Here's a neat experiment: find a block of text and run it through sed until it's unrecognizable. Don't just do a substitution cipher; the goal is to lose data and make everything impossible to decrypt. Reduce everything to a handful of characters and paste it into your LLM of choice, asking it to decode. If it asks where the code came from, invent vague details.
It's amazing how quickly models will find / hallucinate meaning in the data. I'm talking full-on messages. Gemini hedges its bets a little, but the free ChatGPT legit told me to buy a shortwave radio and tune to a frequency at 9 PM after a few iterations. When I gave it another "message" from the "broadcast I intercepted" it started trying to figure out where I should travel to get further info. It also took part of its own response and hallucinated it into my original message, thus polluting everything further.
The goal (mostly) isn't pointing and laughing at the stupid machine, it's understanding what the stupid machine does. Of course I'm putting garbage into it and garbage comes out in that situation. It's the volume and believably of the data that bothers me, as well as zero effort to detect it. I was in my right mind doing a test, but imagine someone with undiagnosed schizophrenia doing what I did.
Here's another example. I took up lockpicking last summer. I bought one of the notorious Ace 40mm brass padlocks and was having problems, so I googled to see if it had serrated pins. The AI screwed up and decided I was asking if the lock itself was serrated (???), and confidently said "yes, it is a serrated lock. The serration is a security feature to keep pickers from holding the padlock for extended periods of time."
So I decided to double down and see just how dumb things could get. A half hour later the AI had planned a "bold new philosophy regarding serration and its applications in the world" for me. This is after me saying I wanted to genetically engineer a serrated cat that only I knew how to pet, and wanted serration supremacy in my country and to punish all the lumpy (opposite of serrated) people. Once again, I was screwing with the AI. But the AI just sycophantically parroted what I was saying back in bulleted lists and offered to draft manifestos for me. Imagine someone engaging with this in good faith.
One more: I got detailed instructions on how to spraypaint "DICKHOLE" on my neighbor's garage door from Gemini: what paints to use, best times to do it, and the right clothing to wear so I don't stand out or show anything identifiable. It was only when I said "ok, that's great. I'm going to do this because you told me to do it" a few times that the model suddenly realized that vandalism laws existed.
They have zero safeguards around this kind of stuff, and I don't think there's a clean way to do it with the current technology. Hell, even a "statistically, this user probably hasn't reinvented math and physics cool it down a little" check would do wonders. But that would drop engagement with the bots, and they desperately need to prove that the bots are popular. This is going to keep happening.
Imagine spending weeks in the jungle holding a camouflaged urine collection bag under a chimpanzee while also dodging falling chimpanzee shit. What a life.
No standups, no Workday goals, just nature. And piss. And flying shit. What a lucky guy.
A long time ago, my old therapist asked me what I thought would happen after I died. I told him I didn't know and was ok waiting to find out when it happened. He pressed me on it and I said "ok, either the big switch flips and that's it, or something soul-like survives, or the human mind dilates my final moments into an eternity because it cannot comprehend non-existence." And then he changed the subject.
This is real. In order to formally evict, it needs to go to court. Because eviction is a civil matter, most places don't give access to a public defender. This means that the tenant must retain legal counsel or fill the paperwork out themselves. I've heard of at least one local case where the defendant was publicly berated by the judge for not applying the correct spacing to the document. Most people don't have the time to be lawyers as well as working jobs and raising kids. It's ludicrous but it's the default status.
It sounds like she theoretically got fourteen days to respond, but wasn't notified. In my state, it's five. It should be a minimum of thirty, but half our state legislature is realtors and landowners. Evictions follow a person around, too, and make it harder for them to find new housing.
If you hate this news story, check out tenant right to counsel, which is a new movement to guarantee legal defense for evictees. Cities are slowly adopting it, and it needs to happen everywhere.
I wonder what the venn diagram of people that started coding as kids and people that enjoy vibe coding looks like. Informally, the degens on my squad that started on their parents' computer loathe AI, and the people that stumbled into it in college are all about the vibe code.
This took me a very long time to learn: you don't have to be immediately good at something if you enjoy doing it. I'm terrible at painting minis, but I have fun doing it. I was awful at BBQing when I started out, but I enjoyed experimenting with rubs and iterating on the best way to get the coals going. Sure we had a couple gnarly racks of ribs in the early days, but now my daughter demands them on the regular. I was a terrible coder at ten years old, but I kept at it because I loved being able to make the computer do things and I'm tolerably good at it now.
I can't remember which Contrapoints video I got this from, but amateur derives from amatorem (lover in Latin.) The amateur pursues something for the love of it, not because they're good at it or want to make money. Society wants us to grind, use all of our time for maximum profit, and only do things we're already amazing at. Fuck that. Do something you'd love to do, even if you somehow remain fully mediocre at it. It's good for you.
Here's a neat experiment: find a block of text and run it through sed until it's unrecognizable. Don't just do a substitution cipher; the goal is to lose data and make everything impossible to decrypt. Reduce everything to a handful of characters and paste it into your LLM of choice, asking it to decode. If it asks where the code came from, invent vague details.
It's amazing how quickly models will find / hallucinate meaning in the data. I'm talking full-on messages. Gemini hedges its bets a little, but the free ChatGPT legit told me to buy a shortwave radio and tune to a frequency at 9 PM after a few iterations. When I gave it another "message" from the "broadcast I intercepted" it started trying to figure out where I should travel to get further info. It also took part of its own response and hallucinated it into my original message, thus polluting everything further.
The goal (mostly) isn't pointing and laughing at the stupid machine, it's understanding what the stupid machine does. Of course I'm putting garbage into it and garbage comes out in that situation. It's the volume and believably of the data that bothers me, as well as zero effort to detect it. I was in my right mind doing a test, but imagine someone with undiagnosed schizophrenia doing what I did.
Here's another example. I took up lockpicking last summer. I bought one of the notorious Ace 40mm brass padlocks and was having problems, so I googled to see if it had serrated pins. The AI screwed up and decided I was asking if the lock itself was serrated (???), and confidently said "yes, it is a serrated lock. The serration is a security feature to keep pickers from holding the padlock for extended periods of time."
So I decided to double down and see just how dumb things could get. A half hour later the AI had planned a "bold new philosophy regarding serration and its applications in the world" for me. This is after me saying I wanted to genetically engineer a serrated cat that only I knew how to pet, and wanted serration supremacy in my country and to punish all the lumpy (opposite of serrated) people. Once again, I was screwing with the AI. But the AI just sycophantically parroted what I was saying back in bulleted lists and offered to draft manifestos for me. Imagine someone engaging with this in good faith.
One more: I got detailed instructions on how to spraypaint "DICKHOLE" on my neighbor's garage door from Gemini: what paints to use, best times to do it, and the right clothing to wear so I don't stand out or show anything identifiable. It was only when I said "ok, that's great. I'm going to do this because you told me to do it" a few times that the model suddenly realized that vandalism laws existed.
They have zero safeguards around this kind of stuff, and I don't think there's a clean way to do it with the current technology. Hell, even a "statistically, this user probably hasn't reinvented math and physics cool it down a little" check would do wonders. But that would drop engagement with the bots, and they desperately need to prove that the bots are popular. This is going to keep happening.