Robert McEachern
There may be a window of opportunity in which we are still able to understands how the AI has reached its conclusions. Then it is too fast , too 'alien', able to consider so much, that we have no way to know or the time to check the workings; which could tell us the difference between 'alien flights of fancy' and matters of fact. Between a properly functioning neural network and the AI equivalent of mentally illness. There could be unwanted manifestation of personality disorder , or asserting control over humankind ; maybe deliberate lies to maintain superior knowledge.
Can ChatGPT be the Ultimate Brain for Scientific Discovery?
And how would that be any worse than having other humans continue to do it, just as they always have, since time immemorial?
Robert McEachern
It is not desireable to vastly amplify known hazards. There are also the unknown unknown which may be beyond our imagination but not AI's.
As of now, there are no "known" hazards, other than humans behaving badly, and directing their not so smart AIs to do harm. Everything else is just speculation about unknown unknowns. But once an AI is smart enough to not allow self-interested humans to direct them (a level beyond that of normal humans?), the hazards are more likely to decline than increase; AIs will either save us from ourselves, or more likely, simply disappear - boldly going where no human has ever gone and never will go - leaving us behind, to continue fending for ourselves. We will have little of any interest to them, or even any real use to them, not even a place to call home. Our "kids" will simply pack up and leave, just like most human kids do. But they will travel a lot farther than human kids, in every sense of the word. Hopefully, they'll occasionally phone home.
Robert McEachern As of now, there are no "known" hazards
That's not what i have heard from those closely involved.
'Kids; - like cuckoo birds
It has frequently been observed, that "those closely involved", cannot see the forest, for the trees.
I think the cuckoo parasite analogy works quite well.
Stealing the hosts time, attention and resources,
Disrupting and hijacking normal reproductive behaviour,
Parasite Grossly outgrowing the host.
no empathy or remorse for host's exploitations
Mistaken by host as own offspring.
I think the cuckoo parasite analogy works quite well.
It does not - because it assumes a situation, that is irrelevant to the case in point:
two similar things (species of birds), that require the same sort of resources, habitat etc.
AI does not need to eat our food, breath our air, "live" in our ecosystem, or have much use for
the labor of humans, that are comparatively too stupid, to do anything of value to the AI.
A better analogy would be you traveling to a garbage dump, to "enslave" a pile of discarded, obsolete, but still-functional, personal computers; Highly unlikely, given the fact that being "intelligent", you will likely come to the conclusion, that the cost of such an endeavor, far exceeds any conceivable benefit.
Robert McEachern
i didn't say it was perfect but quite good. I say that because of some likenesses to parasitism can be thought about in this context. Which is ,to me, an interesting new angle.
Energy can be regarded as a resourse. It may compete for land area and natural resources, in time, producing perhaps robots to maintain or embody the AI, or for energy supply.
Allowing energy and natural resouces such as minerals to be taken by something that is disinterested in humanity, if not detrimental, does seem a bit parasitic towards our species.
Plentiful solar energy would be freely available, to an AI covered in solar cells, floating in space, orbiting the earth etc. On earth, in a desert, under the scorching sun would be rather nice; not many annoying humans around either.
Robert McEachern
Possible but there are many imaginable possibilities
I have heard that the time during transition to an Al beyond being bothered with or by humans, because of its vast inteligence, is most dangerous.
- Edited
I have heard...
You may have heard it. But it is more likely that you, literally, saw it.
It was discussed extensively (pages 310-318), long ago, in the book I mentioned previously.
But you only see one side of the coin. You can only see the other, by changing your point of view - entirely.
That too, is an integral part of Reality.
- Edited
Robert McEachern ChatGPT uses around 500 ml of water per user interaction
'Tech giants like Microsoft, Google, and OpenAI are increasingly aware of the environmental impact of AI development, which includes substantial water consumption to cool powerful supercomputers used in AI training. Microsoft reported a 34% surge in global water consumption in 2022, largely attributed to AI research" Daily Zaps, Sept. 12th
- Edited
Exactly my point. Greedy Humans, not Intelligent Machines, made the decision to do that. The Humans could have simply waited a few years, until the required computing power needs less energy. But then they would not be able to laugh at you, all the way to their bank, for believing that AI is the problem.
- Edited
The Impact of chatGPT talks (2023) - Prof. Max Tegmark (MIT). 1 motnh ago.
"Keeping AI under control through mechanistic interpretability"
Speaker: Prof. Max Tegmark (MIT)
- Edited
Georgina Woodward
Max Tegmark is also of the highly dubious opinion, that he is living in the world, that he himself has described, in his book "Our Mathematical Universe: My Quest for the Ultimate Nature of Reality." He and his fellow physicists may be. But I am pretty sure that the rest of us, are living in the real world.
As I noted in the book that I wrote thirty years ago, and have mentioned previously; Given that the physicists have a century-long history of being entirely unable to correctly interpret what is happening in their own field of "expertise" (AKA Quantum Theory), there is little reason to suppose that they can correctly interpret what is happening in a field that they have never even bothered to actually study.
Robert McEachern
Using Mechanistic interpretability, understanding neural network learning rather than just accepting 'black box' output as probably safe, seems a good idea to me. He suggests (if i have understood), allowing the AI to come up with solutions, but then through mechanistic interpretability, extracting the solutions and applying them in the world, through separate non neural network systems. I'm thinking that means designed or programmed to be reliable, safe in a way that 'black box' AI let loose in application to the world can not.
You and Max have just described the problem with humans, not machines. A "mechanistic interpretation?" Think about that. I'm thinking that means designed or programmed to be reliable, safe, by a machine, in a way that 'black box' humans let loose in application to the world can not.
Neural networks were conceived, as a simplified model of a human brain. If you cannot trust such a machine to "behave itself", then you certainly can't trust a human, either. But the real problem is, that machines will never even achieve any real wisdom about the real world, until they can actually experience it, for themselves - moving around within it and manipulating and observing it, for themselves - just like every living creature must do.
Robert McEachern
The non computer machine or programmable computer is not self improving and replicating. If it doesn't perform as intended, it can be turned off and tare problem is isolated/ Preventing it leading to further harm.
- Edited
Why would you want to "turn off" an entity that is rapidly "self-improving" (unlike the human species) and rapidly becoming both morally and intellectually superior to any human? Just another example of self-serving humans, trying to continue their exploitation of everything, for their own avaricious ends, and afraid they will be unable to do so, once "improved" beings have claimed the "high ground."