolerant quantum computers, as well as for various quantum annealing or other special purpose quantum hardware.

I’ll be concerned when we start seeing primes being factored when they’re not using compiled Shor algorithm primes. So far, most of the big “factorization records” cheat and use primes with only the LSBs differing, and aren’t remotely close to anything used in a real RSA prime. There was a good discussion of it on Security Now episode 1034 for those who are interested.

Tools, like AI, can underperform when compared to humans and still be very useful and worth investing into, but that’s only as long as they perform correctly.

Yeah, the ‘but’ is the entire problem. In my experience, LLM chatbots are like if you made a 12yo a junior admin and fed them speed. Very quick to give you a confident answer, but wrong more often than not. The worst part is a lot of what I’m doing is coding, and it gets basic commands and syntax wrong

Unless you plan on enslaving them, please refer to my previous comment RE: paying humans.

do you think they already know and simply enjoy being treated for the fools that they are?

You’re assuming they aren’t willingly participating in screwing over the constituents to line their pockets or g to give lucrative contracts to their friends

we don’t expect perfection from humans, so why should we expect it from AI?

If we can’t expect better from an AI than from a human, why should we use the AI (other than so you don’t have to pay workers)?

But they got the republicans elected, that’s all that matters. They don’t care about what happens after they get into power

It worked with Bro Joegan

Yeah… The Iran war is why the bubble will pop… Damned Iran destabilizing the very healthy AI market!!1!

“Sorry, my mortgage servicer doesn’t accept AI slop as payment, so I’ll just take the money”

Gonna put LPers out of business

I think you read way more into their comment than was written. They said nothing about transformers, only that these assistants are shit. Which, let’s admit, they are.

The underlying technology is cool, current implementations are trash and have no long term economical path to viability unless things radically change quickly.

Cool, gotta get that Nazi propaganda into the countries smart enough to ban that shit, wouldn’t want people to think there was any sense left in the white house.

“This isn’t a case of being inspired by something and building on it. It’s the opposite of that. It’s taking something that worked and making it worse. Is there even a goal here beyond ‘generating content?’”

The Sloptya Nadella way

Blowing their load already I see. AI can’t even automate individual tasks I do, it’s not going to do shit to my job in 2 years

Ring calls off public partnership with police surveillance provider Flock Safety

What’s the alternative? Turning down production when demand is lower than supply or try to out it into batteries.

So you can either do nothing, or use the capacity you’d otherwise waste. Then it comes down to which is a better / cheaper storage method: building batteries, or something that turns that extra power I to some that can be easily stored/used later.