News AI bubble is worse than the dot-com crash that erased trillions, economist warns — overvaluations could lead to catastrophic consequences

[Eximo]

I can assure you that's not true. If you ask something that it doesn't know the answer to, rather than saying 'I don't know' they tend to just come up with a plausible sounding but incorrect answer. I asked if a particular engine used aluminum or steel heads (I meant to say cast iron) and it assured me it was steel (I don't know if any car used steel heads). Then I asked again if it was aluminum or cast iron and it said it was cast iron (it's not, it's aluminum.) That's not from internet trolls, that's from it not knowing the answer and just making one up.

Bespoke AI wouldn't be that unconstrained. They would feed it extremely specific training data so that it would have a narrow scope of capabilities which would apply to a specific role replacement. Basically advanced automation. You would not be asking it an unfamiliar question, but having it pick a solution/response to a question that has been answered countless times before.

As someone else said. Any 'Cog' that uses a computer.

 

[hwertz]

Bespoke AI wouldn't be that unconstrained. They would feed it extremely specific training data so that it would have a narrow scope of capabilities which would apply to a specific role replacement. Basically advanced automation. You would not be asking it an unfamiliar question, but having it pick a solution/response to a question that has been answered countless times before.

As someone else said. Any 'Cog' that uses a computer.

Indeed. I find the generalized AI to be overhyped and underdelivering; but bespoke AI is quite effective.

That'll be something to watch as the AI bubble continues to expand and eventually bursts -- I figure companies whose AI products amount to "general purpose LLM + some prompts" will have more trouble getting consistently good results than a bespoke AI (like Nvidia's weather AI, wihch gets excellent results) where it has a specific task and is trained precisely for that task.

 

[DS426]

Sorry, but anyone denying the 4th industrial revolution is an idiot. The numbers are there.

These idiots think that the INDUSTRY will leverage AI for profit, while what they don't understand is that it is the GOVERNMENTS that are going to leverage it for their own over complicated business processes.

Anything that is relative to cognitive analysis is targeted. Not to mention that AI is link to the evolution of robotics and cyber wares.

AI is not for selling ADS, it is for advancing science. If all the big tech are throwing money at it is because they know that the people controlling it will control the world.

Welcome to the world of Cyberpunk and the Corpos.

You keep making blanking statements and calling large swaths of people names. Silencing dissenting thinking is an attempt to control the conversation and quite anti-free-speech. Everyone has an opinion on THN, but not all opinions are appropriate to share out loud.

First: why wouldn't the "INDUSTRY" leverage it for profit? Most of the largest businesses in the real are for-profit businesses. Corporations in particularly have a fiduciary financial responsibility to their stockholders. Capitalism -- whether one favors it or not -- just result in both great wealth of nations while simultaneously being an engine for innovation and change; the former and the latter are not mutually exclusive by any means.

Second: we're talking about over-valuating companies and a potential dot-com-like-bubble, which absolutely is feasible. AI is transformational even if it also is over-hyped. Both have fair arguments behind them.

Third: AI is used heavily in marketing. That's not speculation -- it's fact. How accurate/truthful one finds in the marketing is on the individual, but marketing in general tends to be slightly misleading and/or over-the-top. Marketing teams are happy to have such a big bone thrown their way. AI has real value and uses, including to advance science, find anomalies in data telemetry to help find advanced cyber threats, find something important that a medical doctor missed, etc., but it can also be used for cyber attacks (craft convincing phishing emails, indirect prompt injection for phishing/malware, etc.), generating CSAM, misinformation campaigns, and all sorts of nefarious purposes by criminals and adversarial nations. Even if AI's conception had pure intentions, it can't be assigned some kind of moral value or virtue as it's just another technology that will be serve humanity and harm it. Hopefully it will be a net positive... call the Terminator references naive, but the smartest people all over the world have real concerns for really good reasons.

EDIT: Also, this isn't just "analysts" -- it's others as well, including AI CEO's themselves, like Baidu's:
https://www.tomshardware.com/tech-i...en-the-bubble-bursts#xenforo-comments-3858087