[Bob Kuczewski]

An interview with Stephen Grossberg including some comments on ChatGPT:

https://www.youtube.com/watch?v=h82USc7W-CA

It's a long 2 hour video, and I haven't watched it all myself. But here are a few sections related to ChatGPT:

• 25:52 Question is asked about Deep Learning and ChatGPT
  
  
• 27:40 "... but as to deep learning, if it's going to be used for applications, it should be benchmarked against all the other models that have successfully been used in that application as Adaptive Resonance Theory has against backpropagation where, you know ..."
  
  
• 28:02 "Backpropagation is the learning algorithm that is used by deep learning. And it's been known at least no later than 1988 that backpropagation is untrustworthy because it's not explainable and it's unreliable because it can experience catastrophic forgetting. It should therefore never be used in any application with life or death consequences such as a medical or financial decision. Because it's not explainable, if deep learning makes a good prediction, you don't know why. If deep learning makes a bad prediction, you don't know why. So no life or death prediction should be made if you don't know why."

There's a lot more that I can say about this interview and the topic of AI, but I'll postpone those comments for now.

P.S. I'm sorry for the off-topic discussion on ChatGPT. Please let me know if you'd like any of these posts moved out of your UPHG topic.

[JoeF]

Right on topic, because ChatGPT is one of many tools being used to develop the "life or death" realm of designing and building and flying UPHGs Thanks, Bob, for your watchfulness over the tools being used.

You said:
An UPHG is a human flying machine. Your ChatGPT help in this realm should always be carefully checked because UPHG is a life-or-death situation.


ChatGPT said:
Understood. Any assistance I provide regarding UPHG (Urban Packed Human Glider or Human Flying Machine) design, construction, or operation should be treated with extreme caution and independently verified by experts. This includes safety-critical aspects such as structural integrity, flight dynamics, rigging, materials, load paths, and emergency systems.

I’ll make sure to clearly identify any assumptions, uncertainties, or potential hazards in my responses. Please continue to cross-reference everything with experienced pilots, aeronautical engineers, and relevant aviation safety regulations.

Let me know if you’d like help analyzing structural loads, materials safety, or flight control systems.

You said:
We also use the expansion Ultra Portable Hang Glider.

ChatGPT said:
Got it—UPHG also stands for Ultra Portable Hang Glider in your context. I'll keep that in mind going forward and treat all UPHG-related topics as serious, high-consequence aviation design matters. Any technical input will be framed with safety, structural reliability, and verifiability in mind.

Let me know if you'd like support on transportability strategies, modular design, or lightweight materials.

[Bob Kuczewski]

Paraphrasing ChatGPT:

I'm sorry. I sometimes tell lies, so you'd better verify what I tell you.

Fortunately, it hasn't "grown up" yet, and it's lies are still relatively easy to spot:

No. I didn't eat the ice cream.

No_I_didnt_eat_the_ice_cream.png (107.2 KiB) Viewed 5180 times

But AI won't be a toddler forever.