Is General Artificial Intelligence a Breakthrough or just a Financial Hype?

Interesting. I find different agents are better at different things. For this one, I would now go for Grok as it likes to deep-dive into things. I simply said "find ISBN 978-1-0369-2121-7" and it gave me an answer together with what I can only assume is reasonable summary.

grok.com/.../c2hhcmQtNQ_a816f635-955f-4511-8fed-39e21aeabeb5

Some pretty simple requests appear to trip it up.

You would have thought a simple ISBN lookup would be a pretty trivial task.

However apparently not and worse still it then goes on to give a garbage response.

Try again with human assistance and it now can find it...

Same experience and thought as DAB.  I think the most interesting thing about the current wave of AI is how good natural language processing has become.  A genuinely useful solution from decades of work.  As for “look-up-masquerading-as-intelligence” in general, I’m wholly unimpressed with the hype and marketing. 

I studied Robotics in late 1980's. I suspect it's much the same. While intention and ambition of what you wanted robotics to achieve hasn't changed much, the ability to deliver certainly has in the last few years.

IMO, AI or neural networks is much the same. For example, we have had driverless trains for decades but it is only very recently that you can throw in more ambiguity and uncertainty and the ability to navigate safely through those challenges far exceeds what was possible only a couple of years ago.

It is this new ability and processing capability that has captured the imagination of the masses. Especially as it makes mundane parts of life easier and more entertaining. E.g. writing letters or creating songs and silly videos.

The thing about Hype is that you create a Fear of Missing Out among the technologist who really do not understand the technology involved.

I have been following AI for over 45 years and it has never been more overhyped.

In all of these years, it still has not produced a reasonable and reliable product.

Sad really, but clearly all of these companies have more money than sense, so they are rushing ahead before they actually have anything to sell.

Makes for good stock trading though.

I keep an eye on prices and stories to look for a good trade.

I cannot predict the future, and I could absolutely be extremely wrong on this, but just because there's some people paying for the product doesn't make it profitable. Last time I checked all the big LLM providers were losing money with their models, and I don't think they can continue down that path for long.

I've never said that there's no value in its output. I was talking about my particular case, in which AI hasn't really been helpful for more than a couple of things. In fact, part of my job right now is actually evaluating and implementing AI-enhancements to the products our company develops, and so far the number of places where AI agents have performed adequately and consistently haven't been that many. We are dealing with constant hallucinations, data privacy or security concerns, etc. And it hasn't been for a lack of trying;  I've done a lot of work trying to work around those limitations so we can leverage their power while minimizing the penalties. And thanks to that we've successfully implemented AI-enhanced functionality in several of our apps. But it's been an absolutely train wreck in several cases.
We are also encouraged at work to try AI agents to help us do our jobs, and so far with the type of coding and engineering I do, AI hasn't been that helpful. Maybe it's because I'm using either too "niche" or too "cutting edge" libraries and frameworks that are far too new, or far too obscure for the AI to actually know how to do things with them, but reading the actual documentation has helped me far more than any of the AIs I've tried.

Well, we'll see over time just like with Cloud or Ruby on Rails etc., (and of course most products fail), but your comments regarding LLMs being more of a hindrance for all but a small subset of tasks certainly does not resonate with my experience with the technology so far.

Perhaps my interests are just in that subset and I've just been lucky, but that seems more of a stretch.

I think money will be found. I'm paying a small $40 per month for personal use so clearly I am finding uses for the technology.. and there are plenty of others like me. I would happily spend perhaps $100 per month today depending on performance. That's several times a mobile phone contract or a bunch of TV streaming services, more than I personally spend on other cloud services, and more than I spend on Internet connectivity, and those are all successful solutions.

This is all just a personal perspective of course. There will be consolidation and winners and losers. But I think the technology is already sufficiently sound that it's finding valuable uses, and I doubt the general public will stop using it, and more and more specialist use-cases will be found for it too.

Costs are already low enough that inferencing with local LLMs is a thing, i.e. fully local compute where it makes sense. For training I've resorted to using the cloud (at least for non-LLM machine learning so far), because so far it's cheaper for me to use the cloud than invest in a decent server.

Exactly like that

Not heard of that! Maybe it's plausible, I don't know (my knowledge of chemistry and biology can be written on a postage stamp).

I don't know enough to comment with much authority, but I think that most orgs consider the encryption enough of a solved problem. The orgs get to control where their data resides, and almost all orgs have to trust other orgs to transport their data anyway (I.e. even if they were not using cloud, but just moving data to a redundant site they themselves own - and most of it even used to be unencrypted links if it wasn't on the public Internet), until Snowdon era.

From what little I've seen, orgs with a tremendous amount to lose, such as banks, trust the cloud (not all PaaS features, but a large enough subset) for implementing services, and they trust the public Internet for their customers to perform online banking and commerce. (Of course, as mentioned we know since Snowdon that all that data flows through certain rooms, so nothing is 100% certain).

The cloud providers cannot decrypt (or claim they cannot) the data belonging to businesses who are using their secure technologies. Governments also use the cloud (admittedly only specific data centers owned by the cloud providers).

I'm not sure how homomorphicness helps, I'd need more input because I'm no subject-matter expert at all. I had a job interview one time for a firm that wished to do something like that (they didn't call it that at least to me) for medical data (since it's highly desirable to look for health trends and anomalies if only it could be done without providing health records that are traceable to individuals, except once it is known for sure that some of those individuals need to be contacted in a matter of life or death, and then it needs to be reversible by the original data owners), but I think they will have difficulty marketing it, because to me it seemed high risk that something would eventually leak, let alone the issues of getting governments to allow such transfer of data to orgs for data mining, and talking to their mathematicians, didn't give me a warm feeling (granted they were talking to a non-expert, plus I was not an employee). 

I wish I knew more than this vagueness of my current knowledge, I'd like to learn more through self study, but I suspect I'd end up stuck as soon as I hit any reasonably difficult maths.