Reminds me of that scene from I Robot where Will Smith was not supposed to drive his car himself.

Didn't this just happen in an internet company's self driving car with couple kidnappings in the past month? (They were done according to the companies terms of service is what the internet company says if I remember reading correctly)

Judging from the website, Muse seems to be too keen on making me buy things - Book Tickets, Buy a Stroller. Would this personal AI spam me with ads randomly like this?

having used it for a bit, yes, one of the first things it asked me for, unprompted, were things to set it up to buy stuff. i declined but it still consistently wants to spend money; research a trip? as soon as it has any results it prompts to start booking stuff. etc.

Of course it does! Really, though, what it really wants to do is make restaurant reservations. The primary concern of AI researchers has always been, and will always be, making restaurant reservations.

Parsing natural language and solving climate change are downstream of making restaurant reservations.

Universal Paperclips should have been about making restaurant reservations.

but can you imagine the horror of just showing up at a restaurant and asking to be seated? It's surely worth it

The Target marketing team famously could predict when a woman is pregnant but it was too creepy to, uh, Target them with only baby goods in dynamic marketing materials so they would mix in a few masculine things like lawn mowers into the catalog. No idea if that is what Meta is doing here but it's one explanation.

I've always admired people who can live without a smartphone. I met one linux contributor at an open source conference who doesn't have a smartphone. He uses an old dumb phone, doesn't have any socials and always carries cash! In a country like India, where every payment runs on a QR code, it's crazy that he can survive without it. He asked me to look up directions to the metro, memorized it, and left.

One other thing is that he isn't socially awkward. He could go up to people, and ask for help, which I think a lot of us tend to avoid, and instead use apps to solve them.

I wonder if the convenience of a smart phone makes social awkwardness worse. As in: I'm already not very socially comfortable but my phone makes it that I never get any practice either.

This whole tweet looks LLM-generated. It's the same old bland format and speech-styles that chatgpt answers with.

To some extent we are getting more things done as well. In my company (mid-sized startup), they're making us push features every other day now as opposed to maybe 1-2 features per person per sprint. Back when I joined, things were a lot slower. Today, they expect freshers to push new features on day one.

But ofc, slop has increased a lot more as well.

That's very fast. Feels like you cannot really review this code so it's all just "working" AI code with few railguards and few human supervision?

Jumping on the great asymmetry of output between humans and llms, yet still being responsible for the product is pure kafka.

And how many of that features are actually used?

How long do they last until the get replaced by the next feature?

But isn't that the point the writer is making? They'll be self-sabotaging themselves once they do that.

I don’t exactly see how, if there’s no exploit.

Either way there’s other realms it’s relevant like armed drones policing a street. Without a way for the public to activate a shutdown sequence, it becomes very difficult to stop if it starts murdering innocent children on some glitch. Which you cannot have because then the whole purpose is ineffective. Likewise digital communication is purposeless and face competing mediums for sure

hmm fair. so a foolproof backdoor that cannot be exploited by the enemies. I don't know how much of that is possible, but then again, the author is talking about super-secure-no-exploit software. So by that definition, a super secure exploit-free backdoor is also possible.