OT: Are we on the cusp of an “A.I. bubble?”

[PurpleBikeChick]

I work in a medical field where is much interest & potential to use wearables to identify/diagnose illness earlier with the hope/potential for better research/outcomes/treatments.

OTOH, the AI introduced for use in face to face encounters is sorely lacking based on simulated encounters I have fed it.
-If truly AI, this should improve IF clinicians do not rely exclusively on the suggested diagnosis, evaluations/treatments.

Applicants for residency/medical school are using AI for personal statements.

[bicycletricycle]

Why is this on this forum ?

[pinkshogun]

I'd rather know what AI stock or ETF verticaldoug recommends

[vertr]

Originally Posted by bicycletricycle

Why is this on this forum ?

Seems like the norm is anyone can talk about whatever topic they want as long as it's prefixed with "OT"...

[echelon_john]

I worked in AI for the last 6 years. Generative AI is what gets the buzz but deductive AI-which can analyse new data in the context of huge bodies of existing/past data in near real time-is absolutely legit and a game changer. The company I worked with was the first solution built on AI from the beginning to detect fraud in P&C insurance claims and has since branched to have application in health care insurance claims, underwriting, and claims processing.

Using AI to analyse claims for potential fraud results in a ton more fraud identified than with human review only. It just sees anomalies better and cross-references data quickly enough to be very useful.

The opinion on Wall Street a few years ago was that “data is the new oil.” I would go further and say that companies that make enabling technologies to apply AI to data are going to be the biggest performers in the next five years. Several have already enjoyed incredible success (and valuations) over the last several years. Is it a bubble? Not if you talk to c-level execs who are seeing real business problems solved/eased with AI or hear corporate boards putting pressure on executives to figure out how best to invest in AI technologies.

[XXtwindad]

Originally Posted by bicycletricycle

Why is this on this forum ?

Nothing in the rules prohibits it. Don’t like the topic…don’t respond. Not complicated.

[avalonracing]

Some people thought the internet was going to be a fad.
Hell, I thought rap was going to be a fad.

[sonicCows]

As with any gold rush, be the one supplying the picks and shovels, not the one mining the gold. It's the easiest way to ensure earnings.

Of the business problems solved, some will truly be solved, others merely solved on paper (but instead the work is pushed onto other groups).

I still had to use a fax machine the other day. Why is that?

[floxy1]

Kent Beck had a quote I thought was pretty spot on in the software realm:

“The value of 90% of my skills just dropped to $0. The leverage for the remaining 10% went up 1000x. I need to recalibrate.”

[unterhausen]

Originally Posted by Turkle

Its most useful applications appear to be plagiarism and the production of unbelievable amounts of useless trash content at scale.

I asked google's AI for a line drawing of something. It doesn't know what a line drawing is, and what it gave me was someone's copyrighted drawing of of the thing I wanted with bad modifications. Akin to adding extra fingers, but still recognizable as a derivative of the original. I had been searching for images before I asked, so I recognized it right away.

[Fat Cat]

Personally, I think that SkyNet or Colossus would solve many of world's problems.

[MikeD]

AI is in its infancy and is being overhyped by the media, just like Segways were, the Space Program, self-driving cars, Bitcoin, and currently EVs and the internet early on when people were throwing money at internet companies that had no business plan and were doomed to fail and caused the Dot Com crash and a recession. I would not hold onto Nvidia stock long term. I wonder if Warren Buffet invests in AI?

[veloduffer]

A former colleague of mine has this column (free to subscribe) on various risk, mostly financial (he was a former Fed official) but others on climate and technology. Here's a link to his most recent column on AI: Perspective on Risk: AI Stuff

In the column, he mentions an interview by Ezra Klein with Dario Amodei, who led OpenAI creation of GPT-2. He noted: So there are these two very different rhythms I’ve been thinking about with A.I. One is the curve of the technology itself, how fast it is changing and improving. And the other is the pace at which society is seeing and reacting to those changes.

They then list a number of steps where the technology advanced, but public awareness didn’t until the release of GPT-3; the underlying technology is a smooth exponential, but the awareness is a step-function.

I think unless you are deeply involved with AI development, it's hard to fathom the possibilities and the speed at which it is evolving.

[verticaldoug]

I just like watching how FSD reacts to unexpected events which are mundane for us.

As Yann LeCunn says, AI doesn't understand the physical world, so often fails. Even a young child has digested multiples of visual and audio data by a young age compared to the amount of data used to train a large AI system. So a child has an innate understanding of the physical laws of the world, AI does not.

https://www.dailymail.co.uk/news/art...d-Phoenix.html

Waymo behind a truck pulling a tree. Oh, there is a tree in the road, I should swerve so I don't hit it. Rinse repeat while following the tree.

https://www.dailymail.co.uk/sciencet...ing-train.html

Tesla approach a train crossing in the fog with some flat bed train cars passing. It does momentarily look like a clear track.

Pretty cool.

[Wolfman]

Originally Posted by PurpleBikeChick

I work in a medical field...

This. And as echelon john pointed out, it's the deductive AI that's interesting.

I just got back from a conference (deep dive into Echocardiograms) where I heard about a well known clinic that took 15k patient EKGs that had been followed by an Echo within 7 days that was read/coded by an expert-level Cardiologist and fed that into the "AI" that's tacked on to their medical records system... after that training, they're feeding _EKGs only_ into the system and getting likely Echo findings back with a remarkably high level of accuracy.

So, for me, the level of real signal that can be pulled from what used to be noise, especially in imaging, is crazy.