ChatGPT and AI in general. Yea or Nay, good or evil?

[Michael Weber]

Curious about what the memberships opinions are on AI. I’ve been playing with it a little. It’s both impressive because of the depth of knowledge, and scary because AI’s ultimate uses and effects are unknown. I can think of a few. I’ve tried it to write some Arduino scripts, Visual Basic and even LabView code. The latter two were descriptive answers not visual representations as currently it’s dialog based. It isn’t perfect. I was given a wildly inaccurate answer in a simple math (arithmetic really) discussion. But when I told it that the answer didn’t seem right it apologized for it’s confusion and provided a correct solution. It should have taken the inverse of a number rather than the number. Otherwise, answers about programming, process control, drug side effects, children’s literature and meanings of abstract items in a Tom Robbins novel seemed correct and reasonable. If you haven’t tried it the URL is chat.openai.com/chat.

[Alan Rutherford]

Tried it:

ChatGPT is at capacity right now
Get notified when we're back (https://share.hsforms.com/13gyIEVN5SrScw-iVvCgIew4sk30)Explain the status of ChatGPT as a sea otter.

Squeak squeak! Sorry, ChatGPT is very popular right now. Please try again later! Squeak squeak!

Whatever it is, it has a sense of humor.

[Michael Weber]

it does. i ended a convo by typing "Later Gator". It replied "After while crocodile"

[Edward Weber]

The tools themselves are neither good nor bad.
Like any tool, they can create great things or you they can do great damage, depending on who is using them and how they're used.

[Kev Williams]

AI is getting too close the below to suit me...


... [Skynet] goes on-line August 4th, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14 a.m. Eastern time, August 29th. In a panic, they try to pull the plug.

I'm not so worried about the computers taking outright control like in the movie. What I'M worried about are the humans running the computers...

[Edward Weber]

AI is getting too close the below to suit me...



I'm not so worried about the computers taking outright control like in the movie. What I'M worried about are the humans running the computers...

Garbage in, garbage out

[mike stenson]

Garbage in, garbage out

Have you happened to read about the AI bots exposed to the internet in the past. It's amazing how quickly they became the worst of what humanity has to offer.

[Edward Weber]

Have you happened to read about the AI bots exposed to the internet in the past. It's amazing how quickly they became the worst of what humanity has to offer.

Yep, it seems like they immediately go downhill fast.
Doesn't really reflect well on us flesh monkeys pushing the buttons.

When the robots do take over, this was the trigger IMO (about 1:20 in starts our downfall)
https://www.youtube.com/watch?v=rVlhMGQgDkY

[Alan Lightstone]

Not sure if this qualifies as AI or machine learning, but the AI photography retouching programs (Luminar Neo, ON1 PhotoRAW 2023, Topaz PhotoAI, DXO Photolab 6.3, and to some extent Adobe Photoshop and Lightroom Classic) as getting incredibly impressive in what they can already do, and are improving by leaps and bounds yearly. Both awesome to use and scary at the same time.

I'm getting progressively more concerned about Deepfakes, as they may become indistinguishable VERY soon. Right about Presidential election time.

Not surprised that ChatGPT can pass a bar exam. I mean, let's face it, lawyers can too. :D

[Edward Weber]

There have already been several AI generated images that have won photography contests, let the controversy begin.
Example
https://petapixel.com/2023/02/10/ai-image-fools-judges-and-wins-photography-contest/
https://www.cnn.com/2022/09/03/tech/ai-art-fair-winner-controversy/index.html

[Steve Demuth]

Open AI has proven with ChatGPT that you can build an AI capability that is truly fluent in human natural language, and kinda-sorta capable with programming languages, and then train it to more or less carry on a conversation and answer questions. But ChatGPT has no fact base, no means to look up or reference information not in its training set, no ontological capability, no reasoning models, logical or mathematical, other than what got dragged along in achieving natural language fluency. So it's a very fluent BS generator. But we know how to build those other things, and work like Meta's Diplomacy model shows progress in coupling language models with other modes of reasoning. I think ChatGPT is a huge leap forward, but mostly it's just a milepost on the journey to useful general purpose AI. I expect we'll see en ormous progress in curing it's odd liberties with truth and reason in upcoming versions.

As for AI in general, I'm very bullish. The kind of learning and reasoning models outside of language (Deep Mind's AlphaZero calculating how proteins fold, e.g.) show the way to accelerated understanding in many areas of science and engineering. Just before retirement I worked on applying deep learning to reading ECGs in a clinical setting. We were able to learn things from ECGs more quickly and more reliably than with electrophysiologists doing the reading, and most impressively, were able to use ECGs to diagnose and electrically "image" the heart in ways the cardiologists themselves couldn't/ For example, the AI models could detect reduced ventricular ejection fraction - a precursor to heart failure) in patients that had no symptoms detectable to their cardiologists, absent a full imaging workup.

[Lawrence Duckworth]

Curious about what the memberships opinions are on AI.


Here's a link you might like. a conversation with Elon Musk. https://www.vox.com/2016/6/2/11837608/elon-musk-simulation-argument
40 years ago we had PONG. ..two rectangles and a dot :D

BillGates likes CHatGPT says it's as important as the PC and Internet..irwilbe

[Lee DeRaud]

Yep, it seems like they immediately go downhill fast.
Doesn't really reflect well on us flesh monkeys pushing the buttons.

When the robots do take over, this was the trigger IMO (about 1:20 in starts our downfall)
https://www.youtube.com/watch?v=rVlhMGQgDkY
The scariest thing about that clip is contemplating how much they've improved in the 5-6 years since it was shot.

[Bill Dufour]

I suppose they will use it to make annoying phone calls while we will use it to answer them about car warranties they are trying to sell.
Bill D.