Re: [FRIAM] AI Musings

2023-04-03 Thread Roger Critchlow
Wouldn't be the first time I misunderstood what Stu was talking about. It was in the lecture he gave at um6p, ah, here, https://youtu.be/GTQJqpkVd84?t=1820, I guess it's in this paper: https://arxiv.org/pdf/2106.15271.pdf -- rec -- On Mon, Apr 3, 2023 at 4:09 PM Frank Wimberly wrote: > The

Re: [FRIAM] AI Musings

2023-04-03 Thread Frank Wimberly
The real numbers are an uncountable set. --- Frank C. Wimberly 140 Calle Ojo Feliz, Santa Fe, NM 87505 505 670-9918 Santa Fe, NM On Mon, Apr 3, 2023, 3:14 PM Roger Critchlow wrote: > I'm just noodling around. > > So it's a collection of innumerable elements which can't be a mathematical >

Re: [FRIAM] AI Musings

2023-04-03 Thread Roger Critchlow
I'm just noodling around. So it's a collection of innumerable elements which can't be a mathematical set, so you can't use the mathematics founded on sets to analyze it, so what's the paper doing using set based mathematics to count the number of elements in M_t? -- rec -- On Sat, Apr 1, 2023

Re: [FRIAM] AI Musings

2023-04-03 Thread cody dooderson
Google's response to ChatGPT has a waitlist, https://bard.google.com/ . In other news I have been asking chatGPT which jobs are going to be the most secure from AI replacementAI. It seems like plumbing might be the most safe. Health Clinicians look like they are going to be in trouble. _ Cody

Re: [FRIAM] AI Musings

2023-04-01 Thread Stephen Guerin
On Sat, Apr 1, 2023 at 1:52 PM cody dooderson wrote: > The chat GPT demo does not have up to date information, which would > seriously hamper it's abilities to be a good CEO. I asked it about recent > tornados in the Midwest and it replied that it doesn't have any training > data after September

Re: [FRIAM] AI Musings

2023-04-01 Thread Steve Smith
Stephen, Very smart to put my question re: CEOs vs AI to ChatGPT. My question regarded “medium and long-term” capabilities. Your query to the bot was for the present time. And yet, nevertheless, I interpret the bot’s response to you to be, essentially, that there are certain CEO positions

Re: [FRIAM] AI Musings

2023-04-01 Thread cody dooderson
The chat GPT demo does not have up to date information, which would seriously hamper it's abilities to be a good CEO. I asked it about recent tornados in the Midwest and it replied that it doesn't have any training data after September 2021. Surely a model can be made too stay up to date with

Re: [FRIAM] AI Musings

2023-04-01 Thread Grant Holland
Stephen, Very smart to put my question re: CEOs vs AI to ChatGPT. My question regarded “medium and long-term” capabilities. Your query to the bot was for the present time. And yet, nevertheless, I interpret the bot’s response to you to be, essentially, that there are certain CEO positions that AI

Re: [FRIAM] AI Musings

2023-04-01 Thread Steve Smith
It sounds like "the Adjacent Possible"  just ain't possible* in AI land?  With the punchline "you just can't get there from here"? So what is the graph geodesic distance across the adjacent possible between "today" and Kurzweil's /singularity/?  And is the topology of "the Singularity" a

Re: [FRIAM] AI Musings

2023-04-01 Thread Stephen Guerin
And follow up prompt GPT: Stephen: good. Imagine business conditions or models where an AI could serve as a ceo chatGPT: While it might be challenging for an AI to completely replace a human CEO, there are certain business conditions or models where an AI could potentially serve as an effective

Re: [FRIAM] AI Musings

2023-04-01 Thread Stephen Guerin
I've understood the role of a CEO to be generally: 1) set the vision of the company 2) recruit talent 3) foster a culture to execute the vision 4) be the ultimate representive of the company externally 5) make critical decisions affecting the health of the company Btw, chatGPTs understanding is

Re: [FRIAM] AI Musings

2023-04-01 Thread Grant Holland
Good point, Roger. And any “old buddy” system among the board and musical chairs they play with the CEO position would tend to make them hunker down to protect their power structure. It might take the stockholders, once they finally gain confidence in AI, and see how much money can be saved, to

Re: [FRIAM] AI Musings

2023-04-01 Thread Roger Critchlow
I think that it depends on having a board of directors/private owner prepared to take their hands off the wheel. The main problem would be trolls attempting adversarial prompts. However comfortable you might get with the ai's ability to handle the day to day affairs, would you ever feel safe

Re: [FRIAM] AI Musings

2023-04-01 Thread Grant Holland
Good point, Cody! > On Mar 31, 2023, at 9:16 PM, cody dooderson wrote: > > While I think that AI could soon handle the managerial part of a CEO's job, > they may have trouble playing golf. It might not matter if the stock is going > up. > I am very ignorant about what CEO's do 'though. > >

Re: [FRIAM] AI Musings

2023-03-31 Thread cody dooderson
While I think that AI could soon handle the managerial part of a CEO's job, they may have trouble playing golf. It might not matter if the stock is going up. I am very ignorant about what CEO's do 'though. On Fri, Mar 31, 2023, 5:33 PM Grant Holland wrote: > So what do you think? Are CEOs, CFOs

Re: [FRIAM] AI Musings

2023-03-31 Thread Grant Holland
So what do you think? Are CEOs, CFOs etc. and corporate board members at any medium or short-term risk of losing their jobs to machine learning? I like to hear some opinions on this. Thx, Grant > On Mar 31, 2023, at 1:21 PM, Gary Schiltz wrote: > > A... looking more closely, Grant wrote

Re: [FRIAM] AI Musings

2023-03-31 Thread Gary Schiltz
A... looking more closely, Grant wrote CxO not QxO. Google quickly enlightened me on the former. Sorry for the noise. On Fri, Mar 31, 2023 at 2:19 PM Gary Schiltz wrote: > > I must admit my ignorance here, not aided in the least by a cursory > Google search: What is QxO? > > On Thu, Mar 30,

Re: [FRIAM] AI Musings

2023-03-31 Thread Gary Schiltz
I must admit my ignorance here, not aided in the least by a cursory Google search: What is QxO? On Thu, Mar 30, 2023 at 10:59 AM Grant Holland wrote: > > Frank, > > I'm wondering why no-one seems to raise the specter that AI could start > replacing management personnel. And I’m including CxO’s

Re: [FRIAM] AI Musings

2023-03-30 Thread Frank Wimberly
I agree, Russ. I taught at Pitt and Carnegie Mellon for 25 years or so. Frank --- Frank C. Wimberly 140 Calle Ojo Feliz, Santa Fe, NM 87505 505 670-9918 Santa Fe, NM On Thu, Mar 30, 2023, 2:42 PM Russ Abbott wrote: > Look at all the online course companies such as Coursera. They haven't >

Re: [FRIAM] AI Musings

2023-03-30 Thread Russ Abbott
Look at all the online course companies such as Coursera. They haven't replaced college teachers. I suspect that one reason is that online courses have no person-to-person mechanism to help students live up to the discipline necessary to do well in a course. As a retired and somewhat cynical

Re: [FRIAM] AI Musings

2023-03-30 Thread Grant Holland
Frank, I'm wondering why no-one seems to raise the specter that AI could start replacing management personnel. And I’m including CxO’s here; because I’m not convinced that CxO-ing is rocket science or quantum mechanics. Think of the billions saved. After all, if machine learning cannot get

Re: [FRIAM] AI Musings

2023-03-30 Thread David Eric Smith
There’s a nice subtext to why this is true. It isn’t only the mechanics of the task. People care what bulldozer operators do. > On Mar 30, 2023, at 10:58 AM, Frank Wimberly wrote: > > Not particularly relevant to your main point but Raj Reddy, close colleague > of Newell and Simon, once

Re: [FRIAM] AI Musings

2023-03-30 Thread Frank Wimberly
Not particularly relevant to your main point but Raj Reddy, close colleague of Newell and Simon, once said, "It is easier use AI to replace a college professor than a bulldozer operator" or words tho that effect. Frank --- Frank C. Wimberly 140 Calle Ojo Feliz, Santa Fe, NM 87505 505 670-9918