Latest / Brian Thomas / Empower U - October 16th - How to use ChatGPT

Transcript

Empower U - October 16th - How to use ChatGPT · Brian Thomas

• 0:00
  Speaker 1: Out of most of their streaming success.
• 0:02
  Speaker 2: Shows Clay and Buck Today at noon on fifty five.
• 0:06
  Speaker 1: Karc Ato five Here a fifty five KRCD talk station,
• 0:12
  Speaker 1: Happy Wednesday. Sadly no judgment of Polatana. But at the
• 0:15
  Speaker 1: bottom of the are we're gonna hear from Scott Warpin from
• 0:16
  Speaker 1: the Cincinni enquired by the Council Race in the meantime.
• 0:19
  Speaker 1: Empower Youamerica dot Orger is where you find all the
• 0:21
  Speaker 1: empower Use seminar series. They're all wonderful, all very informative,
• 0:26
  Speaker 1: most notably chat GPT. It's a brand new phenomenon for
• 0:29
  Speaker 1: most of us out in the world. It's that artificial
• 0:31
  Speaker 1: intelligence doing the work for us. Michael Mercer is doing
• 0:35
  Speaker 1: the seminars taking place tomorrow night. You can log in
• 0:37
  Speaker 1: from home or I think you can show up yes
• 0:40
  Speaker 1: at the empower U Seminar Classroom three hundred Great Oaks Drive,
• 0:44
  Speaker 1: or you'll hear Michael talk about chat GPT by way
• 0:47
  Speaker 1: of background. He's very informed in this obviously this topic.
• 0:51
  Speaker 1: He is the president of Screen Education, which addresses issues
• 0:54
  Speaker 1: at the intersection of digital technology and human well being.
• 0:57
  Speaker 1: We'll get to that, including smartphone addiction. News meet you, bias,
• 1:00
  Speaker 1: artificial intelligence, thought research seminars, and consultant. He, like most
• 1:04
  Speaker 1: of us my age, started out with world book encyclopedias
• 1:07
  Speaker 1: that he spent a decade as an editor and publisher
• 1:09
  Speaker 1: of college textbooks and another decade in market research before
• 1:12
  Speaker 1: heading on over to president of Screen Education, got multiple
• 1:16
  Speaker 1: degrees and again the seminars tomorrow night, beginning at seven pm.
• 1:19
  Speaker 1: Welcome to the program, Michael. It's a real pleasure to
• 1:21
  Speaker 1: have you on today.
• 1:23
  Speaker 2: Well, thank you, Brian, I appreciate it.
• 1:24
  Speaker 1: And an interesting element of your seminar. Now part of
• 1:27
  Speaker 1: me wants to say, well, if you're using chat GPT,
• 1:31
  Speaker 1: you're not doing the work and you're not learning anything.
• 1:34
  Speaker 1: It's like you sit there in front of it with
• 1:36
  Speaker 1: your mouth open, a little bit of drool coming out
• 1:38
  Speaker 1: of the corner of your mouth. You let it do
• 1:39
  Speaker 1: all the work for you, and then you just regurgitate
• 1:42
  Speaker 1: it back to whatever source you're seeking the information for.
• 1:45
  Speaker 1: But you have a different viewpoint on this. You can
• 1:47
  Speaker 1: use it to learn, right.
• 1:50
  Speaker 2: So that's the point of my talk is really how
• 1:53
  Speaker 2: to use chat GPT to teach yourself anything. That's really
• 1:57
  Speaker 2: the point where I'm coming from with this is I've,
• 2:01
  Speaker 2: as you sort of alluded to this, I've been a
• 2:03
  Speaker 2: sort of an autodidactor. I teach myself things all my life,
• 2:08
  Speaker 2: and I during the COVID I actually spent about thirty
• 2:13
  Speaker 2: five hours studying MR and A vaccines and how they work.
• 2:16
  Speaker 2: I really needed to understand them for personal reasons, and
• 2:20
  Speaker 2: so that was a huge self directed learning project. And
• 2:25
  Speaker 2: I was sort of shocked because I would talk to
• 2:27
  Speaker 2: other people, friends and acquaintances about these vaccines and they
• 2:30
  Speaker 2: knew nothing about it. And I could see that they
• 2:32
  Speaker 2: had no interest in learning about it. You know, they
• 2:34
  Speaker 2: almost felt like they couldn't understand it if they tried.
• 2:36
  Speaker 2: So I realized, like, wow, you know, maybe a lot
• 2:39
  Speaker 2: of people don't teach themselves things. And then when I
• 2:41
  Speaker 2: discovered chatbots, I'll tell you, I asked the man I
• 2:45
  Speaker 2: spent about thirty five hours teaching myself about MR and
• 2:48
  Speaker 2: A vaccines that if I had a chatbot, I think
• 2:50
  Speaker 2: I could have cut it to fifteen. You know. Yeah, well,
• 2:53
  Speaker 2: you know what.
• 2:53
  Speaker 1: I have to interject this because we went back and
• 2:56
  Speaker 1: forth with my son. My son is now thirty one.
• 2:58
  Speaker 1: He was in the computer engineering department at Ohio State university,
• 3:02
  Speaker 1: and I think that speaks volumes to his his intellect
• 3:05
  Speaker 1: because it's tough to get in there. So he spends
• 3:07
  Speaker 1: about a year and a half and then he drops out.
• 3:10
  Speaker 1: And what are you doing? You're not paying for it,
• 3:12
  Speaker 1: Just get the damn degree. I don't want to sit
• 3:14
  Speaker 1: in front of a computer doing coding. And besides, Mom
• 3:17
  Speaker 1: and Dad, I can teach it myself. And we're like, well,
• 3:21
  Speaker 1: well he ended up doing that. He'd got certificates for
• 3:24
  Speaker 1: computer security. He knows how to code. He did it
• 3:26
  Speaker 1: all himself. He taught himself, so it can be done.
• 3:29
  Speaker 1: And I make this point regularly because he convinced us
• 3:33
  Speaker 1: that he was right. Why spend all the extra money
• 3:37
  Speaker 1: and resources when all of this material is readily available.
• 3:39
  Speaker 1: Why would you spend one hundred thousand dollars to get
• 3:42
  Speaker 1: an art degree when there's countless books on art? You
• 3:45
  Speaker 1: can self direct and learn all you want to know
• 3:48
  Speaker 1: about your favorite hobby without having some teacher tell you
• 3:50
  Speaker 1: how to do it. I think people are paying college
• 3:52
  Speaker 1: tuition for the structure, the forced obligation to learn right.
• 3:58
  Speaker 2: Right right, that's you know what I'm going to do
• 4:01
  Speaker 2: in the talk is it's broken into three parts. So
• 4:03
  Speaker 2: the first part, and I feel this is critical for
• 4:06
  Speaker 2: people who want to teach themselves things right the way
• 4:08
  Speaker 2: that your son did. The first part, I'm going to
• 4:11
  Speaker 2: walk through the learning process. So I've broken the learning
• 4:14
  Speaker 2: process into six steps. You set a goal, you gather information,
• 4:18
  Speaker 2: you vet the information, then you have to do the
• 4:20
  Speaker 2: hard work to understand at a deep level that information.
• 4:25
  Speaker 2: Then you structure it, and then you integrate it into
• 4:28
  Speaker 2: your current knowledge base. Right. So I'm going to walk
• 4:30
  Speaker 2: through this self directed learning process so people understand the
• 4:35
  Speaker 2: process and how it works at each stage. Then I'm
• 4:38
  Speaker 2: going to give an overview of chatbots, and specifically I'm
• 4:41
  Speaker 2: going to use chat GPT because it's so popular, right,
• 4:43
  Speaker 2: but i want to show people, you know, you know,
• 4:47
  Speaker 2: this is what a chatbot is, this is the interface.
• 4:50
  Speaker 2: These are the features that are critical and really helpful
• 4:53
  Speaker 2: in self directed learning. But I think another important thing
• 4:57
  Speaker 2: is to teach them how these you know, so called
• 5:00
  Speaker 2: large language models work, because I think if you can
• 5:03
  Speaker 2: understand how large language models work, you can be more
• 5:06
  Speaker 2: effective in using chatbots to learn, you know. And then
• 5:10
  Speaker 2: the third part, I'm going to synthesize the two. Then
• 5:12
  Speaker 2: I'm going to close off by walking through the each
• 5:16
  Speaker 2: steps of each of the six steps of learning again
• 5:18
  Speaker 2: and tying off, how do you know specific ways you
• 5:21
  Speaker 2: use the chatbot to optimize your learning at each stage. Okay,
• 5:25
  Speaker 2: So I think it's really going to be helpful for
• 5:27
  Speaker 2: people to sort of codify this process for people if
• 5:30
  Speaker 2: they do want to jump in and try to teach
• 5:32
  Speaker 2: themselves things.
• 5:33
  Speaker 1: You know, right, if you have a motivation to teach
• 5:35
  Speaker 1: yourself something, this is an ideal class to do it
• 5:37
  Speaker 1: because it is it being the information on whatever topic
• 5:41
  Speaker 1: you want is out there in the world. So the
• 5:44
  Speaker 1: key element that I want to focus on just here
• 5:46
  Speaker 1: for at least a moment, Michael is vetting. You've got
• 5:49
  Speaker 1: vetting in here, and of course Vetting suggests to me
• 5:52
  Speaker 1: you need to understand how reliable whatever given source is,
• 5:55
  Speaker 1: where the source material came from, so you know if
• 5:57
  Speaker 1: it's some sort of like learned treatise or it's just
• 5:59
  Speaker 1: made up whole cloth. Because I keep hearing I mean
• 6:03
  Speaker 1: almost daily, and I'm sure you do as well. Idiot lawyers,
• 6:06
  Speaker 1: for example, who let chat, GPT or some other AI
• 6:09
  Speaker 1: program create case law that they don't even bother. Going
• 6:12
  Speaker 1: back to the original law books, or go to Lexus
• 6:14
  Speaker 1: and Nexus or West Law to find out if it's
• 6:16
  Speaker 1: a real case law and behold, it's not. It makes
• 6:18
  Speaker 1: up stuff. How do you know? How can you vet
• 6:21
  Speaker 1: what you're reading and know whether it's even real or not?
• 6:23
  Speaker 2: Michael, Well, that's a good point, you know, so that
• 6:26
  Speaker 2: you're getting at the in terms of vetting, that's the
• 6:28
  Speaker 2: you're right, the information literacy judging whether something is accurate
• 6:33
  Speaker 2: and correct. Right. So I guess that that's part of
• 6:35
  Speaker 2: the vetting. And one one way you do this is
• 6:38
  Speaker 2: I would say, first of all, don't just trust. You
• 6:40
  Speaker 2: can't blindly trust the chatbot. You can't. So you you know,
• 6:43
  Speaker 2: you've got to go out and cross reference things. You
• 6:46
  Speaker 2: can ask it to give you sources and then you
• 6:48
  Speaker 2: can link to those sources and look at the source
• 6:50
  Speaker 2: and see if you, in your judgment, it's accurate. Right.
• 6:54
  Speaker 2: So information literacy and judging the validity and truthfulness of
• 6:58
  Speaker 2: information is part of the vetting. The other part is
• 7:01
  Speaker 2: is vetting information in terms of determining what what you
• 7:06
  Speaker 2: given your goal, right, since you're the teacher, you're setting
• 7:09
  Speaker 2: the goal, and you have an internal motivation, right, you
• 7:12
  Speaker 2: want to you want to learn what you're studying for
• 7:14
  Speaker 2: internal reasons, right, It's not like you're taking a course
• 7:16
  Speaker 2: and the goal is given to you, right, Right, So
• 7:19
  Speaker 2: you want to use your intrinsic goals as the guide
• 7:23
  Speaker 2: post for judging whether or not the information you're you're
• 7:28
  Speaker 2: finding that could be relevant to the goal is actually relevant.
• 7:31
  Speaker 2: So you're going to find there are things like, Okay,
• 7:33
  Speaker 2: it's tangential, I don't really need to understand that at
• 7:36
  Speaker 2: a deep level given my goal, right. So that's the
• 7:38
  Speaker 2: other part of vetting. So so it's it's sort of
• 7:41
  Speaker 2: you know, both of those things, right, Yeah, judging the
• 7:44
  Speaker 2: relevance but also the validity of the information, and it's
• 7:47
  Speaker 2: it's all going to like primary sources and checking multiple
• 7:49
  Speaker 2: sources to make sure it's it's it's.
• 7:52
  Speaker 1: Truthful, and having a fundamental knowledge of what a primary
• 7:55
  Speaker 1: source is also is a helpful thing to have going
• 7:58
  Speaker 1: into the process, isn't it.
• 8:00
  Speaker 2: Well, that's that's true, that's true. Yeah. So but even that,
• 8:04
  Speaker 2: you know, there's there's a lot of judgment that comes
• 8:07
  Speaker 2: to that because you could have a primary source, but
• 8:09
  Speaker 2: it could be biased, like you know, for example.
• 8:12
  Speaker 1: Like a scientific consensus Exactly.
• 8:17
  Speaker 2: I was going to say a research study going back
• 8:18
  Speaker 2: to COVID. Right, you could have a research study on
• 8:20
  Speaker 2: the vaccines or biromectin or whatever it is, right, and
• 8:25
  Speaker 2: you you have to bring you know, it could be
• 8:27
  Speaker 2: biased science, right, The people doing the study could could
• 8:31
  Speaker 2: have been biased, they could have changed data. So you
• 8:34
  Speaker 2: have to bring some judgment and you know there is
• 8:36
  Speaker 2: some guesswork in estimating. It's not a perfect process, but
• 8:40
  Speaker 2: you can do pretty well.
• 8:41
  Speaker 1: I think you know.
• 8:42
  Speaker 2: So.
• 8:42
  Speaker 1: So for example, if I'm interested in, like whether carbon
• 8:45
  Speaker 1: dioxide is bad for people or something, you're going to
• 8:47
  Speaker 1: have research studies would suggest it is some sort of
• 8:49
  Speaker 1: greenhouse gas, and it'll make that conclusion based upon some
• 8:52
  Speaker 1: other scientific consensus. But then again you can ask it
• 8:56
  Speaker 1: to provide an alternative viewpoint of that, can you not?
• 9:00
  Speaker 1: So it would go to other sources?
• 9:03
  Speaker 2: Yeah? Perfect, Yeah, perfect point. Yes you can. And that's
• 9:06
  Speaker 2: see that's the great thing about the chat This is
• 9:09
  Speaker 2: the thing with the chatbots, right, it's you I basically
• 9:11
  Speaker 2: look at these chatbots as they're they're expert level tutors
• 9:16
  Speaker 2: on every subject, right that are available twenty four to seven.
• 9:20
  Speaker 2: You can access them twenty four to seven and they
• 9:23
  Speaker 2: can speak to you at any level of detail you want. Right. So,
• 9:27
  Speaker 2: so what you're doing is through this iterative process I
• 9:30
  Speaker 2: chat with these chat bots like I'm chatting with a
• 9:32
  Speaker 2: with a teacher or a person right, right. And what
• 9:35
  Speaker 2: it's doing is it's continuously with with every iteration of
• 9:38
  Speaker 2: your conversation, it's continuously refining your it's understanding of your
• 9:44
  Speaker 2: perspective until it gets to a point where it's it's
• 9:47
  Speaker 2: what I call mirror, cognitive mirroring or cognitive achieving, cognitive
• 9:54
  Speaker 2: alignment with with you, right you get?
• 9:57
  Speaker 1: Is that you getting an answer you ultimately want, Like
• 10:02
  Speaker 1: an expert. If I hire an expert as a lawyer,
• 10:05
  Speaker 1: I'm pretty darn certain I can get that expert to
• 10:07
  Speaker 1: reach a conclusion that's beneficial to my client. Can you
• 10:10
  Speaker 1: do that with chat, GPT or other AI resources?
• 10:14
  Speaker 2: You can? It's it's through you know, the more it
• 10:16
  Speaker 2: understands your goal. So what one thing I'm going to
• 10:18
  Speaker 2: lay out in the talk is how you accelerate through
• 10:21
  Speaker 2: the process of having it aligned with you cognitively. Right,
• 10:25
  Speaker 2: Sure you can. So you can almost do like a
• 10:28
  Speaker 2: brain dump at the beginning of the project, so it
• 10:31
  Speaker 2: fully understands what you want to do. And this is
• 10:33
  Speaker 2: the thing, Brian, Once it knows, if it knows you're
• 10:36
  Speaker 2: working on a project, say like I use chat GPT
• 10:38
  Speaker 2: to learn about large language models and chatbots for this
• 10:42
  Speaker 2: for this talk them to give right, And once it
• 10:45
  Speaker 2: new that, I'm like, Okay, he's working on this presentation.
• 10:49
  Speaker 2: He wants to understand them, but he wants to understand
• 10:51
  Speaker 2: them at a way, at a level at which he
• 10:53
  Speaker 2: can explain them to a layperson, right, And he himself
• 10:56
  Speaker 2: is not like a tech expert, right, so he has
• 10:57
  Speaker 2: to understanding of that level. It's it's what you're spying
• 11:00
  Speaker 2: and saying, here's the here's the information, here's how you
• 11:04
  Speaker 2: can present it to your audience at a conceptual level
• 11:08
  Speaker 2: so the average person can understand it. And if you want,
• 11:11
  Speaker 2: I can create a sly for you, you know, to
• 11:14
  Speaker 2: incert in your talk. That's what it starts doing. And
• 11:16
  Speaker 2: so you know, they can be extremely helpful once they
• 11:20
  Speaker 2: really fully understand what you're doing and what your goal
• 11:22
  Speaker 2: is and how you're looking at the project. You know,
• 11:25
  Speaker 2: it's amazing.
• 11:26
  Speaker 1: Yeah, I get that, and I can see it. For example,
• 11:28
  Speaker 1: again going back to my you know my days' is
• 11:30
  Speaker 1: a litigation attorney. Sometimes you have extraordinarily complex case. You know,
• 11:37
  Speaker 1: the facts and circumstances and law and elements of it
• 11:40
  Speaker 1: that you really get used to and you're accustomed to
• 11:42
  Speaker 1: talking amongst yourselves with your other lawyers. Everybody gets what
• 11:45
  Speaker 1: you're talking about, you know, but you've got to be
• 11:47
  Speaker 1: able to explain it to a jury who has no
• 11:50
  Speaker 1: idea where you're coming from. You got to boil it
• 11:54
  Speaker 1: down easily, understandable, a story to tell which connects all
• 11:57
  Speaker 1: the dots, so it makes people that's it sounds me
• 11:59
  Speaker 1: like what you're accomplishing here.
• 12:01
  Speaker 2: That's right. So after the vetting process, once you decide okay,
• 12:05
  Speaker 2: I need these are the five things I need to
• 12:06
  Speaker 2: learn to understand this, then the next step in my
• 12:09
  Speaker 2: learning process is actually understanding it. You know, you won't
• 12:12
  Speaker 2: fully understanding once you vet it. You still have to
• 12:14
  Speaker 2: go deep into it and understand it. That's when these
• 12:17
  Speaker 2: chatbots can be incredible because you can just keep going
• 12:20
  Speaker 2: back and forth and just keep asking it questions about
• 12:26
  Speaker 2: every minute thing that you don't understand, and it'll it'll
• 12:29
  Speaker 2: explain it to you at the level you want to
• 12:31
  Speaker 2: understand it or you need to understand it. It'll clarify
• 12:34
  Speaker 2: the micro elements that you're not getting. It'll give you
• 12:38
  Speaker 2: metaphors that help you under I want to give you
• 12:40
  Speaker 2: an example. So these large language models, it's really complex stuff.
• 12:45
  Speaker 2: So I learned a lot more than the average person.
• 12:47
  Speaker 2: But I'm by no means close to you know, being
• 12:50
  Speaker 2: a quote tech expert on this, right, But I was
• 12:53
  Speaker 2: struggling with this for hours, okay, to understand it large
• 12:57
  Speaker 2: language models. And at some point the chat gave me
• 13:00
  Speaker 2: an analogy that just crystallized the whole thing for me, right,
• 13:04
  Speaker 2: And it was trying to explain, well, it's a large
• 13:06
  Speaker 2: language module. Isn't thinking, it's actually just it's it's using
• 13:10
  Speaker 2: probability to give you the next likely word in its response.
• 13:14
  Speaker 2: So it's a probability machine, right, And it gave me
• 13:18
  Speaker 2: this metaphor it was it said, it's like autocomplete on steroids.
• 13:23
  Speaker 2: It just right, So that's literally what it's doing. It's
• 13:26
  Speaker 2: like on your phone, right when you're typing a message
• 13:29
  Speaker 2: to someone, a text message. It gives you like you're
• 13:33
  Speaker 2: starting to type the next word, and it gives you
• 13:34
  Speaker 2: three options to pick one. It's calculating the most likely
• 13:38
  Speaker 2: next word you're going to use given the context of
• 13:41
  Speaker 2: what you're typing, right, Well, that's what a chatbot is
• 13:44
  Speaker 2: literally doing it, but it's infinitely more complex. The calculation
• 13:47
  Speaker 2: is infinitely more complex. But it's really just a machine.
• 13:51
  Speaker 2: It's it doesn't you know, it's not thinking, it's not
• 13:55
  Speaker 2: it's not a person. You know. It's easy to anthropomorphize
• 13:59
  Speaker 2: these things, right, you really feel like it knows you,
• 14:01
  Speaker 2: but that Chris lives for me, Like, Okay, it's really
• 14:04
  Speaker 2: just a machine. It's a probability generating machine.
• 14:08
  Speaker 1: Really, and that's that's where the human element is always
• 14:12
  Speaker 1: going to be critical and going through this exact process
• 14:15
  Speaker 1: that you're going to be talking about tomorrow night at
• 14:16
  Speaker 1: seven pm. So yes, it is, and it won't end
• 14:20
  Speaker 1: your job. AI Michael Mercer, how to teach yourself anything
• 14:25
  Speaker 1: using chat GTP log in, get registered before you do.
• 14:28
  Speaker 1: Empower you America dot org register seven PM's log in
• 14:32
  Speaker 1: time or the time to be a three hundred grede
• 14:33
  Speaker 1: Oaks driver for the live discussion. It's going to be fascinating, Michael.
• 14:36
  Speaker 1: I appreciate you doing this and helping to educate us
• 14:39
  Speaker 1: and obviously giving us some really positive uses for artificial
• 14:42
  Speaker 1: intelligence when I think most of us are scared that
• 14:45
  Speaker 1: it's going to end our careers. Michael.
• 14:47
  Speaker 2: Yeah, well thanks, I hope it. I hope it helps people,
• 14:50
  Speaker 2: so I really do.
• 14:50
  Speaker 1: So that's what you're all about, Michael, and I appreciate
• 14:53
  Speaker 1: you doing that. Get some help tomorrow night, seven pm.
• 14:55
  Speaker 1: Thanks Michael, have a wonderful day, and good luck with
• 14:57
  Speaker 1: the seminar and folks. And when you get into with
• 15:00
  Speaker 1: Galaxy concrete coatings, because you're going to