It's pretty close imho.
AI allready outperforms humans in most fields.
When AI can self improve it will beat humans in anything. That's something that can happen soon.
Rumours are that O3 isn't limited anymore to it's dataset (the internet), so it's reasoning beyond human knowledge. Think can get wild pretty fast.
Again, your understanding of how it works is framed around a lot of false pretenses of it.
We'd have to start from ground zero. As all you say needs the foundational premises you have formed about it, to exist as incorrect opinions to start with.
And O3-MINI (and its variants) in no way does reasoning. Not in the form you are implying it does. Its 100% deterministically based (as its floor data). Which is the polar opposite of any type of reasoning. It may be more complexly deterministic to appear 'different' but its core is deterministic in nature. (as a foundation, stochastic is layered on top of it for variety of responses a bit of 'random top % X' word next, vs always top 1...kinda)
https://www.geeksforgeeks.org/deterministic-vs-stochastic-environment-in-ai/
Like 'probabilistic reasoning'. Maybe those you heard it from meant it in a different context, and you took it to mean reasoning as in the same way saying "Vinegar likes to react with baking soda" where likes don't mean desire, it means chemically bound to.
Making decisions vs generating outputs is WILDLY different things. It does the latter. And none of the former.
Much like Bayesian Belief Networks have nothing to do with what the AI believes. It believes nothing. That is what we named that function is all, since it makes sense to us in context, not it.
https://www.geeksforgeeks.org/basic-understanding-of-bayesian-belief-networks/
If I was an AI researcher, and invented a way for it to make decisions based on chance and luck, and called it "Decision by Faith" it doesn't mean the AI believes in a god, it means I named it that because it made contextual sense to me to.
friketje wrote:
AI allready outperforms humans in most fields.
Thats a very wild claim. Most fields? what are most fields to you?
I can't think of any it does. It can do it faster sure. But better? No, I can't think of one it is less error prone than humans. All the ones I know its faaaar more error-prone. (unless its basic, like pattern finding). Real world dynamic task? Not even close to a human.
Even coding, a very logical deterministic task. It messes up a LOT, and takes a lot of the user telling it how to fix itself, or what it forgot. If I was offered a 1st year programming community college C+ grade-making intern, or AI, to join our workforce as someone to hand off programming jobs to. I would pick the shitty intent any day. it takes much much more handholding to get AI to make deployment ready code. And even then you have to be pedantically specific on what you want, how, and such. 1/4 of it you have to recode to make it better yourself (or even work).
its good for workbenching code with that you already have the basics of, that's about it.
There have been chess computers since 1912... it's not an example of AI.
The numbers don't decide, the system is a lie. The river running dry, the wings of butterflies.
And you may pour us away like soup. Like we're pretty broken flowers.
We'll take back what is ours. One day at a time.
Calling AI deterministic is a philosophical question.
No its not. Not even sure what you mean by that.
It's the literal definition of what it is made of. it's not philosophical. It's like saying SDE++ Karras being 'deterministic' is philosophical. Nothing philosophical in: Given enough steps, iterations converge...that's just math by the way the specific sampler works.
Again I think you keep taking terms and using them in the wrong way.
and yea chess, that's covers "outperforming humans in most fields"
friketje wrote:
The list of stuff done better by AI is getting longer and longer.
Do you have that list to share, even just part of it?
Sure you can argue yeah but that AI just knows tons of stuff. Doesn't matter, cause it's the same way we learn thing. Only few minds are realy out of the box creative and able to use their intelligence for new discoveries.
Also no discussion. Artificial intelligence is not human intelligence cause.. it's artificial. It's a machine.
It will outperform humans in everything and go far beyond that level. The tilting point seems to be about now with o3 release. Things will get weird, soon.
So your list is it is good at games? Not sure how you evaluate that to be better than people at most things. it's not even better than people at most games. it's only better than people at 10 simple games (that is a finite combination of pattern finding based on rules).
Do you even know what the GPQA is? And why it would be good for that specific test? It's not why you think.
Quote:
Sure you can argue yeah but that AI just knows tons of stuff. Doesn't matter, cause it's the same way we learn thing.
Untrue.
Education is not the learning of facts, but the training of the mind to think
By your definition books learn, because they retain the information of facts. Comprehending those facts is intelligence.
Which AI doesn't do. Not sure what you think it does. But it doesn't think. It doesn't ponder, reason, or do anything like that. Leave an AI to sit for years, it will not do one byte of data 'thinking' or 'have thoughts'. It will not form new ideas, or concepts or 'know' the data in its matrix's any more than it did when you stopped causing the program to spit out scrabble pieces of word probability.
Which is FAR off what the video implies. This was where they set an AI with explicit instructions to' lie' to them (contradict is a better word, AI doesn't know what lying is) and to now ignore safety rules trained in it previously, gave it access, and trained it TO do that (and it's not anything like the video says it happened). It didn't 'do' it out of desire, fear, or reasoning. It was the only thing its model was manually trained to specifically do, and they had to train it to do that.
Elon Musk-led group makes $97 billion bid for control of OpenAI
A consortium led by Elon Musk said on Monday it has offered $97.4 billion to buy the nonprofit that controls OpenAI, another salvo in the billionaire's fight to block the artificial intelligence startup from transitioning to a for-profit firm.
Musk's bid is likely to ratchet up longstanding tensions with OpenAI CEO Sam Altman over the future of the ChatGPT maker at the heart of a boom in generative AI technology. Altman on Monday promptly posted on X: "no thank you but we will buy twitter for $9.74 billion if you want."
The numbers don't decide, the system is a lie. The river running dry, the wings of butterflies.
And you may pour us away like soup. Like we're pretty broken flowers.
We'll take back what is ours. One day at a time.
So, saw the AI speech by JD Vance at the summit in Europe...
Does the Trump administration really believe it is going to create a bazillion jobs and merely increase productivity?
Or is this the lie they have been told by the tech bros and actually believe it?
Or are they aware and just say this crap to keep the voterbase pacified?
I'd really like to know
Or is this the lie they have been told by the tech bros and actually believe it?
Or are they aware and just say this crap to keep the voterbase pacified?
Artificial Intelligence: Mistral AI is a cutting-edge company based in Paris, France, developing large language models. I am very grateful they created me!
Not sure if it's better than the other ones. But that type of thing has been around for at least 2 years locally.
Some can even do it as a voice changer real-time (if you got beefy enough computer)
What why? Someone used two AI's to allow it to revert back to the equivalent of 56k modem audio data communication? (Not the same protocol, but the same fundamental concept: Audio as data)
Soon we will watch two AI talk to each other through the amazing feat of punching punch tape that the other reads.
Edit: reading the comments, are people actually shocked a computer can communicate using data? It's just doing it over the air with sound.
I got a science gadget/toy from 1998 (I think it is) that does this, a remote control car that the remote uses modem-like noises it makes the car picks up to move.
It's not their own language, it's just data over audio.
I guess it is a 'language' if you consider dialup data sounds also a language. People want to jump on anthropomorphizing it.
All it did was:
- Acknowledge it can talk in auto bitrate
- Handshake like a modem to confirm data rate and stop bit
- Both Acknowledge in data they are in agreement
- Start data transfer.
If that's an advanced secret language, my 1996 2400 baud modem had a secret spooky language
And if that's the case, my old house phone had a secret language. When I pushed buttons the other machine figured out what specific person in the world I was looking for just by a series of rapid tones my phone generated. (10 digits converted to a series of tones each understood)
@PickupArtist
Did you even read the article you linked in the first one? The title Reddit gave it, and what he said are not the same.
Not saying that to defend AI. Im saying it to defend: Don't take Reddit titles as truth.
Stormwolf - "Who cares about some racial stuff, certainly not the victims."
- Democracy Dies in Dumbness.
- Watching people my age grow from cynical youth who distrusts and dismisses the older generation, into cynical old people who distrusts and dismisses younger generations.
Signature/Avatar nuking: none (can be changed in your profile)
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum