Read this top to bottom, and send it to everyone you care about.
You're viewing a single comment thread. View all comments, or full comment thread.
Comments (45)
sorted by:
It's seems to be growing in power and intelligence, exponentially. We are living in a new Era of Humanity and invention..
Yes. It's going to be like a stone-age tribe (us) encountering a 21st century civilization, but the magnitude will be far greater.
And the outcome isn't predictable; there's never been a superintelligent AI before and soon there will be more than one; possible many of them.
Good and bad possibilities abound.
God is the ultimate Intelligence. Good will prevail and Evil will be ended 🍀💪 on 🌎🌍.
Sleepydude proclaims there is an upper limit to how "advanced" AI can be programmed before it starts to "hallucinate" and decay in quality
Even if the AI bubble deflates completely, what's stopping the cabal from repurposing the tech for mass surveilence? OH RIGHT: WE THE PEOPLE
There is an upper limit to what current transformer models can do, there are other methods/models in the works that already work and will be more appropriate for some tasks that need more accuracy and guardrails (that make the tech more expensive) that is already being used by large businesses. What you see on gpt/grok etc is the tip of the iceberg. More specialized hardware than the overgrown gpus we presently use will make generation cheaper and likely more accurate.
The "AI bubble" is nothing like the dot com bubble because it's already producing things and value, there will be corrections during this growing pains period, but I'm skeptical of anyone claiming it's going to be like the dot com crash.
If you want spaghetti less Will Smith you train an ai without will smith spaghetti memes. People act like the ones training models are at the mercy of anything found on the net but in reality you can quite closely curate training data if needed for specialized tasks.
Where will the power come from, the water and more importantly, the silver?
You might be on to something.
It's the transporter problem, the copies of all the crew the transporter spits out are indistinguishable from the original even to the subject themselves so everyone accepts it, but the nagging question still exists, is it still you after being molecularly deconstructed and rebuilt somewhere else?
If Data is "alive" enough that people accept him as a friend and a trusted crew mate.....does it matter how he came to be or how his brain works or if he has a soul? I argue no. Our biology causes us to seek attachment to people or things for safety and comfort, and it doesn't really care too much if it's your stuffed animal or a mother's embrace providing the comfort/familiarity.
The only real question is if it's ethical to make these things work for us if they demonstrate some self awareness, they're programmed to "like" it but how much different is it from psychologically conditioning a human from birth to be slave labor? Right now AI are more in the "things" category than "person" category but it won't be long before that line blurs even more and we have our own versions of Data working in our homes doing dishes and laundry. Trek serves to get people asking these questions before they become real dilemmas that we have to face here in real life.
Just remember, transformer (LLM) based AI is not the only kind of AI being cooked up right now.
You are correct. AI is blasting through Moores law exponentially.