7/24/2025 - Live Testing ChatGPT Agent Mode While Planning A Five Day AI Crash Course

Description

Chapters

Transcript

0:00 Na na na. Champy.
0:05 Oh champy. Oh champy.
0:09 Oh champy.
0:12 Load the car and write the note.
0:18 Grab a bags and grab a coat.
0:24 Tell the ones that need to know
0:30 that we that we that we that we
0:35 are coming coming back back.
0:42 Oh my god. Champy, what are you doing?
0:44 What are you doing to my feet? Get out
0:46 of my feet. Get out of my feet, dog.
0:48 Good day, people.
0:53 [Music]
1:04 Standing between
1:06 [Music]
1:08 you and a hard place is insane.
1:14 Standing too near
1:17 you and a fire makes it clear.
1:24 You're trouble to me.
1:27 [Music]
1:30 Real trouble. Can't you say
1:34 [Music]
1:36 leaning in close?
1:39 Smell of your perfume scares me most.
1:45 Leading away. Hey,
1:48 you feel stronger every day.
1:55 You're trouble to me.
1:58 [Music]
2:01 Real trouble. Can't you say?
2:11 [Music]
2:16 >> Oh yes, yes, yes, yes. Champy Shannon,
2:23 what's going on good people? Hope you're
2:25 all doing well this fine Thursday night.
2:27 It's a little warm here in Denver.
2:31 [Music]
2:54 But every time I see you now,
2:57 get that look in mind.
3:01 Every time I see your mouth, I hear that
3:05 smile.
3:07 [Music]
3:08 The early misty morning light. I heard
3:12 the engine turning
3:15 the old for outside.
3:17 [Music]
3:23 Were you leaving me
3:26 again today?
3:29 You would convince me
3:32 again today.
3:35 You're leaving this hard time looking
3:38 for someone else's gold to bring
3:44 to say
3:46 so long Suz. Santa,
3:53 hush now. Don't you cry.
3:59 So long, Susanna.
4:03 [Music]
4:05 >> Don't you cry for me.
4:07 [Music]
4:17 Hello.
4:20 Hi.
4:23 [Music]
4:41 Oh. Yes. Yes. Yes. Yes. Yes. Yes. Yes.
4:43 Yes. Yes.
4:45 I
4:48 don't
4:58 I I got good news.
5:04 Are you having a stroke?
5:08 That That was comedy stuttering.
5:18 Is is he making fun of speech challenged
5:20 people? No. Maybe little. Not that's
5:24 Listen,
5:30 you leave me alone. You leave me alone.
5:35 Comedy. It was like comedy. It's the
5:37 worst it will ever be.
5:44 See, my my philosophy on comedy is just
5:47 try it all and if if you hit one out of
5:49 10, you're doing good.
5:54 Okay, Britney.
5:58 [Music]
5:59 All right. What's going on?
6:02 [Music]
6:32 How are you?
6:35 [Music]
6:44 Freedom came my way that night
6:48 [Music]
6:52 just like a jet plane in and out of
6:56 sight.
6:58 I was calling answered a million miles
7:01 an hour why wondering how hard I'd hit
7:09 when they came into the station.
7:14 [Music]
7:17 They said I was bad beyond repair
7:23 but I got no qualms with my situ. Why
7:27 are you walking away, champion? I'm
7:29 serenating you with a beautiful song.
7:34 YouTube spotlight.
7:36 Uh, what's it say? Oh, good lord.
7:38 There's many words. Hold, please.
7:43 Lori Blox got chat GPT agent just now.
7:47 In five minutes, I found four apartments
7:49 near the beach where we used to live.
7:51 We'll be calling a couple in the
7:53 morning. Excited. Sand and sand, not
7:55 snow. Nice.
7:58 Fantastic. Yeah. Get the hell out of
8:00 dodge. Get Get out of there. Get out of
8:02 You're You're in Maine or some snow
8:05 infested place, right? Go to the beach.
8:08 [ __ ] it. [ __ ] it, man. Now's the time.
8:11 Everything's Listen, everything's just
8:14 getting blown up. Everything's just
8:15 getting blown up, right? Work, the
8:19 definition of work, how we're valued as
8:21 humans. Might as well go to the beach.
8:26 [Laughter]
8:29 Oh my god, I love it. See, slug of doom.
8:33 GPT agent not getting the best reviews.
8:35 One guy had Abacus deep agent kick its
8:39 butt. I listen, I predicted this. We
8:41 We'll play with it tonight. We'll see. I
8:44 predict that it's going to be slow and
8:45 not as good as um
8:48 Jen Spark and Mannis. Producer Brandon
8:50 played with it uh before we got on here.
8:52 He said it impressed the hell out of
8:53 him. did a really cool thing. So, um I
8:56 would imagine that it's going to be
8:58 decent, not great, but you got to figure
9:01 Jen Spark and Manis have had what, six
9:04 months on it?
9:06 Uh Tik Tok question. Human being, nice
9:09 name. With so many AI haters these days,
9:12 how do you communicate to a lay person
9:16 that AI is not evil? Well, AI is not
9:19 evil. AI is just a technology. Um,
9:24 so a couple of things.
9:31 [Music]
9:36 Um,
9:41 here's my philosophy on it.
9:44 I try not to take a stance, good or bad.
9:48 Now that said, I try to be educated
9:51 about what it actually is
9:54 so that if someone comes at me and says
9:56 it steals or um you know it lies, it it
10:01 hallucinates, uh the robots are going to
10:04 kill us, you know, all of the all of the
10:06 tropes. I try to be educated enough
10:09 about what it is to be able to talk
10:12 about it, right? Because
10:16 a lot of times when someone is has an
10:18 entrenched position and here's what I
10:20 know about people that hate AI
10:24 to a person in my experience the people
10:27 that hate AI have not tried it.
10:32 They haven't tried it.
10:35 And and so where I go is I try to the
10:37 first thing I try to do is exercise
10:39 empathy. Now, if they're coming at you
10:42 online, they're going to be a lot more
10:44 emboldened. So, they're probably going
10:45 to call you a a little wiener, no talent
10:48 loser. And then if you're man enough to
10:52 not react to that, more power to you. Or
10:54 woman. I don't know if you're a man or a
10:56 woman, but um if they're coming at you
10:58 online, you know, do your best to have
11:00 empathy. But but in person, I I find
11:02 that people tend to be a bit more a bit
11:05 more reasonable. They're at least
11:06 willing to talk.
11:08 Um, so I try to understand what where
11:10 they're coming from and where their fear
11:12 is actually based,
11:14 right? Is there fear based in um I'm
11:18 going to be made irrelevant? Is there
11:20 fear based in it's going to take my job?
11:23 Is there fear based in um
11:28 I used to get work and now I don't and
11:30 I'm blaming it on on this AI thing.
11:34 Um and so I just try to understand that
11:36 and then where I go is this very simple
11:40 argument which is
11:44 independent of your opinion of it. No
11:46 matter whether you love it or hate it,
11:49 AI is not going away. And I can when I
11:54 talk about that I can talk about the
11:55 early days of the worldwide web. I was
11:57 there for the birth of that. I started
11:58 one of the first digital agencies. So,
12:00 I've got I've got I'm old enough to have
12:02 historical perspective here and I can
12:05 say I can promise you I've been in this
12:08 movie before. Not only is this not going
12:10 away, it's happening way faster and it's
12:12 dramatically more profound than the
12:15 worldwide web was. And we saw how
12:16 transformative that was.
12:19 And so then if it's not going away, you
12:21 have one of two choices,
12:24 right?
12:25 you you pick up a surfboard and you
12:28 start paddling out and learn how to surf
12:31 or you sit on the beach and just wait
12:33 for the waves to come, you know,
12:35 smashing in. Those are the only two
12:38 choices. And where I sit is at a minimum
12:43 at a minimum I want to understand what
12:47 this [ __ ] makes possible,
12:50 right? Because if if I'm just sitting on
12:52 the sidelines with my arms crossed and
12:54 I'm all pissed off and I assume I know
12:57 what it is, like a lot of the people
12:59 that I talk to that are like, "Oh, it's
13:01 not that good."
13:03 They're lines that I heard two years
13:06 ago, they're still using the same
13:08 criticisms of what Chat GPT was when it
13:11 launched.
13:13 I'm like, "Have you used it lately?" And
13:15 they're like, "Well, no, but I it's it's
13:17 the world's greatest plagiarism machine.
13:21 And then I'll do something like, you
13:22 know, what do you do? Well, I'm a
13:24 copywriter. I'm a this or I'm that.
13:25 Yeah. I'll say, ' You want to see
13:26 something cool? And then I'll just show
13:29 them something. Um
13:33 and and very often what'll happen is
13:35 they'll go, "Huh?
13:38 Oh, wait. It can do that. I didn't know
13:40 it could do that."
13:42 You know, um
13:45 sometimes I don't get that far in the
13:46 conversation. you know, they huff off
13:48 and get pissed off.
13:51 But my this whole channel, this whole
13:53 channel is dedicated to independent of
13:58 any of our opinion of AI, it's not going
14:01 away.
14:04 And just me personally, my [ __ ]
14:07 mission
14:09 is I want to be on the surfboard because
14:12 here's what I know.
14:18 It is an incredibly rare experience to
14:22 be this early when a transformative
14:25 technology is being born.
14:28 It is incredibly rare
14:31 and it's exciting and this is when the
14:33 opportunity is and this is when you get
14:35 to innovate [ __ ] that no one else has
14:37 thought of. And this technology is so
14:40 powerful that what's going to happen is
14:43 you're gonna get seduced into exploring
14:45 it and realizing, "Oh my god, this
14:47 amplifies my ideas." And what if I could
14:50 take this idea I had when I was seven
14:52 and gave up on it because stepdaddy told
14:54 me I was talentless. And what if I
14:56 brought this to the world now? Holy
14:59 [ __ ] I can do that.
15:03 That's what this technology is. It's an
15:06 amplifier of your talent.
15:09 It's not a genius that sits off the side
15:11 and replaces you. You can use it like
15:14 that and it it's shitty at that. If you
15:18 just say, "Make me some genius thing,"
15:20 it'll make you a shitty thing. But if
15:23 you say, "Here's who I am, and here's my
15:26 idea, and and here's the part I don't
15:28 know about it, and here's the part I'm
15:30 completely clueless on.
15:32 Can you help?" Holy [ __ ] will it blow
15:36 you away?
15:37 And then you're like, well, I always
15:39 thought that might be a good screenplay.
15:40 And they'll be like, of course it gonna
15:41 be a good screenplay. That's an awesome
15:43 screenplay. What a story. How would you
15:45 like to structure that? You want to use
15:47 the Sid Fields screenplay structure
15:50 format or Save the Cat? Oh, I What are
15:53 those?
15:55 We'll tell you. I like Save the Cat.
15:57 Okay. Would you like me to put your
15:59 story in a Save the Cat format? Sure.
16:02 Holy [ __ ] I got a screenplay outline.
16:06 Like in five minutes, you can you can
16:09 walk someone from I [ __ ] hate AI,
16:12 it's the world's greatest plagiarism
16:14 machine, to holy [ __ ] I I could finish
16:17 the first draft of my screenplay this
16:19 weekend.
16:25 Like one of my great joys in life is
16:27 experiencing people's first Kevin
16:29 Mallister moment which is if you saw the
16:32 movie Home Alone
16:34 that moment
16:38 most people including most people that
16:41 are using chat GPT today
16:44 have no [ __ ] clue
16:47 what chat GPT can actually do today.
16:53 They think it's one tool. They think
16:54 it's the same tool that was launched in
16:56 2022. It's not. It's like a dozen tools.
16:59 It can make movies. It can talk. It can
17:01 understand. It can see images. It can
17:03 make images.
17:06 It can do coding.
17:08 It can build a website for you. It can
17:10 build a dashboard. It can analyze your
17:13 data and make charts and graphs.
17:17 Have you done all that? It can do all of
17:18 that in the same conversation.
17:22 It's [ __ ] insane.
17:25 And there's dozens of other tools like
17:27 that.
17:29 And all of them are just there waiting
17:32 for you
17:34 to have a little bit of intention. Put a
17:36 little bit of meaning behind it. Put a
17:39 little bit of
17:41 taste
17:42 into the mix and it will [ __ ] light
17:46 up your world. So that's generally what
17:48 I talk about.
17:51 Oh man. Taught a vet manis three weeks
17:54 ago and he closed his grant for housing.
18:00 See,
18:02 see this
18:08 one of the things I espouse in here is
18:10 play with these tools. Play play. Figure
18:13 some [ __ ] out.
18:15 And then generously lead. And what
18:17 generously leading looks like
18:20 is teaching what you know or learning
18:23 out loud in this case. Whoever that was,
18:25 what's your name? Connecting
18:29 connecting grace candy attitude. I think
18:34 connecting grace candy attitude
18:38 met a vet who was trying to apply for a
18:41 grant for some housing
18:44 which I don't know if you know about
18:45 applying for grants
18:48 but it
18:50 it's a lot
18:53 and so said hey go to manism man-usim
18:58 and just type in here's my name I'm a
19:01 vet
19:02 and I'm trying to apply for this grant.
19:04 Can you help?
19:07 15 minutes it'll go off and do [ __ ] and
19:09 research [ __ ] Come back, ask you some
19:11 few questions, answer those questions,
19:13 probably fill out the form for you.
19:16 Three weeks later, he's got housing.
19:21 I see this time and time and time again
19:25 that when people
19:28 have the courage because listen here's
19:30 what I here's what I also know about AI.
19:33 It's [ __ ] terrifying.
19:36 It's terrifying.
19:39 Right?
19:41 Four years ago, everyone was talking
19:42 about blockchain this and NFT that and
19:44 everything's going to change the world
19:46 and blah blah blah blah blah.
19:49 And it turns out all those NFT dudes
19:51 were a bunch of douchebags and grifters
19:54 and [ __ ]
19:56 And so we're coming off a technological
20:00 hype cycle
20:02 that was a bunch of [ __ ] Now, do I
20:04 think blockchain and NFTTS ultimately
20:07 are [ __ ] No. I think it's important
20:09 technology. I think four years ago that
20:12 hype was [ __ ] Most of those
20:14 companies were [ __ ]
20:18 They're going to come, but that's going
20:19 to be in the future. So, along comes AI,
20:21 and everyone's like, "Oh, yeah, we had
20:23 blockchain and NFTs. You told us that
20:25 was going to be a big deal.
20:27 AI is a big deal. AI is not going away.
20:30 AI is transformative. AI is going to
20:32 take jobs." Not someone using AI is
20:35 going to take your job. AI is going to
20:37 take jobs.
20:40 And the part that no one talks about, AI
20:43 is going to invent jobs that we can't
20:45 even imagine are jobs right now. How's
20:48 it going to do that? I don't know yet.
20:51 But every technological revolution in
20:54 the history of mankind
20:56 has led to kinds of jobs that were
21:00 unimaginable to the generation before.
21:04 And we're going to get to see that in
21:05 our lifetime. But that's [ __ ]
21:07 terrifying.
21:12 So anyway,
21:16 um
21:18 I've been on vacation for a couple of
21:20 weeks. Wait, you can't go on vacation.
21:22 This is AI.
21:24 Um what did I miss beside Kyle? Kyle
21:27 increasing his forehead and agents. I
21:29 know. My my forehead's looking
21:31 fantastic. Look how big my brain is. I
21:34 didn't realize I was that smart.
21:38 All right, listen people. We've got
21:40 we've got a situation here next week.
21:44 Next week,
21:48 [Music]
21:52 producer Brandon, I need a screen swap.
21:58 [Music]
22:00 There we go. Am I not sharing?
22:05 Uhuh. I am.
22:06 Okay. Um,
22:09 next week on this very channel,
22:14 I'm doing a fiveday
22:16 back to basics AI crash course.
22:20 Um, and you're like, what's a what's a
22:23 back to basics AI crash course? It's
22:26 what it sounds like.
22:31 So when I first started this channel, a
22:33 lot of what I did was just talk about
22:36 how do large language models work? How
22:38 do those image generation tools work?
22:40 Why is AI a big deal? Um I used to give
22:44 presentations on it a lot and then you
22:46 know time has passed and and that stuff
22:48 went away. So, um,
22:51 next week, what it's going to look like
22:53 is five nights a week. Next week,
22:55 starting on Monday, we're going to start
22:57 at 8:00 p.m. Mountain on Tik Tok on the
23:00 AI Learning Lab on Tik Tok. And we're
23:02 going to have a pre-show. We're gonna
23:03 have like a pre-show party. We're just
23:05 going to hang out. I'll probably sing
23:06 with the dog, whatever. Who knows? We'll
23:09 chitchat. You can ask questions.
23:12 And then at 8:30
23:15 on
23:18 YouTube
23:20 at learninglab-ai
23:22 and you can zap that QR code if you
23:24 want.
23:26 At 8:30
23:28 we're going to start the session and
23:30 it's going to be a 90 minute session.
23:31 It's going to be one hour approximately
23:34 of me demoing stuff and then it's going
23:37 to be 30 minutes of Q&A
23:40 and just if you've if you're an
23:42 irregular, if you've been here a long
23:44 time, you know how this is going to go.
23:45 I'll probably run down rabbit holes.
23:47 It'll probably be chat ad like, but the
23:50 general shape of the week is this. Night
23:53 one, Monday night is um here, let me go
23:57 back to this
23:59 slide. Monday night is what's AI? Why is
24:04 it a big deal? Is it a big deal? Is it
24:06 here to stay? How do you know that? Like
24:09 what is it really going to take jobs?
24:11 How does it work?
24:14 I personally feel that understanding how
24:17 these tools work is an incredibly
24:20 valuable insight. You don't need to know
24:22 math,
24:23 but you need to know that it's not
24:25 copying and pasting documents.
24:28 It's not a plagiarism machine. It's just
24:31 not how it functions.
24:34 So that's Monday night. Tuesday night is
24:37 chat GPT 101. And hopefully it's like
24:41 101, 102, 103, 104.
24:44 There's a lot in chat GPT. So, we're
24:47 just going to start with that. Just with
24:48 free chat GPT, you could probably spend
24:51 the next year and still not scratch the
24:53 surface of what that [ __ ] makes
24:54 possible. Okay. Night three is going to
24:57 be the creative night. Making pretty
24:59 pictures, making music, making all the
25:03 stuff, films,
25:06 audio, voice synthesis, avatars.
25:11 Thursday night is going to be like a use
25:13 case mashup where we sort of take a use
25:16 case where we've got lots of different
25:18 tools we bring together and how do you
25:19 figure out which tool to use when
25:22 and then Friday night's going to be just
25:24 a blasting long Friday night date night
25:27 Q&A session. All right,
25:30 so here's my request. My request is I
25:33 put on
25:36 um in the AI irregulars on the in the AI
25:39 salon.
25:41 I uploaded the graphic for the the crash
25:44 course and then I put a comment and I
25:46 have the graphic for the YouTube
25:49 uh thing, the YouTube uh URL. My request
25:53 is go share this with your with your
25:56 networks. I want to get as many people
25:58 here as possible next week. it. This
26:01 doesn't cost a thing. This is free. I'm
26:03 doing like five nights of full-on
26:06 [ __ ] workshop for free.
26:09 Why?
26:11 Apparently, my relationship with money
26:13 is flawed.
26:15 [Laughter]
26:19 [Music]
26:25 Hey, Mr. T. What's happening?
26:29 [Music]
26:40 is his place. I can rest my forehead.
26:46 Yeah. My thoughts in sweet silence
26:52 is this place where the feelings aren't
26:55 dead. It's like it's like poor man Bob
26:57 Dylan. Um,
27:03 we don't need no stinking consonants.
27:09 Oh, Kyle, can I do a video duet
27:15 to promote back to basics? Sure. You
27:17 I've got a I've got a Tik Tok video up
27:19 there. Do the [ __ ] out of it. I think
27:21 you can do that anyway.
27:25 Can you give examples about how to how
27:27 to use chat GPT agent? Um uses, best
27:31 practices, do and don't t talk. Thanks
27:35 to Rick Rubin of AI. Um we're going to
27:37 play with agent tonight. I will admit I
27:40 have not played with it yet. Um I have
27:42 some ideas.
27:44 Um
27:48 come in.
27:50 So, we're going to go play with Agent
27:52 and I think that'll probably take the
27:53 whole night because I think those things
27:55 are slow. They're slower than [ __ ] Um,
27:58 and I got to be a little judicious about
28:00 them because I think I only as a plus
28:02 member I think you only get 40 agents a
28:05 month.
28:07 So about one or two a day
28:13 [Music]
28:17 in a westerly direction.
28:24 This car is my train.
28:27 [Music]
28:29 Been driving. I've been wondering
28:33 what it is I'm running from again.
28:38 feel like an ity old man. Um, if you're
28:43 wondering, am I ever going to do
28:46 anything here tonight?
28:51 I might.
28:53 This is one of the joys of having this
28:55 channel is
28:59 the agenda is
29:02 primarily driven by my ADHD adult brain.
29:12 Oh my god. So, doing a back to basics
29:14 night should be interesting next week,
29:16 but I think I can hold it together. We
29:18 got talk about AI night on Monday, chat
29:20 GPT on Tuesday, some creative [ __ ] on
29:22 Wednesday, bringing it all together on
29:24 Thursday, and hopefully some people have
29:27 some questions on Friday. We'll also do
29:29 a Q&A every night. So, I'll I'll work
29:32 for about an hour and then we'll do 30
29:34 minutes of Q&A at the end, which if
29:36 you've ever been here, my
29:40 one small question could turn into a 45
29:43 minute ramble for me.
29:48 So,
29:50 [Music]
29:56 look, here it is. Here it is. We got it.
29:59 See right there? Agent mode. 40 left.
30:02 More available on August 23rd. That's
30:05 actually nice that they did that. That
30:07 they gave us your your count and when it
30:10 resets right in there. That's really
30:11 nice.
30:16 [Music]
30:30 This is never going to work.
30:32 [Music]
30:37 All right, let's get going.
30:40 Let's get going. Anybody have any
30:42 questions before we start? Anybody got
30:45 any questions? What are we doing
30:47 tonight? We're going to work with chat
30:50 JPT agent mode.
30:53 It's going to be really exciting. Unless
30:56 it's not.
30:59 One of the things we like to do here at
31:01 the learning lab. Is it learning? No.
31:04 No. No. We don't. Not a lot of learning
31:07 going on here.
31:14 Uh, we're gonna play with it. We're
31:16 gonna play with it. And it, Listen, if
31:19 it sucks, we'll call it. If it's great,
31:22 we'll call it. My prediction is it's
31:24 going to be somewhere in the middle. I
31:26 think it's going to be interesting. All
31:28 right.
31:30 Are we ready? I only have 35 left. I
31:32 know. I'm a little anxious to uh to do
31:35 it. All right. Here's agent mode.
31:37 suggested
31:39 reports.
31:43 Oh, they got a whole bunch of [ __ ] down
31:45 here. Okay. I I I I'm I'm picking up
31:48 what you're putting down there, chat
31:50 GPT. Uh if you have some If you have a
31:52 scrolling list below
31:56 a box, could you not eat up the entire
31:59 top of the screen with nothing? Chat
32:02 GPT,
32:03 anyone?
32:05 Any any UX designers in the house?
32:10 Why is there not a cranky Gen Xer
32:12 sitting in the corner at Open AI? Uh,
32:15 guys.
32:25 I'll do it.
32:29 Hey, Sam Alman, in case you didn't know.
32:34 Sam, um,
32:37 listen, I know you lost some top
32:39 engineers to Meta for like hundred
32:42 million dollar signing bonus and a $300
32:44 million base. I'll sit in the corner at
32:47 Open AAI and [ __ ] about your UX for
32:50 like 10 million a year, like onetenth
32:53 the price, and your product will suck
32:55 less. And and I'll get to like [ __ ]
32:58 about things and get paid for it, which
33:00 is my dream.
33:02 What do you say?
33:05 You in? All right.
33:08 Liberal arts major, Gen Xer,
33:12 missing from Open AI.
33:14 Yeah, you got a bunch of zenials and
33:17 millennials
33:18 and neurosicy
33:21 superbrains.
33:23 You need to balance that out with some
33:25 humanities and a little bit of rage.
33:28 Okay. Okay. All right, Sam. Thanks. All
33:32 right. Did you want to play with this
33:34 thing?
33:37 I don't I don't know why one of you
33:39 suggested I talk to Sam directly.
33:45 Sharon Crawford will do it for 9
33:46 million. Listen, we got a bidding war
33:48 over here at the AI learning lab. Sam, I
33:51 think you could pick up three or four
33:52 Gen Xers for 510 million bucks. I think
33:56 you could do it. It's a [ __ ] bargain.
33:59 I mean, you're worth $3 trillion. Come
34:01 on. All right.
34:07 Oh, lordy.
34:11 Okay. Find the best San Francisco Giant
34:13 tickets on StubHub. Research carbon
34:15 capture costs. Audit fast fashion versus
34:19 slow fashion wardrobe cost and impact.
34:22 Good god, that sounds awful, but you can
34:25 have an agent do it now. Research the
34:27 top refurbished stand mixers. Oh,
34:30 suggested.
34:31 Build a discount discounted
34:34 cash flow model. No,
34:38 these are suggested based on what?
34:42 All right. Actions. Order pizza delivery
34:45 for 10 people on Uber Eats. Book dog
34:48 friendly hip camp and hot tub near San
34:51 Francisco. What's a hip camp?
34:56 Hire a piano turner tuner verse via
35:00 thumb stack for Saturday. Conduct
35:02 conduct a calendar audit based on Okay.
35:06 Oh, I know what we're going to do.
35:10 Um
35:21 trying to think of something simple.
35:27 Shut up, producer Brandon.
35:30 I'm I'm getting trolled by my own
35:32 producer. I want you to know if you
35:33 think this is easy, it actually is. It's
35:37 pretty easy. You just kind of sit in
35:38 front of a computer and uh talk. It's
35:41 It's not hard. It's not hard.
35:46 Haven't researched different AI stuff
35:48 and put it in a spreadsheet. I was going
35:49 to do that. I was going to do like a uh
35:52 go on go on eBay and research the prices
35:55 for like
35:57 um used Dyson vacuums or
36:03 I don't know something like that. That's
36:05 kind of boring. Um all right,
36:09 spreadsheets presentations. Tom on
36:11 YouTube. Low restructure the US
36:13 government.
36:19 Um,
36:32 let's let's
36:34 turn off agent. I'm going to turn off
36:35 agent for a second. We're going to do
36:37 chat GPT 03.
36:40 So, by the way, if you're new to chat
36:42 GPT,
36:44 chat GPT 40
36:47 is
36:50 is not better than 03.
36:54 03 is actually better than 40, but for
36:56 different things. They're actually two
36:58 very different models on two very
37:00 different tracks. So 40 is the evolution
37:03 of the [ __ ] that launched in 2022. 03 is
37:07 the second version of the OpenAI
37:11 reasoning model. And you're like, well,
37:12 wait, if it's the second version,
37:14 shouldn't it be called O2? Yes, it
37:17 should. But someone else had the
37:19 trademark, so they just named it 03
37:22 instead of naming it chat GPT reasoning
37:24 2, which would have been fine. Would
37:28 have been fine.
37:31 They didn't have a cranky Gen Xer
37:32 sitting in the corner office going, "No,
37:37 you can't name it 03 when you call the
37:40 other one 40. It will confuse people.
37:43 That's stupid."
37:46 But they didn't have that and so
37:49 engineers said well if the O if the O
37:52 precedes the digit that means open AI
37:54 and that that in implies infers that
37:57 it's a it's a reasoning model and if the
37:59 O comes after the number then then that
38:02 obviously stands for om omni omni as in
38:06 as in multimodal it's it's it's chat
38:09 GPT4 omni or or multimodal clearly
38:13 different than 03 which is open AI3
38:15 three, which is the reasoning model.
38:17 It's it's very logical. You you just
38:19 need to pay attention to whether the O
38:21 comes before or or after the number.
38:25 Oh, if it's not an omni model, then we
38:27 just have the number and there's no O.
38:28 Well, there's 4.5. That you you could
38:31 say there's 4.0, but zero is different
38:34 than O. 40 is
38:42 Yeah.
38:45 I'm sick of them, too.
38:51 So, we're going to use 03, which is the
38:52 reasoning model. And what we're going to
38:54 say is
38:56 um
38:59 I just got access
39:02 to open AI
39:06 agent
39:09 and I want you to research
39:13 how people have been using it
39:20 for interesting
39:24 use cases in the humanities. Now, why
39:28 I'm putting in the humanities is because
39:30 no one's using it for the humanities as
39:32 my guess. So, while that's working, I'm
39:34 going to go over to X and I'm going to
39:36 go to Grock
39:39 and I'm going to say um based on
39:45 what people are talking about,
39:50 open AI chat, GPT
39:55 agent,
39:57 new functionality,
40:02 Um
40:05 character or no categorize categorize
40:09 categorize the
40:15 best use cases where people are seeing
40:21 good success.
40:24 Uh, and if you didn't know this, Grock
40:28 is the only of the large language models
40:31 that can actually search Twitter.
40:34 So if you want to see what people are
40:36 talking about in real time, go use
40:37 Grock.
40:40 I think that's cool. Okay, back to here.
40:43 Are we done? It thought for 35 seconds.
40:47 Culture, cultural heritage super
40:49 interns, rapid transcription and
40:52 enrichment of archives.
40:54 University special collections teams
40:57 piped scanned manuscripts
41:00 into
41:06 that's cool. Okay. Bulk literature
41:09 review scholar in a box. Hist agent.
41:14 Wait, these aren't
41:17 these aren't talking about Okay,
41:20 I'm going to say you misunderstood.
41:26 Open AI has a new chat
41:32 GPT feature
41:35 called, what's it called? Agent
41:46 Wait, all of a sudden I don't have agent
41:47 anymore. I got to reload the page, I
41:49 think.
41:51 [Laughter]
41:57 Wait, why do I not have it anymore?
42:01 Oh, because I know why. Ah,
42:04 here's a welcome to chat add.
42:08 If one of your questions, if you're new
42:11 here and one of your questions is, "Will
42:13 Kyle ever finish a sentence?" I might.
42:18 Um,
42:21 there's a behavior in chat GPT that that
42:25 is really [ __ ] annoying. And it is
42:29 if you switch models midway through a
42:32 conversation and it
42:35 whatever model you were using before
42:38 doesn't work with a new feature, you
42:40 can't use that feature anymore at all in
42:42 this chat. So I can't turn on agent in
42:45 this chat, which is okay. I don't need
42:47 to. You must understood. OpenAI has a
42:49 new feature called
42:51 agent
42:53 which is an autonomous
42:58 agent like Manis or GenSpark.
43:04 I want you
43:07 to
43:08 find me use cases for that tool. Okay.
43:16 Uh uh uh
43:18 I saw others using chat GBT agent to
43:22 search on X. That's interesting.
43:26 Will cook for friends.com.
43:30 Elephant garlic. What it? Huh?
43:35 What are you doing? Jeff Flanigan. Oh,
43:36 you're talking to Tom Nodler.
43:40 All right. Hey everybody. Ann Murphy.
43:42 What's shaking? What's shaking? Bacon
43:52 Silver Fox. It's great.
43:56 Oh, the Theo Von interview with Sam
43:57 Alman. If you have not seen the Theo
44:00 Vaughn interview with Sam Alman, I think
44:03 it's must mustwatch television.
44:06 It really is. It is very much like the
44:08 1996 interview with Katie Kirk and and
44:11 Brian Gumble talking about the internet.
44:14 What is internet anyway? That one.
44:17 Remember that one? You know, it's it's
44:19 the it's the A with the circle around
44:21 it. I think it stands for about
44:33 uh Theo Vaughn plays Bryant Gumble and K
44:37 Katie Kirrick combined. He's got Katie
44:38 Kirk's hair and Bryant Gumble's
44:40 intellect. Uh I actually Theo Von's
44:43 actually really smart and and um he's
44:46 incredibly sweet and and well-meaning.
44:50 Um, but man, is it a fascinating
44:52 interview. It It's
44:58 I think Theo Vaughn represents like the
45:01 everyman of, you know, people that don't
45:04 understand AI and just assume that
45:06 people in San Francisco are a bunch of
45:08 evil [ __ ] out to take over the world.
45:12 Uh, Tik Tok comment.
45:15 Well, [ __ ] now I have to go do that,
45:16 Kyle. I know, Ann. Sorry, Ann. I gotta
45:19 tell you, totally. If you haven't
45:21 already, smash down a gummy or two, let
45:24 it kick in a little bit and then go
45:26 watch Theo Vaughn. I swear to God, the
45:28 the interview opens up. We should go
45:29 watch it. The interview opens up with
45:31 Theo Vaughn saying to Sam Alman, "So, do
45:35 you think this is the last 40 years
45:37 where we'll have babies inside of human
45:39 bodies?"
45:43 And Sam Alman goes, "You mean like we'll
45:47 have them in vats?"
45:51 The Van's like, "Yeah."
45:53 And Sam Alman's like, "I mean, I guess
45:56 we maybe could because that would be
45:58 bad, right?" And Theo Vaughn's like,
46:01 "Yeah, that's what I thought, but I
46:03 thought that would be the the position
46:06 that you would take." Oh my god, it's so
46:09 funny. Tik Tok pin. Hey Kyle, you always
46:11 have entertaining Oh, you're always
46:13 entertaining and helpful. Thank you,
46:14 Darren Drew. I appreciate that. Um,
46:18 okay. Let's go see if uh if chat GPT's
46:22 research did anything here. Autonomous
46:24 multi- multi-step task execution. Well,
46:28 I understand it can do that. That's what
46:30 it is. Core building block. Okay. Ah,
46:33 potential humanities use cases.
46:36 Research, synthesis, and analysis. deep
46:39 archival exploration.
46:42 You know what would be cool?
46:49 I have a fun one.
46:53 All right, I got one. New chat. This
46:56 one's going to be weird.
47:04 Agent mode.
47:15 I'm
47:18 curious about
47:21 ancient
47:23 texts texts,
47:28 especially those
47:31 that are not
47:34 yet that wait that are
47:38 digitized
47:40 but not yet translated.
47:53 I want you to go find
47:59 a university
48:05 or
48:07 museum
48:09 archive of scanned
48:13 ancient texts.
48:23 and
48:27 translate them.
48:37 And put a link to the
48:42 original scanned image.
48:46 as well as the translation
48:52 and uh
48:55 and relevant
48:58 metadata
49:01 about that document
49:05 into a spreadsheet.
49:14 You can stop after you
49:18 complete
49:23 10 of them.
49:25 Let's do
49:29 five from Egypt
49:33 and five from some other culture.
49:38 All right, that good. Any any
49:41 suggestions to add to that?
49:45 When Kyle met Sam. Imaging the
49:46 irregulars. Okay, I'll go look for that.
49:49 Yeah, finally agent has arrived. Secret
49:52 ancient man. Okay. Any any uh that
49:56 sounds good, right? So,
49:59 have it go find an image with like
50:01 Egyptian hieroglyphs on it, translate
50:04 those, put the translation into a
50:06 spreadsheet, and do 10 of those. Uh oh.
50:09 Producer Brandon Producer Brandon
50:11 approves. Okay, good. All right, we've
50:13 we've got approval. We got approval from
50:16 upstairs. We're going in. We're doing
50:18 it.
50:20 All right. Thinking setting up my
50:23 desktop.
50:26 Okay, understood. I'll search university
50:29 or museum archives for 10 digitized
50:31 ancient texts. Five from Egypt and five
50:33 from other cultures that have been sc
50:36 Oh, say now cool. So, look at the
50:38 background of this thing. This is cool.
50:41 So, this is very much the same
50:43 background as um is in the the little
50:47 talkie advanced voice thing when you
50:50 talk to chat GPT.
50:55 And I can go back on this, right? How
50:57 does this work? Yeah. So, this is this
51:01 is a play bar down here. So, I can play
51:02 what it's doing.
51:05 All right. The results are not set.
51:07 Language support. I'm opening the result
51:09 from Claremont College. Bad gateway.
51:12 It's searching.
51:16 Right. All right. So, this thing is off
51:18 doing its thing. That's a great query.
51:21 Thank you, Silver Fox. Appreciate that.
51:23 Yeah, it's a good one. Right now, I'm
51:25 going to do another one. I'm going to I
51:27 because I think the way these things
51:28 work, right, we're in the middle of a
51:30 chat here, right?
51:34 So, I'm going to say um
51:38 create a uh a list.
51:43 Let's see. Create uh create
51:46 create a presentation
51:52 of the
51:55 top 10
52:00 um ancient
52:03 document archives
52:06 in the world.
52:10 With
52:14 one
52:16 uh uh I'm going to put hero document.
52:20 >> Hey Kyle.
52:21 >> Yeah.
52:23 >> Uh you'll need to enter that information
52:26 into a new chat window. If you push it
52:29 there, it'll interrupt what
52:32 >> it interrupts the other one. Ah, okay.
52:34 All right. Well, then that sucks. But
52:36 that's okay because I'll I'll do
52:38 something different if I if I have to do
52:39 it in a different window. Um, let's let
52:42 that let's let this thing go. Activity
52:45 take over the browser stuff. Okay, I'm
52:47 not going to have to log into anything.
52:48 It'll just go find open source stuff.
52:50 Okay, so let me do a new chat
52:54 and I I'm going to do this. I'm going to
52:56 say um I want you
53:00 to analyze
53:03 the branding
53:06 elements at storyvine.com which is my
53:10 company and create a
53:16 um brand
53:20 guidelines page
53:24 complete
53:26 with logos,
53:29 uh, brand colors,
53:33 fonts,
53:35 um,
53:38 graphic
53:44 elements like uh, corner,
53:48 radius,
53:50 et
53:52 corner, corner,
53:55 radus,
53:59 etc.
54:01 Um,
54:04 build this brand
54:09 guide into a presentation.
54:17 Let's just let's just see how it does.
54:19 So, wait. So, I got to turn on agent
54:21 mode.
54:23 And then bang, off it goes.
54:26 I'm down to 38 people.
54:30 Was that God speaking to Kyle? This
54:32 agentic ethereal voice from nowhere.
54:39 Uh, are you on a pro plan or a plus
54:42 plan? I'm just on the plus one. I used
54:44 to be on pro, but then I gave it up.
54:46 Eddie, why not just talk instead of
54:48 typing? Uh, because I'm on I I I can. I
54:51 just I one of the things I realize is
54:54 that when I type and talk at the same
54:56 time, it bugs the [ __ ] out of people.
55:02 No, I should do that. I I have this bad
55:06 optimistic view that I'm just going to
55:08 type in a short little sentence and then
55:10 because I have diarrhea mouth, I just
55:13 keep typing and talking and it is clear
55:15 that I should have just talked it. Uh so
55:17 I will do that next time. I promise,
55:19 Eddie. I I should not torture you like I
55:22 have been, but it does kind of give me a
55:25 sadistic kind of joy that that it bugged
55:27 you.
55:33 Oh, man. Alive. Man alive. Man alive.
55:37 Man, man. Man, man.
55:39 Okay. Uh uh uh. Oh, let's go look at Oh,
55:42 wait. What's this doing? Thank you for
55:44 your request to analyze the elements. To
55:45 clarify, do you have any specific
55:46 preferences of layout or is it okay if I
55:49 browse it? I'm going to say um
55:53 use pro uh let's see use branding. Look,
55:57 I'm not using my talkie thing bits. I'll
55:59 do it now.
56:04 Best practices.
56:07 Oh, come on. Best practices. Yeah, it's
56:11 not working anyway. Use branding. best
56:14 practices
56:17 uh use branding best practices uh for a
56:21 brand book
56:25 thinking.
56:31 Should it be able to also do external
56:33 stuff like book a discussion
56:36 for the brand results and suggest travel
56:39 plans for the meeting? Yes, it should,
56:41 Gwen. And and what it will likely do is
56:44 when it comes to things like, hey, you
56:46 need to put in your credit card or you
56:47 need to log in. Look, there it is on my
56:49 website. That's kind of cool.
56:52 So, one of the things that agent does
56:55 really well or what they claim it does
56:56 really well is it it successfully
56:59 combines a couple of elements. They had
57:02 this thing called deep research which is
57:04 really good at reading the content of
57:05 websites and analyzing them. And then
57:07 they had this thing called uh operator
57:10 which was good at visually surfing an
57:12 interface. And what they've trained this
57:16 agent model to do is to understand based
57:20 on this where it is in the process which
57:23 of those tools to use or if it needs to
57:25 create its own tools using Python it can
57:27 do that too. Um the interface here I
57:30 think is quite good. The fact that it's
57:32 inline in a chat is really good. This
57:34 window is there and available. I can
57:36 kind of see what's going on.
57:38 We need to gather logo things. It's
57:40 grabbing logo.png, which is 900 by 1299.
57:45 Look, it's going in and reading.
57:48 Oh, now it's Oh, it's it's reading.
57:50 Yeah. Content.
57:54 It grabbed the logo. All right. It's off
57:56 doing its thing. Look, there's our logo.
57:57 Look at there. You see that? Any English
58:01 teachers in the room are probably
58:02 slapping their foreheads about talking
58:04 and speaking. Or am I the only English
58:07 snob here? Okay, I'll shut up for now.
58:10 What did I say that was out of line?
58:12 Speaking or talking?
58:14 Tell me. I I will learn. I am not
58:16 closed-minded.
58:21 All right. Ancient text search. How is
58:23 this going? Still going.
58:29 I'll open line 340 to view the
58:31 translation. Oh, wow. It's doing
58:33 translations.
58:41 Wouldn't it be cool if we found some
58:42 [ __ ] crazy ass document and and we
58:46 get credit for it? We get the irregulars
58:48 get credit for
58:51 for for unlocking some new new found
58:55 text of the Bible or you know
59:01 oh man um
59:04 get the full picture on YouTube. All
59:06 right so that's doing that. Um
59:12 let's go back to our branding
59:14 exercise. I'll quickly view the file as
59:17 requested.
59:19 So if I go back here, yeah, I can see
59:20 what it did. Oh, this is so cool.
59:32 Absolutely crazy.
59:43 I'm starting by generating images for
59:44 the brand guidelines document.
59:48 That's pretty cool.
59:54 Does it show the time it will take? It
59:56 doesn't show the time it will take Gwyn
59:58 because it doesn't know like like it
1:00:00 sets out the way these things work is
1:00:02 they kind of set out a plan and then
1:00:04 they go start looking for [ __ ] and then
1:00:08 they modify the plan. Like when Brandon
1:00:11 did his, he said at one point it
1:00:13 recognized that it went from page seven
1:00:15 to page nine and it realized on page
1:00:18 nine, oops, it looks like I missed page
1:00:20 eight and it went back to page eight and
1:00:22 and kind of started over. So like that
1:00:24 that kind of stuff it can't anticipate.
1:00:28 The Egyptian one says drink more oval
1:00:30 teeth.
1:00:33 Um
1:00:36 all right it's off doing its thing.
1:00:38 Nebuchadnezzar cylinder Morad details
1:00:42 ununiform digital library initiative.
1:00:45 This is so cool.
1:00:49 This is groovy. So yeah if you if you
1:00:51 don't know what's going on here. So, so
1:00:53 the way this thing works,
1:00:56 I'm in a chat, GPT chat,
1:01:00 I said I wanted to go do this ancient
1:01:03 thing, right? And I have the agent tool
1:01:05 turned on. This this blue window is a
1:01:09 virtual machine. It's meaning
1:01:13 it's think of it like a hermetically
1:01:15 sealed fake computer inside my chat.
1:01:20 And inside that fake computer, it's
1:01:23 actually running code. It's got its own
1:01:25 browser. It can write Python and execute
1:01:28 it. It can it can actually log in. It
1:01:32 can take you to a website where you need
1:01:34 to log in your stuff. So, if you want to
1:01:38 go book travel tickets, maybe it goes to
1:01:40 Expedia, it finds you all of your
1:01:42 destinations and when it goes to book
1:01:44 the ticket, you actually have to log
1:01:47 into your account. So, then you can sort
1:01:48 of take over log into your account,
1:01:51 which I understand is kind of janky and
1:01:53 slow understandably for now.
1:01:57 And then once it's logged in, then it
1:01:59 can go take actions on your behalf. And
1:02:00 they've got it set up so that it's it's
1:02:03 not it's not going to take any action
1:02:06 that would be something like spending
1:02:08 money on your credit card without like
1:02:10 double and triple checking with you. So,
1:02:12 it's probably going to be annoying, but
1:02:15 it's just off doing all this stuff. All
1:02:17 right.
1:02:19 I'd love to see these projects working
1:02:21 themselves out over the world in real
1:02:24 time. A digital gallery. That's actually
1:02:26 a great idea, Silver Fox.
1:02:28 I don't
1:02:33 You could certainly do that. Well, you
1:02:36 know what's probably going to happen,
1:02:38 Silver Fox, is there's probably going to
1:02:40 be an API for this that you'll be able
1:02:43 to call these things within an
1:02:45 application. failed to get upload status
1:02:48 of a file.
1:02:50 Did it just kill the
1:02:54 Oh.
1:03:07 Uh, let's see.
1:03:14 Egyptian test texts Michigan papyrus
1:03:19 receipt for tax grainis 276 AD. This
1:03:24 Greek papyrus
1:03:26 records the delivery of tax grain in
1:03:30 Karanis during the first year of emperor
1:03:33 Probus. The API's translation note that
1:03:37 it reads year one of our lord's probus.
1:03:41 Wait, the API's translation notes
1:03:46 fragment. This summary explains that six
1:03:48 documents.
1:04:02 Yeah, here's the translation.
1:04:04 It didn't do what I wanted to to do.
1:04:07 All right, I'm going to yell at it.
1:04:12 Okay. So, I'm gonna say I have two
1:04:16 issues.
1:04:20 One,
1:04:23 the CSV
1:04:25 isn't there.
1:04:31 Hyperlink
1:04:33 is dead.
1:04:36 And two,
1:04:38 you just pulled translations
1:04:44 from
1:04:49 the library or museum um
1:04:57 archives.
1:04:58 I
1:05:00 explicitly
1:05:02 asked you
1:05:04 to find images
1:05:10 that have not have
1:05:13 not been translated and use your vision
1:05:20 capabilities
1:05:21 [Music]
1:05:23 to translate
1:05:26 them. A fresh.
1:05:31 All right. Duly yelled at.
1:05:34 Hate those dead links it delivers. Can
1:05:37 you yell at it in canvas mode or is that
1:05:40 not available when you use agent? It
1:05:42 didn't seem to put that in a canvas.
1:05:45 All right, it's thinking right now.
1:05:47 Let's see if it thinks well enough to do
1:05:49 it right this time. All right, let's
1:05:52 just make sure that spins up an agent.
1:05:55 Okay, spinning up an agent. I'll address
1:05:57 the file sync issue. The previous cursor
1:06:00 something.
1:06:09 All right. It's off.
1:06:18 Okay. So, it's going to do its own
1:06:19 translation. That That'll be
1:06:21 interesting.
1:06:26 But it's it's still not what I asked
1:06:28 for. So, it's being of a bit of a
1:06:30 [ __ ]
1:06:32 Um, new chat. Is this my storyline one?
1:06:59 I'm trying to figure out
1:07:04 I feel like there there are going to be
1:07:06 some professions where it's really
1:07:08 obvious what to use this for. Like
1:07:10 research feels like a biggie to me.
1:07:15 And there going to be other areas where
1:07:17 it's not going to be as obvious when you
1:07:18 would use this. But
1:07:22 one of the reasons to join something
1:07:24 like the AI salon,
1:07:26 which
1:07:29 if you've not joined the AI salon,
1:07:32 go to community.thesal salon.ai
1:07:37 and join us because
1:07:40 this is a community of people trying to
1:07:42 figure this [ __ ] out at the same time
1:07:45 together and try to support one another
1:07:47 in that. Um,
1:07:50 if there was a pin there, I missed it.
1:07:52 Kyle, I respect how you still type out
1:07:54 instead of talking voice. I need to do
1:07:56 that. Shut up. Shut up. My My text my
1:08:00 voice to text isn't working right now.
1:08:03 But, but I I I have in the past used it,
1:08:07 but right now it's not working. So, but
1:08:10 but I appreciate the the uh semi the the
1:08:13 pseudo compliment.
1:08:15 Pseudo compliments are good. I like
1:08:17 them.
1:08:20 I've watched you for three years. I'm
1:08:22 now going to join. There you go. Um,
1:08:26 it really is. It's it's it's a really
1:08:28 powerful community. I I gotta say,
1:08:31 surrounding yourself with people that
1:08:34 are trying to figure this AI [ __ ] out is
1:08:36 probably one of the smartest moves you
1:08:38 can do right now.
1:08:41 Being ready for AI, being prepared for
1:08:43 AI is no longer like a nice to have.
1:08:47 This is a requirement. A lot of jobs
1:08:49 right now do not require college
1:08:52 degrees, but do require you to have a
1:08:54 clue about AI.
1:08:57 Think about that.
1:09:00 You can go work at Facebook without a
1:09:02 college degree. If you have a brain,
1:09:06 you can't go there and work and work if
1:09:08 you don't have your [ __ ] together with
1:09:10 AI, though, right? Tik Tok pin talking
1:09:12 about back to basics. Next week is going
1:09:14 to be fire. Yes, Mary. Mary, next week
1:09:16 is going to be fire. So, next week, by
1:09:18 the way, if you're new here,
1:09:27 all week next week, starting at 8:00 on
1:09:30 Tik Tok, we'll do a half hour pre-show.
1:09:33 And then at 8:30 Mountain on YouTube,
1:09:37 five nights,
1:09:40 what is AI night one chat GPT 101 night
1:09:43 two
1:09:45 creative deep dive or no creative
1:09:48 shallow dive night three Wednesday
1:09:51 putting it all together multi-
1:09:55 use cases that have to use multi
1:09:57 different models and how do you choose
1:09:59 which ones to use and when and why and
1:10:03 And then Friday's going to be a full-on
1:10:05 Q&A.
1:10:06 All right.
1:10:08 So, 90 minutes every night. And if
1:10:10 you're on the East Coast, yes, it starts
1:10:13 at 10:30 at night and ends at midnight.
1:10:17 Why did I do that? I don't know. You're
1:10:21 in New York. Figure it out.
1:10:27 You're not going to go out to dinner
1:10:28 till midnight anyway. And sure, not
1:10:31 everyone east of the Mississippi lives
1:10:34 in New York City, but that's where I
1:10:36 moved from. So that's my point of
1:10:38 reference. So if you're in South
1:10:39 Carolina, sorry, it still starts at
1:10:41 10:30. And you're like, "But wait, can I
1:10:44 watch it on demand?" We we will probably
1:10:46 make this available on demand, but it's
1:10:48 not going to be free. The free course is
1:10:51 going to be next week 8:30 to to 10
1:10:54 o'clock on uh
1:10:57 on YouTube. All right, mountain time.
1:11:03 Spread the word. Yeah, please do spread
1:11:05 the word. And if you go into the AI
1:11:07 salon into the irregulars channel, this
1:11:09 graphic and the YouTube graphic, this
1:11:12 one, which is where it's going to stream
1:11:14 live, those are available if you want to
1:11:16 share those on your social. Share them
1:11:18 on your socials and tag me. I'm Kyle
1:11:20 Shannon on LinkedIn. I'm Kyle Shannon on
1:11:23 X. Um, I'm Kyle Shannon on Instagram.
1:11:28 I'm at Kyle Shannon on Facebook or no
1:11:31 Facebook. I'm like Kyle Shannon 96 or
1:11:34 seven or I don't know some [ __ ] [ __ ]
1:11:36 like that. But I don't check Facebook
1:11:38 anyway.
1:11:41 I posted the training
1:11:43 locally EST and my contacts were like
1:11:46 what? Yeah, exactly. It's okay.
1:11:49 If if they want to get their [ __ ]
1:11:51 together with AI, this is I'm telling
1:11:53 you this is going to be a really good
1:11:54 week. Like even if you've got your [ __ ]
1:11:56 together with AI, I think I think going
1:11:59 back and just going, "Okay, here's why
1:12:02 we're here. Here's how this [ __ ] works.
1:12:04 Here's chat GPT. Here's these image
1:12:07 tools. Here's how to apply this [ __ ] in
1:12:08 business. Let's ask about it." Like that
1:12:11 progression in five days, I think, is
1:12:13 going to be quite quite powerful. Razer
1:12:15 wants to see the VO version of you.
1:12:18 Where's the VO version of you?
1:12:21 Oh, this thing. This dude. Let's run
1:12:24 over to VO and take a look. All right,
1:12:27 that thing's still working. How's brand
1:12:30 guidelines coming? Is it done? Oh, no.
1:12:31 That one's still working. Is that one
1:12:33 still going?
1:12:35 Oh, what' we get? Do we have a
1:12:44 My translation reads, year one of our
1:12:46 Lord Probus.
1:12:48 I, Aurelius Ediman, have delivered two
1:12:52 arti of wheat addressed at 400 Dmmy.
1:12:57 This is cool. We're translating Egyptian
1:13:00 documents here, people. Okay. Um,
1:13:07 you already did this. It's on Twitter.
1:13:09 Uh, yeah, it's on Twitter. We can go
1:13:11 look on Twitter.
1:13:14 [Music]
1:13:24 [Music]
1:13:25 There he is.
1:13:38 We could go watch the beginning of the
1:13:40 Theo Vaughn thing where he talks about
1:13:42 babies will be born outside of human
1:13:44 bodies.
1:13:47 Mountain on YouTube. Curious about AI?
1:13:50 Now's the time. Join me starting July
1:13:53 28th at 8:30 p.m. Mountain on YouTube.
1:13:57 Curious about a Now, how I did that,
1:13:59 that's not my voice, obviously. Um, I
1:14:03 just uploaded that graphic into V3 on
1:14:06 flow.google
1:14:09 and uh I said, "Turn this into like a
1:14:12 spokes model, dude. AI, now's the time.
1:14:15 Join
1:14:15 >> and I wrote the script.
1:14:16 >> Join me starting July 28th at 8:30 p.m.
1:14:20 Mountain on YouTube. Curious about AI?
1:14:23 Now's the time. Join me starting July
1:14:26 28th at 8:30 p.m.
1:14:28 >> Great minds, brother. Oh, yeah. Let's
1:14:31 watch that.
1:14:37 Can I get an agent that'll sleep for me?
1:14:43 Mountain on YouTube.
1:14:45 >> Curious about AI? Now's the time. Join
1:14:48 me. Starting July 28th at 8:30 p.m.
1:14:51 Mountain on YouTube. Curious about AI?
1:14:55 Now's the time. Join me. Starting July
1:14:57 28th at 8:30 p.m. Mountain on YouTube.
1:15:01 Curious about AI? Now's the time.
1:15:06 This This is This is like those
1:15:08 infomercials, right? the ones where
1:15:11 you're like, you know, you know when
1:15:13 you're watching a TV show and like the
1:15:15 series ends and then an info infomercial
1:15:18 starts, but you don't realize that
1:15:21 you're watching an infomercial because
1:15:22 you just zombieed out from watching TV
1:15:24 for four hours.
1:15:26 And it's not until the thing repeats
1:15:29 itself like this that you go, "Oh [ __ ]
1:15:32 this is a commercial."
1:15:37 Call 1800 carpets to get your carpet
1:15:40 samples today. Call 1800 carpets to get
1:15:43 your carpet samples today.
1:15:48 That happens to me all the time.
1:15:55 Oh man. While you're sharing your sound,
1:15:57 uh, Pika. Okay.
1:16:00 Uh,
1:16:03 let's go to AI Salon. Hey.
1:16:14 All right. So, Danielle um has made some
1:16:18 videos in uh in PALabs. They've got a
1:16:22 new social video network that's all AI,
1:16:25 which I think is interesting. Sam
1:16:28 Haltman,
1:16:29 you're just a cranky Gen Xer.
1:16:34 That does. It looks like Sam Waltman.
1:16:36 Champy's like, "Uh, I wouldn't [ __ ] with
1:16:38 him. He's a little cranky."
1:16:44 All right, that one's That was weird.
1:16:45 This one's really cool. And then what do
1:16:46 we have here? Oh, agent. Okay, Brandon
1:16:49 put a thing up about agent. So, this is
1:16:51 Oops. I got to copy this. Hang on. Go
1:16:55 back here. Change this.
1:16:59 I'm starting to figure this [ __ ] out.
1:17:01 All right.
1:17:02 My turn.
1:17:27 These animators, they got to train these
1:17:29 models to get they need to understand
1:17:30 rock and roll. when they start ripping
1:17:32 the vocals, you gotta make the face, you
1:17:34 know. Um, but yeah, this is the new Pika
1:17:37 app. It's If you don't know what this
1:17:40 is, think Tik Tok, but all AI effects.
1:17:45 That's what that's what it's going to
1:17:46 be. They're like, instead of having, you
1:17:48 know, rules that you can't use AI, we're
1:17:50 going the other direction. How about you
1:17:52 go all AI? Um, so that's that one. Uh,
1:17:56 super janky. Yeah, it's really fun. Very
1:17:58 cool. Yeah, it's going to be janky for a
1:18:00 while. It's going to be fun. And you
1:18:02 know what? It's going to feel weird
1:18:06 for a month or two and then it's just
1:18:08 going to be normal. All of the all of
1:18:11 this [ __ ] about
1:18:14 we've got to label stuff AI and this AI
1:18:16 and that AI, it's going to start to fade
1:18:19 to the background. There are certain use
1:18:21 cases where we need to label stuff AI.
1:18:24 If you're trying to do a news thing and
1:18:27 you're faking the video, that needs to
1:18:29 be labeled. If you're doing something
1:18:31 creative, it doesn't [ __ ] matter.
1:18:34 It just doesn't. It's just another tool,
1:18:37 right? So, I think we're going to start
1:18:39 to see the whole thing about is it AI or
1:18:41 not AI. I think that's going to drift in
1:18:43 the background except in use cases where
1:18:45 it's actually relevant. In most cases, I
1:18:48 don't think it's relevant. That's my
1:18:50 That's my two cents. That's just my two
1:18:52 cents. I'll tell you what, that's my two
1:18:55 cents. Chad GBT.
1:18:58 Hi, name's Kyle Shannon.
1:19:01 I'm a trainer of Chat GPT. If you
1:19:03 haven't heard of Chat GPT, it's a
1:19:05 fantastic tool.
1:19:08 It's a reflector of all of the knowledge
1:19:10 of all of humanity.
1:19:13 But no, really, it is.
1:19:16 They've taken every document uploaded on
1:19:19 the internet in the past eight decades
1:19:20 and smashed it into a softballsized ball
1:19:23 that you can talk to.
1:19:25 No, seriously, they have.
1:19:28 They've done what's called embedding
1:19:30 where they take these documents and
1:19:31 shatter them into trillions of little
1:19:33 shards that get embedded into some
1:19:36 mathematical space that's thousands of
1:19:39 dimensions.
1:19:40 And then when you type in your prompt,
1:19:42 it creates a probability and reassembles
1:19:44 those shards into some original
1:19:46 document.
1:19:48 No, really.
1:20:02 Ah, unlabelled to load conversation.
1:20:06 Seriously?
1:20:08 Oh, maybe that's because it's still
1:20:10 alive over here.
1:20:13 No,
1:20:16 maybe.
1:20:18 Oh, we got brand guidelines. Hang on,
1:20:21 people.
1:20:24 Let
1:20:27 [Music]
1:20:28 mama.
1:20:32 [Music]
1:20:41 Okay, I'll go look at Theo Von.
1:20:44 Um, we have
1:20:49 we have StoryVine brand guidelines. This
1:20:52 is fascinating.
1:20:54 All right. Do we think this is going to
1:20:56 be any good? It's sort of got the right
1:20:59 colors on the on the This is not This is
1:21:03 nowhere close to our brand sort of
1:21:06 style. And I I don't think this is our
1:21:09 font,
1:21:10 but let's see. Let's see what we got
1:21:13 here.
1:21:15 Brand essence and miss mission.
1:21:18 Use our video guides to structure. Okay.
1:21:24 Well,
1:21:26 sort of
1:21:28 sort of.
1:21:34 Um, guided video orange F2 7421.
1:21:39 Those are close. They're not right
1:21:40 though.
1:21:43 Proxima Nova and Monserat. Those are
1:21:46 correct.
1:21:47 Graphic elements. It sort of pulled
1:21:49 those. Those are not bad.
1:21:52 Imagery and photography. That's
1:21:53 completely wrong. But wait. Authentic
1:21:56 and human. Yes. Candid moments. Yes.
1:21:59 Color harmony. Sure. Context. Include
1:22:02 smartphone frames or video cues to
1:22:05 anchor the narrative. Not true.
1:22:09 Customer testimonial, video card, social
1:22:11 title. These are bad.
1:22:15 All right. Well, it
1:22:19 sort of did something.
1:22:26 I mean, quite frankly, this is better
1:22:28 than I thought it would do.
1:22:30 It It should have been able to pull
1:22:32 those colors from the logo.
1:22:35 Um, it shouldn't have. This is
1:22:38 distorted. This is completely wrong.
1:22:41 That's correct. That's correct. That's
1:22:43 correct. That's wrong. That's wrong.
1:22:45 These are overlapping.
1:22:48 But, you know,
1:22:51 not horrible.
1:22:54 Usable. No.
1:22:58 All right. Let's go. Let's go to
1:23:00 YouTube. Let's go watch. We're gonna go
1:23:02 watch some Theo Von. Oh, wait. I'm not
1:23:04 sharing
1:23:05 I'm not sharing audio here, am I? So,
1:23:07 let's go here.
1:23:22 I just want to watch the beginning of
1:23:24 this again because
1:23:27 I think it's important. Actually,
1:23:31 >> I think I stumbled into something.
1:23:34 [Music]
1:23:35 They're coming after our family.
1:23:41 >> Nobody too. Rated art only in theaters
1:23:42 August 15th.
1:23:44 >> Today's guest is uh well, dude's a
1:23:46 straightup tech lord. Let's be honest.
1:23:52 >> Dude's a straightup tech lord. [ __ ]
1:23:55 Theo.
1:23:56 Oh my god. All right,
1:23:59 we just I just want I just want you to
1:24:01 get the first meaty question here.
1:24:07 He's uh he's one of the leaders, the
1:24:09 world leaders in the development of AI.
1:24:13 >> AI
1:24:15 like he's never said those letters
1:24:17 together before.
1:24:18 >> Um he started open AI which is known for
1:24:22 uh having chat GPT.
1:24:25 Uh, we had a fascinating chat about the
1:24:27 pros and cons, um, the fears and hopes,
1:24:30 everything I could learn about, uh,
1:24:32 about artificial intelligence and where
1:24:34 we're headed. TBD, baby. Today's guest
1:24:38 is Mr. Sam Alman, and I'm very thankful
1:24:41 for his time.
1:24:42 >> Yeah, no [ __ ] This is This is a hell of
1:24:44 a get.
1:24:48 [Music]
1:24:56 place and uh
1:24:57 >> I think it was perfect because
1:24:58 everybody's
1:24:59 >> staff very sweet nice people. Um you
1:25:03 have Thanks for hanging out, man.
1:25:04 >> Absolutely. Thanks for appreciate it. Uh
1:25:07 yeah, I haven't seen you since I fell
1:25:08 out of my chair chair at the
1:25:09 inauguration.
1:25:10 >> That was really like quite a way to meet
1:25:12 you.
1:25:12 >> Yeah, I felt so embarrassed and you were
1:25:14 one of the faces that I looked up and
1:25:15 saw and I was like God. And that was my
1:25:17 first moment like AI build us a better
1:25:19 chair to be honest with you.
1:25:21 >> And you did nothing, right? You were
1:25:22 just sitting there and it just
1:25:23 collapsed. Nothing.
1:25:24 >> I remember that.
1:25:25 >> And it was just so embarrassing. I was
1:25:27 like, "Oh, of all people me." And here I
1:25:29 am in this place. And uh
1:25:30 >> I think it was perfect because
1:25:32 everybody's got to have some sto when
1:25:33 people are like, "Oh, what was the
1:25:34 inauguration?" Like everybody's got to
1:25:35 have some story to tell.
1:25:36 >> Yeah.
1:25:37 >> And that was an incredible story for us
1:25:38 all to tell.
1:25:39 >> That's a good point. I do remember
1:25:41 looking at people for help though. And
1:25:42 oddly your eyes I I was like, "Oh my
1:25:45 god, he could help." You did look like a
1:25:47 beacon of help in the distance.
1:25:48 >> I tried to help.
1:25:48 >> Um, you have a baby. You have a new
1:25:52 >> Yeah.
1:25:52 >> child.
1:25:53 >> It is.
1:25:56 >> Okay, we're going to skip him talking
1:25:57 about his baby. We're going to get to
1:25:59 the follow on question. The best
1:26:01 question in the history of interviewing.
1:26:04 >> So intense.
1:26:05 >> So it's really like almost like a coffee
1:26:07 for your heart or something kind of.
1:26:09 >> I don't even know how to find I've tried
1:26:12 to like come up with analogy to tell
1:26:13 because now I'm like telling everybody
1:26:14 you got to have a lot of kids. It's
1:26:15 really important. Yeah.
1:26:16 >> And I've been looking for an analogy of
1:26:18 what to explain and and I always just
1:26:20 say like I I don't know how to explain
1:26:21 this. It's just it is the best thing
1:26:22 I've ever done by far. I feel like
1:26:26 >> a completely changed person. And I was I
1:26:28 was like thinking the other day like
1:26:32 there used to be all these other like at
1:26:33 this point all I do is work and hang out
1:26:35 with my family. I like I don't
1:26:37 >> I don't like really get to do a lot of
1:26:38 hobbies anymore. It's busy time at work.
1:26:40 >> I don't get to hang out with my friends
1:26:41 that much. Uh
1:26:44 and I and I don't you know there were
1:26:45 like all these things where people tell
1:26:46 you like oh you got to baby come you got
1:26:48 to go you know take that spontaneous
1:26:49 international trip cuz you're not going
1:26:50 to be doing that again.
1:26:51 >> Yes in like a nice
1:26:53 >> okay here's the question
1:26:54 >> about that
1:26:57 process more
1:26:59 but they're like intrigued and stuff
1:27:01 they start to like smile or process
1:27:03 more. I don't know how you guys say it
1:27:05 but um
1:27:05 >> yeah he's totally like turned on though
1:27:07 like really aware understands things.
1:27:08 It's super cool. I have a thought
1:27:10 sometimes that this will be one of the
1:27:12 last like maybe 40 years that we
1:27:15 conceive children in the body. Did you
1:27:17 have any thoughts about that?
1:27:20 >> I've definitely heard a lot of people
1:27:22 say that. Um
1:27:25 I haven't thought about it hard myself,
1:27:27 but yeah, I guess it does make sense.
1:27:29 Like
1:27:31 >> I guess that does make sense.
1:27:32 >> Like God, you were in your mom's butt.
1:27:34 It's crazy, you know, you pervert or
1:27:36 whatever. Like I think in the future
1:27:37 people will be it'll be kind of done
1:27:39 like in a
1:27:40 >> in a vet or something.
1:27:41 >> Yeah.
1:27:44 >> In a vet.
1:27:46 >> In like a nice vet. You can go see it on
1:27:48 the weekends or whatever. And like
1:27:49 >> doesn't that just feel like off to you?
1:27:51 Like I can totally intellectually like
1:27:53 understand that that may be the better
1:27:55 way to do it.
1:27:56 >> Oh yeah. It feels way off to me. I was
1:27:57 trying to I thought you would like it.
1:27:59 >> Okay. That's really important. I thought
1:28:02 you would like it.
1:28:05 I think Theo Vaughn absolutely here
1:28:07 represents everyone that has that is
1:28:11 terrified of AI that hasn't played with
1:28:13 it.
1:28:15 The minute you play with it and realize
1:28:17 that it's an amplifier of your ideas,
1:28:20 it's this incredibly empowering thing.
1:28:22 But when you sit on the outside of it,
1:28:25 he's like Theo Vaughn constructed this
1:28:28 whole thing. the le the next 40 years
1:28:30 are going to be the last time we
1:28:31 conceive babies in human bodies. Sam
1:28:35 Hton was like, "Huh?"
1:28:38 He goes, "Doesn't that feel wrong to
1:28:40 you, Theo?" And Theo's like, "Oh, yeah.
1:28:42 It feels way wrong to me. I thought
1:28:45 you'd like that." Right.
1:28:48 Anyway, it's it's
1:28:52 worth watching the whole interview just
1:28:55 because I think that's
1:28:59 I think there's a lot of people that
1:29:00 have deeply deeply irrational fears
1:29:05 about AI. I think there are some fears
1:29:07 about AI that are justified. There's a
1:29:10 lot that are just in this weird surreal
1:29:13 place because people haven't played with
1:29:16 it. And so anyway, that's the mission of
1:29:18 this channel. Play with this [ __ ] you
1:29:22 pervert. I know, right?
1:29:26 Oh my god. Um,
1:29:37 Lovable Agent dropped today. Did it? I
1:29:41 went and looked at it yesterday.
1:29:44 Also,
1:29:49 fastest run rate to hund00 million.
1:29:53 Yeah.
1:29:56 Yeah, I saw that
1:30:04 tabs if I Yeah, I'm showing something
1:30:06 right now.
1:30:17 So, I had it I asked it to make an
1:30:20 arcade. Here's what I don't know,
1:30:22 Brandon, is
1:30:25 I don't know the difference between
1:30:29 lovable just lovable and lovable agent.
1:30:36 Does anybody know agent does what is
1:30:38 this? Like am I looking at agent right
1:30:40 now? And I know I
1:30:42 >> I am. So So how is this different than
1:30:45 what it was before?
1:30:48 >> Yeah. So the the key differential here
1:30:50 is that lovable agent has the ability to
1:30:54 think within itself and um throttle your
1:30:58 credits. So it it looks at different
1:31:00 options uh and different ways of
1:31:03 executing what you're asking to do and
1:31:05 then it will deploy
1:31:08 on a more efficient path. So it's it's
1:31:10 taking time to think about before it
1:31:12 just burns through your credits doing
1:31:14 the wrong thing.
1:31:15 >> So So it's it's essentially transparent
1:31:17 to the end user like
1:31:19 >> Yes.
1:31:20 >> Like like the only thing an end user
1:31:21 will notice is that they don't run out
1:31:24 of credit so fast. They might notice it
1:31:26 talking. So, it sounds like it's
1:31:28 switching back and forth between chat
1:31:29 mode, which doesn't cost you credits,
1:31:32 and programming mode. Is that is that
1:31:34 Yes.
1:31:34 >> a fair assessment? Okay. All right.
1:31:36 Cool. So, if you have not played with
1:31:38 Lovable, um you should go play with it.
1:31:41 I also saw today um right before I came
1:31:44 on
1:31:46 and if you don't know what Lovable is,
1:31:48 Lovable is is one of the the top what
1:31:52 they call vibe coding apps where you can
1:31:55 just talk an app into existence. So if I
1:31:58 go into my um
1:32:03 area, how do I get in here? How do I get
1:32:06 Oh, workspace.
1:32:14 How do I get in there?
1:32:30 Oh, scroll down. When in doubt, scroll
1:32:33 down.
1:32:45 I think. Where where did I do my my
1:32:49 asteroids game
1:32:53 here? This one.
1:33:02 This was a version of Asteroids I made
1:33:04 in in lovable with a single shot. So a
1:33:08 single shot basically just means one
1:33:09 prompt.
1:33:23 [Music]
1:33:27 Ah!
1:33:34 [Music]
1:33:43 Ah!
1:33:52 [Music]
1:33:59 Okay, I'm sucking. The game's good, but
1:34:02 I suck.
1:34:04 [Music]
1:34:06 But it, you know what? It's got a lot of
1:34:08 the same adrenaline as when I was a kid.
1:34:11 Anyway, all right. That's lovable. And
1:34:13 then I saw today a new one of these
1:34:16 things launched
1:34:18 that claims to be better. It's called
1:34:20 Mocha.
1:34:22 Get get mocha.com.
1:34:26 I don't know if you all have seen this.
1:34:28 I haven't. So,
1:34:30 >> we're not going to see anything unless
1:34:31 you change your tabs.
1:34:31 >> Yeah, I know. I'm going to change it.
1:34:33 Calm down. Calm down. I got
1:34:35 >> the uh also pointing out Vicki called
1:34:38 out that chatting with Lovable in the
1:34:40 chat does cost you credits. So, be
1:34:42 careful with that. Um, one of the things
1:34:44 that I've found is uh helpful is
1:34:49 chatting about the project in another
1:34:51 tool like Gemini or Chatg GPT and then
1:34:55 doing your front work there and then
1:34:58 taking the prompt into chat into
1:35:00 Lovables to save your credits. Uh,
1:35:02 Lovable co-founder was on the cover of
1:35:04 Forbes Daily today, though. So,
1:35:06 >> Wow, nice. All right, cool. And they're
1:35:09 French, right? Isn't it a French
1:35:10 company? Lovable. I know that you're
1:35:12 somewhere. Yeah. Um, okay. So, this is a
1:35:16 this is a new tool tool called um Mocha
1:35:20 and I I Mocha 1.0 is here. The first
1:35:23 complete app builder, read the blog
1:35:24 post. So basically what they're claiming
1:35:26 here, okay,
1:35:29 if you're not a coder, if you use
1:35:31 something like cursor um or replet,
1:35:35 it writes all this code stuff and then
1:35:37 it it incorporates with things like
1:35:41 GitHub where you can store code and then
1:35:43 you have to set up your databases and
1:35:46 your authentication and things like
1:35:47 that. And you still kind of have to know
1:35:50 programming, right? It does the coding
1:35:52 for you, but you need to know sort of
1:35:54 application management.
1:35:58 Lovable was the first one that says,
1:35:59 "Hey, I can just build you a a web app
1:36:03 that you can then launch and share." You
1:36:05 could also do that in Claude and you can
1:36:07 also do that in chat GPT. But lovable is
1:36:09 kind of this standalone. We're going to
1:36:11 make you web applications. And then they
1:36:14 made it very easy to incorporate a
1:36:16 database with superbase and
1:36:19 authentication.
1:36:21 Basically making it easy to set up. What
1:36:24 this Mocha thing claims is that it's
1:36:26 going to do all of the setup completely
1:36:29 transparently.
1:36:31 I'll believe it when I see it. But so
1:36:33 let's let's just you know what I'm going
1:36:36 to do? Let me go back to lovable for a
1:36:39 second.
1:36:41 I'm gonna go get my original prompt from
1:36:44 this Asteroids game.
1:36:47 Make me an Asteroids clone very close to
1:36:49 the original arcade game and make the
1:36:51 physics great and the look and feel
1:36:54 great and the sounds great and
1:36:55 playability awesome. That's not very
1:36:57 good English, but it sure beats having
1:37:01 to type that again. Okay, there's our
1:37:04 prompt. Um, and I don't know even if I
1:37:07 have enough credits to do this. Oh, I
1:37:10 got to sign in. Hang on.
1:37:14 Maybe he should clear his throat before
1:37:16 he starts giving a lecture on how to use
1:37:18 artificial intelligence.
1:37:22 I thought this was some sort of
1:37:24 professional situation. Okay. Oh,
1:37:28 can I just say something that's actually
1:37:30 impressive here?
1:37:33 So before
1:37:35 we were out on the web page where we
1:37:39 hadn't signed in and I and I typed in a
1:37:42 prompt and then I realized, oh, I've got
1:37:45 to sign in. And so I signed in and it
1:37:47 took me to Google authentication
1:37:49 and then it brought me back to a page
1:37:51 that said, okay, you're signed in. And
1:37:53 then I said, make a new app. And it
1:37:55 remembered my prompt.
1:37:58 For those of you here who are engineers
1:38:02 and and you're coders, if your app isn't
1:38:05 that smart, it pisses Gen Xers off.
1:38:10 Okay? You don't want us angry.
1:38:14 We're bitter and we enjoy it.
1:38:18 This is this is good. This is promising.
1:38:21 Okay, here we go. So, you just saw the
1:38:23 Asteroids game I just played. This was
1:38:25 the prompt that generated it. So, let's
1:38:27 see how this mocha does.
1:38:30 Working on it. All right.
1:38:33 [Music]
1:38:44 Green Acres is the place to bay. This is
1:38:47 Mocha. Get Mocha mocha.com.
1:38:52 I'll create a classic Asteroids game
1:38:54 with authentic physics, retro vector
1:38:57 graphics, engaging gameplay. The game
1:38:59 will feature smooth ship controls with
1:39:02 momentum based movement. Yes, realistic
1:39:05 asteroid physics with proper collision
1:39:07 detection and classic sound effects.
1:39:10 I'll focus on gathering the original's
1:39:12 addictive gameplay loop with multiple
1:39:15 asteroid sizes, score systems, and
1:39:18 progressive difficulty. That's good.
1:39:21 All right, writing use game engine.ts.
1:39:26 Read our Jet docs. Join our community.
1:39:28 So, this is yet another vibe coder. Um,
1:39:31 I don't know if it's any good at all. I
1:39:32 just saw it before I came live and
1:39:34 that's it.
1:39:36 That's it, people. That's all I got
1:39:38 going. That's all I got going for me. I
1:39:40 got really nothing. So, tomorrow
1:39:42 tomorrow's Friday. So, tomorrow night we
1:39:44 got Friday night date night.
1:39:47 Microsoft also doing this. Oh, in
1:39:49 Copilot. Yeah. Yeah. Microsoft just
1:39:52 dropped a vibe coding app. Here's Mocha.
1:39:54 I think there's like 10 of these things
1:39:56 now. Um, Lovable's the most usable one.
1:40:00 We'll see how this thing does. Maybe it
1:40:01 maybe it'll inspire. Um, if you go to
1:40:06 GitHub, GitHub Spark is the new
1:40:08 Microsoft offering.
1:40:11 Um,
1:40:13 tomorrow, Friday night date night. So,
1:40:16 tomorrow is not the first Friday. So, it
1:40:19 should be 8:00 tomorrow night. And then
1:40:22 tomorrow at 11:00 a.m. Mountain time,
1:40:26 you've got your uh what's it called? Um
1:40:31 AI office hours.
1:40:34 So, on LinkedIn, if you go to my account
1:40:36 on LinkedIn, I'm I'm Kyle Shannon on
1:40:38 LinkedIn.
1:40:40 And just go to events in my profile and
1:40:43 you'll see AI office hours. Just copy
1:40:46 that. Something went wrong. Please try
1:40:47 again.
1:40:50 Uh, okay.
1:40:56 Should I say continue?
1:41:03 Initializing.
1:41:08 Um, if you haven't been to an office
1:41:11 hours before, it's it's a really smart
1:41:13 group of people. If you're trying to get
1:41:15 your head around this AI stuff, get your
1:41:18 ass in community. Join the AI salon.
1:41:20 Come to AI office hours. Turn on your
1:41:23 [ __ ] camera
1:41:26 and let the good people see who you are.
1:41:28 Connect with people. Tell people what
1:41:30 you're passionate about. But I'm eating.
1:41:34 Exactly.
1:41:38 Uh, have you used Oh, is that's so
1:41:41 funny. Real real Gunny retired. Um, have
1:41:44 you used wind surf? I have not. I
1:41:46 thought I thought I thought that said,
1:41:47 "Have you used Windows?"
1:41:50 Mocha crashed. Yeah, it did. I see this
1:41:53 is a fresh starter app. Just loading a
1:41:55 spinner. Yeah, that crashed. That
1:41:57 crashed up hard, people. Let's Let's
1:42:00 restart the prompt.
1:42:04 Create app.
1:42:05 Paste. Go.
1:42:10 All right, we'll we'll let this go for a
1:42:12 little bit.
1:42:14 I don't have a LinkedIn account. You
1:42:15 don't have a LinkedIn account.
1:42:19 Um,
1:42:25 you know what I'm going to have to do?
1:42:26 I'll make a uh I'll make one of those
1:42:29 slides. Yeah, something went wrong
1:42:31 again. I don't know. Oh, you know what?
1:42:32 I bet went wrong
1:42:35 because I said it was an Asteroids
1:42:36 clone. I bet it's got some copyright
1:42:40 [ __ ] in it. that that because I used
1:42:42 the word asteroids, they're like, "Oh,
1:42:44 that's an Atari copyright, so we can't
1:42:46 we can't do that." I bet that's wrong
1:42:48 with this. All right, whatever. You can
1:42:50 go play with Mocha if you want. Maybe
1:42:51 it's a free way to do some vibe coding.
1:42:54 Um, Hedra now has real-time avatars,
1:42:57 um,
1:42:59 which I'm looking at right now, but they
1:43:01 were pretty bad, so not worth looking
1:43:02 at.
1:43:04 Um,
1:43:06 in the life hacks club in the AI salon,
1:43:09 they've been playing with lovable. So,
1:43:12 yeah, they're working on a project over
1:43:13 there. I do not. Should I Should you
1:43:15 have LinkedIn, Cam Ken? I don't know. I
1:43:23 LinkedIn definitely has
1:43:28 professionals,
1:43:31 but primarily what it is is a place for
1:43:34 salespeople to try to find other people
1:43:36 to buy their [ __ ] And it's pretty bad.
1:43:39 But the AI office hours is pretty good.
1:43:44 Yeah, the office hours is a Google Meet.
1:43:46 Oh, you know what I'll do? Um, hang on.
1:43:48 I know. I know what I can do. Cam Ken,
1:43:50 I'm gonna throw this in irregulars.
1:43:54 I'm going to throw the the Google Meet
1:43:55 URL for the for the meeting tomorrow
1:44:00 and then I'll make a proper
1:44:09 You are called out. You gave me the
1:44:11 idea.
1:44:15 Okay. events.
1:44:20 Office hours with Kyle Shannon. Hi, I'm
1:44:23 Kyle Shannon. Come to office hours with
1:44:26 me, Kyle Shannon.
1:44:30 All right. So, I'm going to go to
1:44:33 [Music]
1:44:42 So, I am going to go to
1:44:47 the Irregulars channel. Hello
1:44:50 irregulars.
1:44:51 Good day to you sir. Good day uh sir.
1:44:53 Yes. Carrying on rather cheerio. Um pip
1:44:57 pip. Yes. I was wondering sir with the
1:45:00 AI learning lab and you positioning
1:45:02 yourself in the professorial position. I
1:45:05 was I was wondering sir by chance pray
1:45:07 tell what your qualifications might be.
1:45:12 My qualifications?
1:45:13 Yes. I was just uh curious
1:45:16 if there were any what they might be.
1:45:20 I have none. I have none, sir. Why? Why
1:45:23 would you call my integrity into
1:45:25 question, sir?
1:45:27 Is this some sort of challenge? All
1:45:29 right. Um boom. There's that.
1:45:33 And then I'm gonna go uh
1:45:40 here is here is the
1:45:47 URL for AI office hours on LinkedIn
1:45:57 and then I'll put in parenthesis but you
1:45:59 can also just join the Google meet here.
1:46:06 Colon close parentheses. There it is.
1:46:10 And then I'll put 11 a.m. MT on Fridays.
1:46:17 Always the
1:46:20 same
1:46:21 URL.
1:46:24 Put it in your calendar exclamation
1:46:29 point. All right. and post.
1:46:33 I know we don't need to notify the world
1:46:34 about that.
1:46:37 Apparently Musk heard about Pika Elon
1:46:39 Musk on X. We're bringing back Vine but
1:46:43 in AI form. I guess my company Storyvine
1:46:46 should re reopen my trademark dispute
1:46:50 with from 13 years ago
1:46:56 where we we were we were I think they
1:46:59 ceased and assisted us that like you're
1:47:02 a video app and you should go away and
1:47:05 then we responded and said we were here
1:47:07 before you nanny nanny boooo and then
1:47:10 they said no we were here before you and
1:47:13 then we proved it and then they said,
1:47:16 "Oh,
1:47:17 well,
1:47:19 we're Twitter, so we're going to keep
1:47:21 using it. So, if you want to come after
1:47:23 us, go [ __ ] yourself."
1:47:26 So, I guess I could open up that can of
1:47:28 worms again, just as a hobby.
1:47:35 Oh, man. Good night, Quinn. Thank you,
1:47:36 everybody. All right, so that's that.
1:47:39 There you go. Happy that. So, tomorrow,
1:47:42 office hours at 11:00 a.m. I think you
1:47:46 got the afternoon free. We got Friday
1:47:47 night date night, 8:00 PM tomorrow night
1:47:49 here. We're prepping for next week, the
1:47:53 back to basics AI crash course. Five
1:47:55 days. Get your [ __ ] together. Bring your
1:47:58 friends. Bring your friends that don't
1:48:00 want to hear about AI,
1:48:03 right? Let's have some Kevin Mallister
1:48:05 moments. Okay. Okay. Deal. All right,
1:48:10 everybody. Beautiful DQ Blizzard. Thank
1:48:12 you, Kyle. See you tomorrow. All right.
1:48:13 Peace out, everyone.
1:48:15 Have a good evening.