Artificial intelligence discussion

Something I've noticed during the last few months is that many researchers have acknowledged that AGI may be still out of reach.

While powerful, the Chatbots / LLMs which are based of the transformer architecture, seem unlikely to achieve AGI (Artificial General Intelligence), at least by simply scaling the models (was the past strategy.)

For two main reasons, at least that's what distilled in my mind:

1. Lack of Generalisation

2. Lack of Continual Learning

They still will have most of the issues and gains you write about, but this is an aspect which is discussed in many areas as well. Said in their language, the singularity (when AIs reach AGI, and maybe beyond, to super-intelligence) isn't going to be reached with the current architectures.

That is my read from the ones I follow most commonly: Ilya Sutskever, Andrew Karpathy, Yann LeCun etc.

Others have somewhat different view.

Just thought this tangent could be of interest for some members.

Sorry Lotus been thinking about the resources aspect, it is an issue and I guess a growing one, I suppose AI might be able to find its own solution? It’s a very interesting area.

• I think it’s a work in progress and like any tool whether it’s good or not depends on how it’s used. I take photos of my artwork and ask it to appraise it. As it’s not personal I find this helpful it gives advice on composition, style and technique.

I use it at work and at home. 

At home I use it for two things: firstly I find the info I receive back is more tailored response than say just a google search , secondly I use it as a sounding board for checking my thoughts or situations which is like a online therapist. 

It's almost you're having a 2 way conversation - and definitely check the response - sometimes it definitely gets things crossed up and for work I'd never do a copy and paste from it, other than refining say a sentence or two, but even then not without checking.

It's just another tool in the toolbox, as far as I'm concerned - but something whose development and capabilities will need to be monitored + controlled

My thoughts as a tech enthusiast, but having only basic AI usage:

At the moment, there is a lot of hype about AI, some of it is justified, and others not.

A good use case I read about was as a "peer" for reviewing medical scans, where 2 doctors are already required to review, and there is already too much work to do. This seems like a good idea, and could save a lot of lives. First use it as an additional peer. As you gain confidence in it, replace one of the peers to save time. Potentially going further, but only if the technology is good enough. From what I've read, they can sometimes detect things that the doctors can't. Another option might be as a quick regular screening. If it takes a doctor even a few minutes to check a scan, but an AI only a few seconds, you start getting into the realms of just sending it unrelated scans to see if it can detect somethimg before there are even any symptoms.

It can also be good for fun. I recently had it generate a cartoon image of me and my dog. It's a very good image and was quick. This is however potentially a big environmental impact just for fun, and an economic impact, taking work away from artists who might be struggling to get work. The thing really don't like is it being pushed where it feels like it's there just to say "We have AI". For example, why is it in whatsapp? So I can chat to a bot? I'm fairly sure that won't be allowed in the T&Cs, but it is for Meta.

There is a much more difficult question about it's use in the military, which I think it's best not to go into.

At the moment, any use of the technology I think needs either review of answers, or a lot of trails if that isn't possible (e.g. Self driving cars can't have their "answers" reviewed in the moment).

The other side of this I think is if we are ready for AI as a society, related to the economic impact mentioned earlier. Ideally, AI would help to do less desirable work so that we have more time to do things we enjoy. Essentially allowing everyone to reduce their hours, but maintain the same or an improved lifestyle. What I believe would happen instead, based on reactions to 4-day week and responses to requests to go part time I've heard of, is an increase in unemployment instead.

Neither would I

Id argue it's going to change all jobs. I'm 20 years into my career now and very much invested in it so changing now is a difficult task, but I'm doing my best to find a way through. Touch to pay bills and retrain though

Deleted User said:

I would not count on Elon supporting a humane transition

It isn't often you hear Elon and humane in the same sentence LOL

JT said:

The only part I really care about is the transition from where we are now to AI doing all the jobs.

Find a field where AI cannot do the job.

I've moved into property development as AI can't plumb, install electrics, plaster, build a partition wall or eat all the owners biscuits at tea break.

I don't need to interact with people much either which is a bonus, but if you don't like unpredictability and loud noises / mess then it probably isn't for you.

My last job was preparing a multi billion pound companies IT support service for replacement by AI so I most certainly saw the writing on the wall.

I would not count on Elon supporting a humane transition 

Yes, it may be time to stop stigmatisong and punishing the unemployed. 

The only part I really care about is the transition from where we are now to AI doing all the jobs. Why aren't governments putting plans into place for the impending purging of staff as AI Hoovers up the work??

I am very worried about this phase

I also did a module in creating AI systems as part of a computing course.

Although AI can be useful.  The big problem is that it lacks compassion,  intuition and common sense.

The classic problem was how an AI system and a human doctor would identify a dead body.  A human doctor would come to the conclusion almost instantly.   An AI system would ask question after question trying various hypothesis to figure out a possible illness.

Deleted User said:

Now, how do you feel about people like Elon Musk, Mark Zuckerberg or Jeff Bezos having thousands or millions of brains vastly smarter than Einstein under their control? Imagine what they’d do with that.

Amodei thinks we have less than 18 months.

What all this means is that o3 (which already exists but hasn’t been released) will likely be capable of generating prize-winning novels, musical compositions and scientific discoveries.

Dario Amodei (CEO of Anthropic) used this analogy - imagine a data centre full of instances of a system like this. It would be like having an entire country of geniuses at your finger tips.

Now, how do you feel about people like Elon Musk, Mark Zuckerberg or Jeff Bezos having thousands or millions of brains vastly smarter than Einstein under their control? Imagine what they’d do with that.

Amodei thinks we have less than 18 months.

I totally agree with all of these points 

I think it has its place, it's already used to undertake certain kinds of precision surgery. 

The question will always be making sure we use it rather than being a slave to it. In many countries, Smartphones are being banned from classrooms because of the potential for contrant distraction they bring for kids. Tablets with no access to the internet might be a way they could be used better in the classroom, for example. Or with an intranet exclusive to the school.

Robots in the house? Well. I  could already go out and get a robot vacuum leaner today if I wanted. Robots would have to be better coordinated to do much more though. 

I do get mad when I hear about humanoid robots selling their artwork or tens of thousands of dollars, when I can only sell mine, refined after decades of practise, for peanuts because I don't have enough of a Name, and am mainly self taught. But the same goes for animal artists. I've half a mind to rebrand myself as a pig, or an elephant. I could make me a nice dangling trunk and reidentify. I could rebrand as a robot dog even! The possibilities are boundless! I could tape a rotten banana skin to the wall on my next Exhibition! And no, I wouldn't be interested in coming out as neurotypical, with all the condescension that would bring. 

Casting pearls before swine, and yes I know that artistic self expression is often a spiritual thing. 

The big question of SF is again, will the robotorcsentient computer become a Frankenstein monster, or will they become infused with the divine spark? Some scientists fear that in the case of A1 we are already there. 

Interesting times then. 

I’m not sure the word “intelligence” applies to the current system.

Deleted User said:

I think the forum admins could do with some help from a decent AI.

Or can't they just contact the provider about the one in use?

Too simple perhaps

I think the forum admins could do with some help from a decent AI.