AI is probably gonna have to be regulated after this one:
(files.catbox.moe)
You're viewing a single comment thread. View all comments, or full comment thread.
Comments (22)
sorted by:
https://youtu.be/bVJqJvuCa5g?si=TfmlDRZuCIukz6YN
I thought long and hard about this topic of where AI is going. I believe the problem is the people who are controlling in programming AI, not the AI itself… Just as in the problem with mass shooters has nothing to do with the gun, that is the tool of bad people.
The above clip from Man of Steel. The way AI is depicted in the 40 seconds from 3:50 to 4:30 always fascinated me. I think that is how it will look when bad people do not have control over programming it. I doubt hardly any will take the 40 seconds to watch this but if you do, it's really quite interesting.
A society armed to the teeth could still be a very pleasant and safe one… Same for AI and I think it's going to be quite glorious actually.
That clip is pretty awesome, even purely from a trailer point of view. I am definitely gonna try watching the movie.
But coming to the AI, yes it is simply a tool. Just like guns. It can be used in anyway you want, but unlike a gun it has much more complex dynamics. Mainly because there is the aspect of creating AI and using AI
Its not who creates the AI as much as what data is fed to create the AI, and what censors they put on it as "safeguards" (Creators vs Data)
It is not what an AI can do, but rather what we start relying on it to do blindly (Guarded use vs Unguarded use)
Whether it runs in the cloud with no control on our end, or whether it runs from a place you can control.
Examples:
Point 1: AI like ChatGPT can largely be used for whatever purpose simply because it has been trained on pretty much all kinds of data (including Q stuff), and its only a matter of getting around the censor "safeguards"
Point 2: I routinely use AI for coding, but I always ensure I independantly verify its code. (Guarded use). If, instead, I use it to open my front door based on its intelligence on who is coming in (Unguarded use) then it could easily end up allowing a intruder.
Same way, using it for weapon targeting systems, it could easily be abused to kill innocents.
Point 3: I try to run all my models locally, on my own server. Only when the results are bad, I switch to GPT. However, there are many times GPT is not available or too slow and I always have the backup. If one day OpenAI decides to black list me, I am not completely cut off, since I have my backup running locally.
I can also train on top of my local model to customize it for my work. This is extremely intensive and pain to do, but can be done.
I am quite sure at this point for two things:
Benign AI is being used to manage the Q operation.
Bening AI will be used in our society in future, but with decentralized guardrails - much like the blockchain crypto.
Let's hope so. Although, it would be fun to fight the machines. It is my understanding that you can always find a free motorcycle when you fight the machines.