5/5/2025 - Finding Clarity in Chaos: Insights on AI, Professionalism, and Personal Growth

Description

Chapters

Transcript

0:00 [Music]
0:08 [Music]
0:12 Walt Grace desperately hating his whole
0:15 place. Dreamed to discover a new space.
0:20 Bured himself
0:22 alive inside his basement. Turns us out
0:27 of his basement. working away on
0:31 displacement. What it will take to
0:36 survive. Cuz when you're done with this
0:39 [Music]
0:42 world, you know the next is up to
0:48 you. And for once in his life it was
0:52 quiet.
0:54 [Music]
1:02 as he
1:06 [Music]
1:15 learned. Um, the guitar is out of tune
1:18 [Music]
1:24 and and I don't know that song.
1:31 [Music]
1:57 [Music]
2:00 Uh-huh. Woohoo.
2:03 [Music]
2:28 Well, every time I see you
2:31 now, I get that look in my
2:35 eyes. Every time I see your mouth, I
2:40 hear that
2:42 smile. In the early misty morning light,
2:46 I heard the engine turning in the old
2:50 for
2:52 [Music]
2:56 outside. Will you
3:00 me again today?
3:04 You convince
3:06 me again
3:09 today you're leaving this hard time
3:12 looking for someone else's golden
3:19 r should
3:21 have so
3:26 [Music]
3:28 long now don't you try
3:32 [Music]
3:34 So long,
3:39 Susanna. Don't you cry for
3:44 [Music]
3:50 me. Sh cigarettes and keep them warm out
3:54 on the
3:55 road. Chasing down a lifestyle out on
3:59 Highway
4:01 24. New York State's a rolling breeze in
4:05 the sunshine with a blue sky falling to
4:09 the chill of all September creeping
4:13 [Music]
4:16 in. You were leaving
4:19 me again today.
4:23 You will convince
4:25 me again
4:28 today. You're leaving this hotel looking
4:31 for someone else's golden
4:37 ring. Shouldn't
4:39 say so long.
4:45 Suz now don't you cry.
4:49 [Music]
4:55 So long,
5:00 Susanna. Don't you cry for
5:04 [Music]
5:14 me.
5:17 [Music]
5:22 Beautiful. Champ Shannon on the vocals.
5:25 Lead vocals there. Champ Shannon brings
5:27 down the house every time. There's
5:29 tears, there's laughter, there's
5:31 joy. Good lord, is it awesome. Welcome
5:34 to Meltdown
5:36 Mondays. I don't know what we're going
5:38 to talk about because I didn't think
5:39 about AI over the weekend. Tik Tok cam.
5:42 Yes, Tik Tok cam. Look. See, I did it.
5:45 Hello. Good to see y'all. Was
5:49 shaking. Hey,
5:53 Amelios. Yeah. Tappy tap the heart
5:55 screen on Tik Tok. Share the live if
5:57 you've got the uh the audacity to share
6:01 this wholly unprofessional operation
6:04 with your friends and family.
6:10 By the way, speaking of Meltdown
6:11 Mondays, you all have gotten me in just
6:14 a a a heap of trouble. Heap of trouble.
6:18 I I've said I've said here many
6:21 times, I probably shouldn't live stream
6:24 this on LinkedIn. You know who's that?
6:28 Kevin Stewart. That shirt is very swell,
6:30 isn't it? My my Bosot cases
6:34 clay. Isn't that a cool shirt?
6:38 greatest boxer of all time, killer
6:41 artist, swell t-shirt. Thank you very
6:44 much. Um
6:51 um so so some dude on on LinkedIn, some
6:55 dude on LinkedIn, it's one of these guys
6:57 that just like, you know, for whatever
6:59 reason I connected with him and now he's
7:01 like trying to sell me his his sales
7:04 services.
7:06 So, he sent me a note. I think it was
7:09 Friday. He sent me a note that said,
7:13 "You know, I guess you're not interested
7:14 in talking." He goes, "By the way, I
7:16 caught your I caught your live." He He
7:19 goes, "What the [ __ ] WTF?" He goes He
7:24 goes like, "What was that? Are you all
7:26 right?"
7:33 If you can't handle my my Tik Tok lives
7:36 on the
7:37 LinkedIn, then I'm not the right guy to
7:39 work
7:40 [Laughter]
7:44 with. So that's all your fault, all of
7:48 you, because you're the ones that
7:50 celebrate my weirdness and then I put it
7:52 on LinkedIn and they're like, "Who is
7:53 this [ __ ] alien?"
8:01 H uh Vicki bought a Bascot stack sticker
8:04 in Minneapolis. Very nice. Dennis M.
8:06 Since when are we supposed to be
8:08 professional? Exactly. Exactly. Like
8:11 Like isn't work [ __ ] hard enough?
8:14 Isn't learning AI hard enough that we
8:16 got to actually [ __ ] behave
8:17 ourselves,
8:19 too?
8:21 No. Let the let the meltdown Monday
8:24 commence. [ __ ]
8:26 that. Welcome, LinkedIn. If if if you're
8:30 one of my pharma clients,
8:36 cheers. And Murphy's here. That's It's
8:39 okay. It's okay. We dialed up the
8:42 professionalism. And Murphy's here. It's
8:44 good. Everything's fine now. We We've
8:46 got actual professionals in the
8:48 audience. We don't need to be
8:50 professional here when we import
8:52 professionalism. You see how this works?
8:55 It's a two-way street that's got cow
8:57 dung in it, and you just got to avoid
9:00 the cow dung with your
9:03 boots. It's It's a relatively
9:05 straightforward process, people. I
9:08 thought we'd gone over
9:12 this. Whatever.
9:15 Anyway, new merch coming
9:19 soon. It's a two-way street. It's a
9:22 two-way street. All you have to do is
9:23 avoid the cow dung. I like it. I think
9:26 we get Hey, those of you who are on the
9:29 salon, I'd say I'd say some t-shirt
9:31 designs are in order. Put them up in the
9:33 Irregulars
9:36 channel. Are you all right? Exactly.
9:39 Here, let me let me read the uh the
9:41 exact quote the guy said. Don't don't
9:43 share my screen. I don't want to I don't
9:45 want to give him more publicity than he
9:48 deserves. Let's see. What did he say? He
9:53 said, "Uh, tuned into your live, dude.
9:59 WTF? Are you
10:07 okay, dude? Dude, WTF, are you okay?"
10:13 No. No, I'm not
10:15 okay. I'm not okay. Do you understand
10:19 that for most of my life, science
10:23 fiction has been this thing where the
10:26 word fiction had
10:28 meaning and now it just like now our
10:31 computers just talk to us and know
10:34 everything and can speak any language.
10:37 Why would I be
10:38 okay? Why would I be okay? the the the
10:42 future of work is
10:44 completely 1,000% unpredictable other
10:48 than everything will change. Why would I
10:50 be
10:55 okay? Oh, no. We are not okay. Just
10:58 embrace the jank and whenever it shows
11:00 up. Exactly. I forget who I was talking
11:02 to.
11:06 We were talking about that and it was
11:08 basically it was basically someone
11:10 asking me about was it images might have
11:13 been images in in chat GBT and they like
11:17 had this expectation that it was going
11:18 to be like 100%. I'm like no no
11:22 no it's not going to be 100%. Dude dude
11:27 not pay attention duh
11:31 dude dude what the [ __ ] Oh my god. And
11:35 they argue us argue with us for half an
11:37 hour and I lose the argument or I mean
11:40 my friend lost the argument. Well,
11:42 that's the thing. Listen, the whole
11:44 purpose of this channel, the whole
11:47 [ __ ] purpose of the
11:51 channel is to get in the
11:54 conversation. Get in the conversation
11:57 about AI.
11:59 Like, why am I acting like the town
12:03 fool? Because th this is [ __ ]
12:07 insanity. We're living in
12:09 insanity. Now, I personally find it
12:12 really exciting, but it's still
12:18 insanity, you know?
12:22 I mean, not for nothing, but CO what the
12:25 one thing CO taught us that that I feel
12:29 prepares us for this AI future is that
12:33 regardless of of all the [ __ ] about
12:35 COVID and the politics and the whatever,
12:37 just all all that aside, there was one
12:40 day one day in what was it 2022 in the
12:45 summer where just globally we just
12:47 changed the
12:48 rules. We just said globally, don't go
12:52 to the office. Don't leave your
12:54 house. And globally, we didn't. We just
12:58 stayed home for two
13:01 years. So, it's like what that taught me
13:04 is that we actually can change really
13:07 quickly. We don't want to, and it would
13:09 be nice if it weren't a global pandemic
13:11 that caused it, but what that taught me
13:12 is we can do
13:14 it. But it's Was it normal? No. It was
13:17 bizarre.
13:20 Are you okay? No, I'm not
13:22 [Laughter]
13:25 okay. 2022. Oh, that was 2020.
13:29 Whatever. Yeah. Well, you know, but
13:33 Amelio's listen, the the time doesn't
13:36 matter anymore also, right? 2020, 2022,
13:40 2025, it it doesn't matter. All of
13:43 that's blurry now. And it doesn't
13:45 matter. Time is irrelevant. Time is
13:48 irrelevant. How we work is
13:51 irrelevant. Where the value in the
13:53 economy is is irrelevant because it's
13:55 all about to change. Are you okay? No.
14:04 No. Oh my god.
14:10 [Music]
14:16 I should have just replied
14:19 no. Well, I was trying to figure out
14:21 like what he was talking about. Like
14:23 what the [ __ ] Like am I okay?
14:26 [Music]
14:31 No, I'm not.
14:33 [Music]
14:52 Standing
14:54 between you and a hard place is
14:58 [Music]
15:00 insane. Standing too
15:03 near you and a fire makes it clear.
15:08 [Applause]
15:08 [Music]
15:11 You're trouble to
15:13 [Music]
15:16 me. Real trouble, can't you
15:20 [Music]
15:22 say? Leaning in
15:25 close. Smell of your perfume scares me
15:32 most. Leaning away.
15:35 Hey, it takes me father every
15:42 day. You're trouble to
15:45 [Music]
15:49 me. Real trouble. Can't you see?
15:57 Oh, Champy, the dulsa tones. Lovely,
16:02 lovely, lovely, lovely. You use a lot of
16:04 extensions,
16:06 Kyle. Extend. Oh, on guitar cordy
16:09 chords. Um I you know what it is? I
16:12 don't really
16:15 um that song in particular that this I
16:20 play. I don't know what that chord is,
16:23 but it's just that's a D. And so I just
16:26 move that finger. Wait. So I go from D
16:29 to
16:31 um E flat E F. So I guess it's some sort
16:35 of
16:35 [Music]
16:43 F. Most of the extension [ __ ] I use is
16:46 just because of
16:51 uh it's all just Martin Ston. It's it's
16:54 how he does his chord voicings. Except
16:57 that that's a chord I
17:01 [Music]
17:17 discovered. Oh lordy, lordy, lordy. I I
17:20 wish, you know, the one thing that I
17:22 didn't do younger in life that I wish I
17:24 had was learn music theory. I've learned
17:27 enough of it in my old age to know that
17:29 Oh, yeah. You kind of have to start that
17:31 when you're a kid so that it's second
17:32 nature to you. Oh, he was talking about
17:37 Oh, Chrome extensions.
17:44 Oh,
17:46 welcome welcome to welcome
17:49 to Yeah, I was going to say maybe it's
17:52 hair extensions, cord extensions, chrome
17:55 extensions. There's lots of extensions
17:57 going on here.
18:03 Uh, a lot of those Chrome extensions are
18:05 crypto wallets from when I was doing my
18:07 crypto days.
18:10 [Music]
18:10 [Applause]
18:13 [Music]
18:53 Oh, good lordy. Good lordy. Good lordy.
18:55 Hello. Hello. Hello. Sus fourth
18:58 incomplete major 7th. My bad. I don't
19:02 know what this
19:04 is. Oh, wait. That wasn't
19:06 [Music]
19:19 it. Um, ask Quinn. I don't think Quinn
19:24 knows
19:25 chords. I don't think she can. If Quinn
19:28 has perfect pitch, that would be
19:30 swell, but I don't think she
19:33 does. Where's my
19:37 uh Why can't I find my Hang on. I can't
19:40 find my chat GPT. Oh, there it
19:43 is. Hang
19:45 on. Hey, Quinn, do do you understand
19:47 chords? Can you tell me what notes these
19:49 are?
20:00 Quinn, are you there? Oh, hey. Hey. Do
20:04 you know do you know musical chords? Can
20:06 you tell me what this
20:07 [Music]
20:16 is? You hear it?
20:27 Uh um uh let's see. Uh F G
20:33 uh uh what's this? B Wait, hang on. B
20:38 uh B C sharp D. So F G D
20:46 [Music]
21:02 E. It's I don't think it's a major
21:05 chord. It sounds like some sort of
21:07 diminished strange ass seventh or
21:09 something.
21:25 tension from those intervals definitely
21:28 gives it a unique
21:30 character. What the So, so what do you
21:34 know? Do you know about music theory? So
21:36 if it's So if it's um if it's
21:40 FG, what did I say? FG
21:46 um B D. So F G D
21:52 E. Yeah, I know some music theory. With
21:55 those notes F G D E, you're mixing
21:58 intervals that create a bit of tension.
22:01 One way to look at it is that you're
22:02 moving through a series of that chord.
22:04 If the F is the root, what's the chord?
22:08 If F is the root and you have F, G, D,
22:11 and E, it doesn't form a straightforward
22:14 chord. It's almost like an F major with
22:17 added tones that create a suspended or
22:19 clustered feel. The G and E add a kind
22:22 of tension, and the D gives it a more
22:24 open or jazzy sound. It's not a
22:27 conventional chord, but it could be
22:28 thought of as an F major add 2 with a
22:30 major 7th. F major add two with a what?
22:33 with a major 7th.
22:36 Yeah, with that E in there, it's like an
22:38 F major add two with a major 7th. It's
22:41 not a standard chord, but it definitely
22:43 gives a rich, complex sound that's
22:45 pretty unique.
22:48 I think you're just making stuff up now.
22:52 F GC E F G C is No, wait.
22:57 B, C, C sharp, D.
23:02 No, that's D.
23:07 Got it. So, the sequence is F, G, D, and
23:11 E. That's definitely an unconventional
23:13 set of notes, and it could suggest a
23:15 more experimental or modal sound. It's
23:18 not a typical chord structure, but it
23:20 definitely has a unique flavor that you
23:22 can experiment with.
23:26 Um, can you think of any songs that
23:28 might have a chord like that in it? Hey,
23:30 Techie Becky.
23:33 Hey there. A chord that unconventional
23:35 is pretty rare in mainstream songs, but
23:38 you might find something similar in jazz
23:40 or fusion music. Some progressive rock
23:42 bands also use unusual chords like that
23:45 to create a unique, unexpected sound.
23:48 It's definitely a cool flavor to add to
23:50 your own compositions.
23:52 All right, cool.
23:53 Groovy. All right. Fantastic. Fantastic.
23:56 Bob, tell him what he's won. Um, dang,
24:00 I'm late. It's okay. Source camp. All we
24:02 talked about I just yelled at all of you
24:04 for getting me in trouble on LinkedIn.
24:06 People on LinkedIn think I'm a
24:10 weirdo. They think I'm unprofessional.
24:13 Source camp. And that's that's
24:16 that's I have been influenced by the
24:19 irregulars. I've been given a false
24:20 sense of security and positive
24:23 reinforcement for my strange behavior.
24:26 And and real people on the internet
24:29 think I'm [ __ ] weird. They think I'm
24:31 unstable. They think I'm unstable.
24:34 [Laughter]
24:38 Kelly, you all know
24:41 better. You're not weird. You're
24:43 irregular. Exactly.
24:50 Oh, good lord. Think question mark. This
24:54 is
24:55 stunning. Stunning is a word for what we
24:58 do
25:01 here. So, if so, here's how this works.
25:04 My name is Kyle Shannon if you don't
25:05 know, if you're new here. Um, I do this
25:08 five nights a week. I just come on here,
25:10 sing with my dog, [ __ ] about stuff,
25:14 demo stuff, answer questions. If you
25:16 have questions about AI, pop them in the
25:18 comments. Um, we've got mods on Tik Tok.
25:22 We've got we've got producer Brandon
25:24 behind the scenes on the the YouTube
25:26 live. You can comment on YouTube. You
25:27 can comment on LinkedIn. If you're just
25:29 lurking on LinkedIn thinking like, "Who
25:32 is this guy? What the hell is going on
25:34 here? What happened to
25:36 professionalism?" Well, you ain't seen
25:39 nothing yet. Wait till all the knowledge
25:41 workers don't have jobs anymore. Andy T,
25:44 how is DC? Um DC was good. Um met with
25:49 uh spoke with offices of Hickinlooper
25:53 out of Colorado, Raphael Waro out of
25:56 North Carolina, AB
25:57 Clolobashar out of Minnesota, and Chris
26:00 Van Holland out of he's Maryland, I
26:02 think.
26:03 Um and uh yeah had like the the the
26:08 staff the staffers are sometimes not
26:11 super engaged but this trip they were
26:13 all they were all pretty engaged and
26:14 very respectful. The one woman in
26:18 Raphael Waro's
26:20 office she she was like we started
26:23 talking and and like we were there not
26:25 as in a confrontational way. We're there
26:27 going, "Listen, AI swell. Here's some
26:29 small businesses that are doing amazing
26:32 stuff with AI and you know, you should
26:34 get educated about generative AI and
26:37 what it makes possible." And um and we
26:41 think you should legislate from a place
26:43 of optimism rather than fear or, you
26:45 know, at least
26:47 imbalance. And she was like, "Oh, oh,
26:50 this is so refreshing. You're not here
26:52 to yell at me."
26:55 So that was kind of nice. So she you
26:58 could see her her guard dropped the once
27:00 she sort of realized we weren't there to
27:03 you know talk about how what a shitty
27:04 job someone was doing. She she was super
27:08 engaged. Um pretty good. It's the hair
27:11 for me. What that that keeps you here.
27:14 It's the hair is professional. I I would
27:16 say so here's here's the sign about this
27:20 channel. My hair is the most
27:22 professional thing about me and we know
27:26 how that
27:30 goes. I made the Tik Tok video from the
27:33 the Senate office with the big Calder
27:36 sculpture and and I'm like I'm like
27:39 looking at the Tik Tok, but you know in
27:41 Tik Tok if you ever do it, you're kind
27:43 of framing yourself up. And I'm like
27:45 looking at the sculpture and this and
27:46 that and I'm like halfway through
27:48 recording the video and then I noticed
27:50 my hair and it's just like it's just
27:52 like a muskrat had had nested on my
27:59 head. Oh, good lord. I didn't re-record
28:03 it, of course. Yes, the hair keeps
28:05 coming back night after night. Hey, at
28:07 least I got some hair, you know? I think
28:09 that's a that's at at my age now, we're
28:13 doing something right.
28:18 All right. Um, who's working on anything
28:20 interesting? Where do you wanna Where do
28:22 you want to play? There's a There's a
28:23 lot of things. T-shirts and irregulars.
28:25 We can go look at that. Let's go do
28:28 that. Go ahead and throw up the
28:31 uh the AI salon poster or what you call
28:36 it.
28:39 Um, if you're new here, if you've not
28:41 been to the AI salon, go to that URL,
28:46 the
28:47 salon.ai, click join our community, and
28:50 that will jump you over to this site I'm
28:52 on right here. Um, which is the AI salon
28:56 community, and uh, it'll it'll land you
28:59 on this welcome page, and we've got this
29:02 cool AI readiness cycle. Wait, why is
29:05 this not sharing? Oh, because I got the
29:09 wrong tab shared. Hang on, hang
29:13 on, hang on. I'm going to lose my ever
29:16 loving [ __ ] You always lose your ever
29:18 loving [ __ ] Okay.
29:22 Um, so if you go to the AI salon, you
29:25 end up on this this welcome page.
29:27 There's a welcome video with myself and
29:29 Leah Faston, who's the co-founder and
29:32 co-host. There's the AI readiness cycle.
29:35 There's the five stages of AI adoption,
29:37 but down the left hand
29:39 side and
29:42 Moy.
29:46 Yes. 03 after
29:49 regulars. Talk about 03. Okay. Okay, I
29:52 can do that.
29:59 Um, I've got to find Where did I put
30:02 that article? I think I put it on
30:03 LinkedIn.
30:05 Yeah. So, what we'll do, okay, and so I
30:08 wrote an
30:10 article specifically for 03. So, so let
30:13 me let me show let me to toul around
30:16 here in the salon for a second and then
30:18 we'll go talk about 03 because I think
30:20 03 is really important. Okay. Um, so
30:24 anyway, so so you you you come to the
30:26 salon. Um, there's sort of like a
30:28 sevenstep onboarding process. Step two
30:31 is introduce yourself. So if you join
30:33 and you're new, please introduce
30:34 yourself. Getting in community, getting
30:37 in an AI optimistic community like the
30:40 salon, I feel is going to be all of our
30:44 best opportunity to keep working. Um,
30:48 you know, become some of the the AI
30:50 ready people on the planet so that as
30:52 people are like, I got to figure this
30:54 stuff out, you're the ones they come to.
30:56 That's sort that's sort of the whole
30:57 idea here. And then down the lefth hand
30:59 side there's a community area, a news
31:01 area. Um there's kind of a showand tell
31:03 area. And then there's this this area
31:05 called clubs and hubs. And in clubs and
31:07 hubs is the AI learning lab and
31:10 irregulars. And this is where you can
31:12 come in to this group and just share
31:15 stuff with this group. So you all right?
31:18 Oh, I like that one. That one's really
31:21 good. Although the actual quote was,
31:23 "Are you okay?"
31:27 It's a two-way street if you can avoid
31:29 the
31:32 horseshit. That's solid. That's solid.
31:35 Who did that? That's a Steo. That's a
31:36 classic. Very nice. Okay. Beautiful. You
31:41 will be missed. All right. That was when
31:43 I was traveling. All right. If you got
31:44 more more t-shirts, here's the AI
31:47 dashboard I built to help everyone learn
31:50 about AI. Oh, cool. Vite and React.
31:55 What's VITE? Why have I never heard of
31:58 Vite? Oh, here you go. Here you go, Ann
32:03 Murphy. Model comparison
32:08 OpenAI. Wow, this is very
32:13 cool. So, this is
32:21 llm-compare-tool.web.app. Here's
32:22 Google's models.
32:26 Yeah, this
32:28 is this is really important right now if
32:32 you actually want to understand what's
32:33 what's the difference between all these
32:35 models or wait 6 months, nine
32:39 months and we're going to have a
32:41 simplification layer sitting on top of
32:43 this [ __ ] Um it'll that we'll have
32:46 model choosers and I think open AAI will
32:49 lead with that.
32:53 Okay. Ann coming in soon to talk. Oh,
32:56 okay. Cool. To talk podcast. Very
32:59 cool. Vite is like a web server that
33:02 works with React. Okay.
33:03 [Music]
33:08 Cool. If you select the models, you can
33:11 see them side by
33:13 side.
33:16 Oh. Oh, that's cool.
33:23 Huh? Wow.
33:26 Dang. Good lord. Look at all this
33:32 [ __ ] This is This is why I don't do
33:35 open
33:36 source. Open source is really important
33:39 and it's super valuable and we've got to
33:42 protect it, but it just exhausts me. Oh,
33:46 the compare button. Where's the compare
33:48 button?
33:54 Um, compare. I see. So, we can't use 41,
33:58 but we could use Let's
34:01 compare 03 and 04
34:06 mini. Oh, very
34:09 cool. Yeah, this is very slick. Very,
34:12 very
34:13 slick. Unbelievable.
34:17 You know what's amazing and I I'll get
34:20 to o3 in a second here an um or if you
34:23 come up is is an coming
34:27 up the reason the reason the singular
34:33 reason here let me show you something
34:36 I'm going to go to join me if you will
34:40 platform platform.openai.com
34:48 openai.com/playground.
35:03 Okay. Okay. So, this is the OpenAI
35:06 playground. So in here there's prompts
35:10 down the side and and images. You can
35:13 make images and you can do real- time
35:15 voice development. And then over here
35:17 there's all these settings and
35:18 parameters like threshold and prefix
35:21 padding and silence duration. And that's
35:23 just for the real time thing. But if you
35:25 want to do prompts, you got all these
35:26 different models you can choose from.
35:28 You choose all these different models
35:30 and then you choose a model and then you
35:32 choose your settings and you you compare
35:36 and then you do your temperature and you
35:39 got tools and you so it's a lot right.
35:44 It's a lot like it's a lot to figure
35:47 out. You can put your system prompt here
35:50 and then your prompt prompt here and
35:51 your conversation will show up here and
35:54 then what do you do with this? Well, you
35:57 work on your prompts and you figure out
35:59 how much they're going to cost and you
36:00 figure out which model is going to work
36:02 best and then you take your API key and
36:05 you build an application and then it
36:07 sort of works like chat GBT but it's
36:09 your application so it's kind of cool
36:11 but no one's using it but you spent like
36:13 four months on it and you build it up
36:14 right this is what large language models
36:19 were for
36:20 years for years and years and
36:23 years chat GPT comes
36:29 out.
36:30 And here, let me close all this
36:33 up. Ignore all these buttons here. It
36:36 was just the chat
36:37 hole and and you just asked it a
36:40 question and it gave you an
36:42 answer. And it was that simplicity. It
36:44 was the fact that we didn't have to
36:46 choose models is why chat GPT got to a
36:50 million users in five days and a 100red
36:53 million users in six weeks.
36:56 And
36:59 now you've got chat gvt operator suno or
37:05 sora. Uh you've got your image library
37:08 which shows you all the images you've
37:09 ever made. Did you know that that
37:10 existed? That exists. Um you've got all
37:14 these custom GPTs. You've got projects.
37:17 You've got all your chat history. You
37:20 also have all these different models.
37:24 Does this look familiar? Does this look
37:26 a lot like the playground that we just
37:29 [ __ ] came out
37:30 [Laughter]
37:34 of? So, if you're wondering why
37:38 uh why AI isn't being mass adopted, it's
37:43 because because it's moving
37:46 quickly and because we as consumers are
37:49 hungry for the non-janky AI. We keep
37:53 demanding, "Give us more models. Give us
37:55 new [ __ ] Give us new [ __ ] We want the
37:57 one that works. We want one that writes
37:58 better. We want one." Right? All that.
38:01 They're they're doing all this work at
38:04 our request and it's it's creating
38:07 feature bloat and model bloat. And so
38:09 we're just going to have to deal with it
38:10 for a while. Hashtag chat hole. Exactly.
38:13 Um, wait. The playground is free or not?
38:17 The playground is free.
38:19 Um,
38:22 oh, I think you have to I think the
38:25 playground's free, but if you want to do
38:27 anything
38:29 substantive, you've got to basically do
38:31 an agreement of
38:34 like, you know, you're going to you're
38:36 going to you can spend up to $5 before
38:39 they'll give you a warning or up to $20
38:42 before they give you a warning. And then
38:44 the different models have different
38:46 um have different costs per million
38:49 tokens
38:51 basically. All right, Ann's here. Let's
38:54 let's talk to Ann. Let's let's
38:56 let's it's time to it's time to upgrade
39:00 the conversation here.
39:07 Wait, wait, wait, wait. I just told the
39:11 people and you show up like a trucker.
39:14 You're
39:16 and and while you were saying that, I
39:18 hit my head on the tree and fell.
39:23 That's beautiful. Beautiful. Well, at
39:26 least we're keeping the brand
39:27 consistent.
39:30 What does it say? Oh, nice. Marisol got
39:34 it for me. She leads AI. That's
39:36 gorgeous. I love it. That's her design.
39:38 Yeah, that's her design. That's so
39:40 clear.
39:42 So, you can't tell me anything? No, I
39:44 can't tell you anything. No, I can't.
39:46 Did I don't know if you were here, but I
39:48 got I got yelled at by some salesman on
39:50 LinkedIn. Okay. All right. I just went
39:52 to try to find him. Yeah. To get to have
39:55 a word with him. And I can't find him. I
39:59 can't find Which post is it? Oh, it's
40:02 not it's not a po It's not a public
40:03 post. It was just a
40:05 private a private thing to me. Oh, what
40:08 a baby. Yeah, he he was like, "I caught
40:11 your live, dude. What the [ __ ] Are you
40:16 okay?" Oh my god.
40:19 [Laughter]
40:22 Refer him to me. Exactly. Are you okay?
40:26 No, we're not okay. No. No. Everything's
40:29 changing. No.
40:32 Where where are you walking? You're like
40:34 in Are you in some Oregon? Wait, is it
40:37 sunny in Oregon? Okay. It's um Yeah,
40:42 that's Oregon. Yeah, that's Oregon. It's
40:45 where I live. Shouldn't it be like rainy
40:47 and depressing? Well, yeah. That
40:50 [Music]
40:51 No. Yeah. Yes, it should. It's gonna
40:56 probably get really cold soon because
40:57 that's what it does. Oh, yeah. until
40:59 it's not summer till um July 5th.
41:05 Hey, listen. I had a reason for coming
41:07 up here. You're wasting my time here.
41:09 Please. Yes, we wouldn't want to do
41:11 that. Just kidding. You've got you've
41:13 got branches to walk into. I've got
41:15 branches to walk into.
41:16 Um I mean, there's a million things I
41:19 want to talk to you about, but
41:21 I know we got to catch up. I got to tell
41:23 you, being out for a week is just bad
41:25 for my brain. Oh my
41:27 god. combulating yet. Yeah, I'm still
41:30 recombobbulating.
41:32 Yeah, it's
41:33 brutal. Um, so Hale Lima Muhammad is our
41:38 guest on Wednesday and our guest. So on
41:42 Wednesday, Ann and I do the AI readiness
41:44 project podcast and it's 400 PM
41:49 um mountain. Yes. Right. Yes. 400 PM
41:53 time on Wednesday. um is is the podcast
41:57 uh that that Ann and I host. Okay, go
41:59 ahead. So, I'm just like I met her and
42:03 texted you immediately. I was like, "Oh
42:05 my god, she's amazing." So, she just a
42:09 little tiny bit of background, but
42:11 mostly I want to tell you in the nice
42:14 people why I'm so excited about having
42:16 her on our particular show.
42:19 So, she's a beautiful AI artist. Like
42:25 total
42:26 like just stunning. Absolutely stunning
42:30 artwork. She
42:33 herself absolutely stunning. Oh, it's
42:36 great. And she has been in the game for
42:41 just six months. Wow. And her and her
42:44 artwork is off the charts. But here's
42:47 the reason why she's so perfect for us
42:50 to have this conversation with. She did
42:53 not get famous virtually. She walked her
42:58 art into
43:00 places. She didn't even have an online
43:03 presence. No. Yes. So, she's got murals
43:08 and she's winning contests and she's got
43:11 greeting cards and she's got [ __ ] in
43:13 galleries. Wow. That's amazing. She
43:16 walked into the galleries and she was
43:18 like, "Yeah, I'm Hel Lima.
43:22 This is, you know, Ann, I'm I'm I'm so
43:24 glad you came up and said that. I just
43:26 like
43:27 I the skills moving forward like clearly
43:31 she's got a vision, right? She's got a
43:33 vision for she's got something in her
43:36 head she wants to get out. She's got a
43:38 point of view, right? A allowing her to
43:41 get that out. But that hustle, that
43:44 entrepreneur
43:46 thing where she was like, "Screw it. I'm
43:48 going to go talk to people and pitch it
43:50 and sell it." That's the thing that I
43:52 think I think is I think that's a really
43:54 big part of AI readiness because I do
43:57 too because it's you know what's going
43:59 to end up happening is like like Joy
44:01 Perty I know Joy Pertie's on the I saw
44:03 her posting here. I know she's here
44:05 tonight. You know she's been a a a sleep
44:07 researcher for for yeah 30 years and now
44:11 she's a filmmaker. Right. Yeah, exactly.
44:13 Her res her resume is not going to be
44:15 film making, right? It's gonna be it's
44:18 gonna be sleep researchery. And so for
44:21 her to work, she's just going to have to
44:23 hustle it up and invent a career. But
44:25 she's got these tools to do it now. But
44:27 that's that's I'm that's really exciting
44:31 to me because that it's just a good
44:33 reminder of like to people that are not
44:36 paying attention to what AI makes
44:38 possible,
44:40 you can appear like a wizard in this
44:42 world, right? Yeah. Yeah. That's
44:45 awesome. Okay, good. I'm super excited
44:48 now. So then, so you add that to a
44:51 couple of other things, a couple of
44:52 other examples like that and I am just
44:56 more sure than ever
44:59 that it's about authenticity
45:04 and what um Adnan and Brian talk about
45:08 with the AQ action quotient or something
45:10 like that. EQ EQ IQ AQ. Yep. Yeah,
45:15 exactly. Yeah. And that's the AQ taking
45:18 action. Yep.
45:21 And uh and then just get your [ __ ]
45:23 together. Just just move stuff forward.
45:25 I was I was also talking to Hunter who
45:28 we interviewed last week. I was talking
45:30 to him today and you know he was asking
45:33 me about his business and what he's got
45:35 going on and and one of the things that
45:37 we talked about was not getting
45:41 too married to whatever he builds right
45:44 now because he's building a new business
45:46 and then within the business he's
45:48 building new tools to do video editing
45:50 all sorts of stuff and he was like you
45:51 know what are the best tools I should
45:53 use and and I was like well quite
45:55 frankly like it's kind of irrelevant
45:58 like the best tools today are going to
46:00 be just completely different than what
46:02 are the best tools six months from now
46:03 and like the capabilities complete be
46:06 completely different. So what we're
46:08 talking about is just kind of like build
46:10 the sort of minimum stuff that you need
46:12 right now that that is really getting in
46:14 the way of your being able to do your
46:16 business. fix that stuff with AI, but
46:19 don't get too attached to it because as
46:22 stuff, you know, changes, you're, you
46:26 know, if you're too married to it,
46:28 you're you're going to be left behind.
46:29 And that's like that's a skill that
46:32 doesn't I I don't think that's an easy
46:34 one. Like for me, like I feel like when
46:36 I build something and I'm really proud
46:37 of it, I'm like, here's this thing I
46:39 built. Yeah. But it's like now three
46:42 months from now, it's going to be like
46:43 who cares because the tools have moved
46:45 on, right? And it's like that's
46:48 something we're gonna have to practice
46:49 as humans, I think. Yeah. Yeah. I you
46:52 just like you have to hold things much
46:56 more loosely.
46:58 You can't get attached because it's just
47:00 everything's so impermanent. Yeah. Think
47:03 about
47:04 like how lucky we are to have taken a
47:07 generalist approach.
47:10 Yep. Yeah. With, you know, 50 years
47:13 late, right? you know, for for all of
47:14 our lives, it's been it's it's been a
47:17 burden. But, you know, now now it's like
47:19 now it's good superpower.
47:23 Finally. Finally, the generalists. Yes.
47:26 The revenge of the liberal arts majors.
47:28 That's it.
47:31 All right. Well, I just wanted to tell
47:34 you about her. Yeah. No, I'm so glad you
47:36 came up. I'm I'm super excited. So,
47:37 yeah, put it in your calendars
47:39 Wednesday. You can also go to a
47:41 readinessodcast or AI readiness
47:45 project.com and you can see past
47:47 episodes. You can see the episode last
47:49 week where I planned poorly and and Ann
47:53 had to do it solo but Brandon helped out
47:55 and made it all work. Thank you,
47:57 Brandon. Yes. Thank you, Brandon. And
47:59 thank you for covering.
48:03 Thank you for that. Can I Can I say
48:06 something about Brandon real quick
48:08 before before I let you go? So,
48:12 the other day, no, not the other day, a
48:14 couple weeks ago,
48:16 I got to go on a podcast where I went to
48:19 a physical place
48:22 and they had like people doing like
48:26 podcast producer [ __ ] and everything and
48:28 like a little like a little room,
48:30 microphones and
48:33 Yeah, I've never had anything like that
48:35 before and it was just so fun and they
48:38 were
48:39 so nice and so
48:42 lovely.
48:46 And when Brandon is the producer on your
48:51 show, he kind of has that same vibe of
48:54 like you've been welcomed into a space
48:58 that is being held by somebody. Y even
49:01 though it's not physical. Yeah. So
49:03 Brandon,
49:06 you're awesome. No, it's great. He is.
49:09 Yeah, it's really good. It's It's good
49:10 to feel like just know that someone's
49:12 paying attention because you know me.
49:16 Yeah, I may or may not be at any given
49:18 moment.
49:22 Oh, man. Well, it's great to see you.
49:24 Say hi to Donnie for me. I will. I will.
49:27 Okay. And then and I'll see you
49:29 Wednesday. Yeah, I'll see you Wednesday.
49:31 You and all the nice people. Okay. Okay.
49:34 Bye. Bye.
49:37 Watch out for the
49:38 tree. That was from
49:42 Brandon. Um, all right. So, let me um
49:47 Ann was asking about 03. Let me I'm
49:49 going to go find a post that I
49:56 [Music]
49:57 did. Was it a post or was an article? It
50:00 was an
50:03 article.
50:07 Wasn't it an
50:11 article?
50:19 Yes. Yes.
50:24 Okay.
50:26 Okay.
50:29 So, is there a
50:32 way? I wish there was a way. There
50:34 probably is a way without having this
50:37 giant comment section over to the right.
50:40 So, if you go to to
50:42 LinkedIn and search for
50:46 um how I used a reasoning engine to use
50:49 a reasoning engine by Kyle Shannon. Um
50:53 or let's see, I
50:57 can put this I'll put this in the
51:00 YouTube chat as well.
51:07 So, what this is is Okay, so there so
51:09 there's a couple of things. Let me let
51:11 me jump over to chat
51:14 GPT. Right
51:16 now, you've got a bunch of different
51:18 models here and it's very confusing and
51:22 it's I pay attention to this stuff and
51:25 it's very
51:27 confusing.
51:30 Um 4, which stands for four
51:34 omni, is the multimodal sort of core
51:39 model that that you use for most stuff,
51:41 right? Great for most tasks. This is my
51:44 daily driver. Um I've I've heard of some
51:47 people using 03 as their daily driver,
51:49 but
51:50 um but for me, 40 is just it's really
51:55 good at writing. um you know it can do
51:58 search. It's just good. It's fast
52:00 fastish depending on the
52:03 day.
52:06 4.5 is an anomaly that I don't really
52:10 use all that much. But if 40 is not
52:14 giving me interesting results,
52:17 4.5 is a more empathetic model and it's
52:22 a more creative model in a lot of ways.
52:26 From what I understand, 4.5 is is likely
52:30 what's going to
52:33 be
52:35 the orchestrating model when GPT5 comes
52:39 out. So GPT5 is going to take all these
52:42 models and kind of hide them from us.
52:45 And there's going to be this
52:46 orchestrating layer that we interact
52:48 with that's going to figure out which
52:50 models to use. From what I understand,
52:52 that's what 4.5 is going to be used for.
52:55 So it says research preview here. It's
52:57 weird. It's like I would say like all AI
53:02 stuff, play with it. Play with it. Um
53:04 there might be things it does that you
53:06 love it. I find it weird and sometimes
53:09 good, sometimes not. I don't know when
53:10 to use it. So, I tend not to if I don't
53:13 know when to use something. Okay. Then
53:15 you've got 03 04 and 0 O 04 mini and 04
53:19 mini
53:21 high. And you're like 40, 04, same
53:24 thing, right? No, completely
53:26 different. The O after the number means
53:30 omni. The O before the number means
53:32 reasoning engine.
53:40 So these three things are reasoning
53:42 engines. Now what's a reasoning engine?
53:45 A reasoning engine is just like chat GPT
53:50 except when you ask it something instead
53:53 of it just answering you right which is
53:56 that's called next token prediction
53:59 right it's it's a when when you're in
54:01 forro you give it a prompt and it just
54:04 answers you sort of one token at a time
54:06 or you know group of tokens at a time
54:08 and it just writes the answer that's why
54:10 you get all this hallucination and
54:12 weirdness right what a reasoning engine
54:15 does is it talks with itself first
54:18 multiple times. So, you ask it a
54:21 question and it goes, "Huh, let me think
54:23 about that." And it says, "The user
54:25 asked us for this and maybe if they ask
54:27 for that, maybe we should go to the
54:29 internet and search for things." And
54:30 then it'll go search for things and then
54:32 it'll find things and it goes, "Huh, out
54:34 on the internet, I found this." And
54:36 right, it's doing all of this thinking
54:38 and multi-step
54:42 processing. 03. Um, let me see. Let me
54:45 do a thing here. Um, let's see. find
54:49 the uh 2024
54:52 uh Denver
54:55 budget
54:57 actuals and
55:01 um
55:03 categorize
55:06 anomalies where you think it would be
55:11 interesting to understand what happened.
55:17 Okay, that's a fairly horrible
55:19 prompt and I spelled anomalies wrong,
55:22 but I now this thing is going to go off
55:26 and start thinking. So, see it says
55:28 thinking here and then at some point
55:30 it's going to have thought enough to
55:32 show us a little something.
55:47 Whack. There we
55:49 go.
55:52 Okay. So, the user asked for Denver's
55:55 budget actuals and any anomalies. Next
55:58 bullet point. I'm thinking the steps to
56:01 find de Denver's I'm thinking through
56:03 the steps to find them. Searching the
56:05 web. Here's six different sites that it
56:09 searched. I'll organize the anomalies
56:12 into
56:13 categories. Colorado General Assembly
56:16 did something. There's a schedule of
56:19 things. It's searching the
56:22 web. It's doing more [ __ ] So, it's all
56:26 still thinking, right? Right. These are
56:28 not answers. This is it
56:32 thinking. More searching on the
56:35 web. More searching on the web. It can't
56:38 find it. may need to consider the
56:40 mayor's 2025 budget. At some point here,
56:43 it's going to find numbers and it the
56:46 other thing that this this these models
56:50 can do is they can make and use their
56:53 own tools. So if an if it finds a PDF
56:58 with the actual numbers in it, it might
57:00 need to do some analysis on those
57:02 numbers. So there was a thing a year or
57:05 two ago called uh uh code interpreter
57:09 which which was kind of a standalone
57:11 mode within chat GPT where it could
57:13 write and execute Python code and do
57:15 number analysis. Well, 03 and 04 mini
57:20 and mini high can now all do that in the
57:23 context of thinking, right? And so it's
57:26 really good for math. It's really good
57:28 for science. It's really good for
57:31 research. Um, it's still thinking. It's
57:34 still going on. This was a good one. I'm
57:36 glad we did this one because it's it's
57:38 taken it a while, right? It's having
57:39 trouble finding the
57:41 actuals. Ah, official 2024 budget
57:44 actuals haven't been published yet.
57:47 So there you
57:51 go. Um All right. So it's done. Below's
57:56 the clearest picture I can
57:58 find. Right. And it's still thinking.
58:01 It's giving us hot spots and
58:06 underspends. All right. So thinking
58:08 model. And then the difference between
58:10 03 and 04 mini basically just, you know,
58:14 the 04 ones are newer, but they're
58:16 smaller. The 03 one is kind of the Mac
58:18 Daddy, you know, most state-of-the-art.
58:20 Uh, but think of the ' 04 ones as as
58:23 essentially previews for what's coming.
58:25 Um, they're really quite powerful. Um,
58:27 and if you run out of credits in 03, you
58:29 can you can do the ones in 01. So that
58:32 article that I wrote was
58:35 this. when they when they rolled out 03,
58:40 um the examples that they used were
58:44 here's a crazy math equation that I had
58:47 a tough time solving in college and then
58:50 it solved a math thing and then the
58:52 other one was like here's a programming
58:54 thing that is a big challenge and it it
58:56 did some programming and and you know
59:00 they did a physics thing and a math
59:02 thing and like they did all this sort of
59:03 high science [ __ ] right Because these
59:06 tools are smart enough now that they're
59:08 starting to get really good at math. In
59:11 fact, David Shapiro said of 03, they've
59:14 solved
59:16 math, which is a big [ __ ] statement
59:18 and it's a big [ __ ] deal,
59:21 right? If you need
59:26 math, but what if you don't need
59:30 math? What if you are a marketer? What
59:33 if you're a knowledge worker? What if
59:35 you're not doing statistical analysis,
59:37 but you're doing strategy and marketing
59:41 and copywriting and, you know, business,
59:46 you know, I don't know, positioning, how
59:49 can you use it? So, what I what I wrote
59:51 in the LinkedIn
59:53 article was I went to
59:57 Grock and and I said to Grock, so so one
1:00:00 of the things, oh, let me let me show
1:00:02 you Grock. If you don't know Grock, so
1:00:04 if you go to X, formerly
1:00:15 Twitter, you got your you got your, you
1:00:17 know, your Twitter stuff,
1:00:20 right? Everyone's making movies
1:00:23 now. And then down the side here, you
1:00:26 got this thing called Grock. And Grock
1:00:28 is just like chat GPT except it's Elon
1:00:31 Musk's and he hates Chat GPT and he
1:00:33 hates Sam
1:00:36 Alman. But one of the things that Grock
1:00:39 does is it can search all of Twitter,
1:00:43 right? It can it can tap into all of
1:00:45 what Twitter said. So I went to Grock
1:00:47 and I said, "Tell me how people are
1:00:50 using this 03 reasoning model for
1:00:52 humanities, for things that are not
1:00:55 science, not STEM.
1:00:58 and it basically couldn't find
1:01:00 anything. And so then I had to come up
1:01:03 with 10 use cases for just pure
1:01:05 humanities, 10 use cases for
1:01:09 um humanities stem hybrids and 10 use
1:01:14 cases for image generation because 03
1:01:16 can also see and analyze images. And so
1:01:20 that's what that article is is basically
1:01:23 30 different use cases.
1:01:26 So, story and screenplay structuring,
1:01:31 um, character consistency and world
1:01:34 building, right? Budgeting, production
1:01:36 planning. So, if you're trying to figure
1:01:38 out how to use 03, this article is
1:01:42 designed to just teach you how to do
1:01:44 that. Hang on, I have to do an
1:01:45 intelligence
1:01:48 test. I feel feel like I'm the president
1:01:51 of the United States. We're going to
1:01:53 need you to figure out which of these
1:01:54 letters is the same. I didn't realize
1:01:57 there'd be a
1:02:01 quiz. I'm just trying to run the country
1:02:08 here. All
1:02:09 [Music]
1:02:10 right. Confused still. Yeah, I know. I
1:02:15 Well, I I absolutely agree. Here's the
1:02:18 here's the deal. So, so Cindy [ __ ] I
1:02:21 wrote this article. Cindy [ __ ] took a
1:02:24 two hours on a Saturday and she did
1:02:26 every one of these prompts. Actually,
1:02:29 she she did something really cool. So,
1:02:32 so she took this entire article and she
1:02:34 put it into chat
1:02:36 GPT and then she took all the
1:02:38 information about a client job that
1:02:40 she's working on and she put that into
1:02:42 chat
1:02:43 GPT and she said, "I want you to rewrite
1:02:46 all of these use case prompts
1:02:49 specifically for my client." And it did
1:02:52 it. And she said it was [ __ ]
1:02:53 brilliant. And so it created for her 30
1:02:57 new, you know, use cases with with
1:03:00 simple, medium, and advanced prompting
1:03:03 for her client. And she took two hours
1:03:05 on a Saturday and just went through all
1:03:07 30 prompts. So if you want to if you
1:03:09 want to
1:03:10 explore what 03 is good at, you know,
1:03:14 just literally copy and paste out of
1:03:16 that article. All
1:03:19 right. Are we talking about maths? Thank
1:03:22 goodness for Gen AI because me and maths
1:03:25 don't play well together. You know
1:03:26 what's funny, Source Camp? I
1:03:30 um someone else was talking about
1:03:32 [ __ ] math the other day and like you
1:03:35 know when you have an acting degree at
1:03:37 some point in your career you might play
1:03:39 a
1:03:40 mathematician and then you're going to
1:03:42 have to semi give a [ __ ] about it. But
1:03:45 for the most part, it's just it's just
1:03:48 not on the same side of the planet as
1:03:55 math. So I asked um I think I asked
1:03:59 03. I said, oh, someone someone
1:04:04 said I think it was
1:04:07 um Google DeepMind I think came out with
1:04:10 a new they came out with a new model and
1:04:12 it's a new math model and basically it's
1:04:15 going to be it's it's a math agent
1:04:17 that's going to come up with new
1:04:20 conjectures and hypotheses and whatever
1:04:23 the [ __ ] they call the things in math,
1:04:25 right? The the theories that you need to
1:04:27 make proofs for. It's going to create
1:04:29 new ones of those and then it's going to
1:04:31 also be able to design ways to prove
1:04:34 them and then it's going to prove them
1:04:35 and then we're going to have new
1:04:37 mathematical proofs and theorems
1:04:40 basically. And so I was like I feel like
1:04:44 I should be excited about that but I
1:04:46 just don't like I have no idea zero what
1:04:51 that actually means to me. like what
1:04:54 does it mean that we come up with some
1:04:56 [ __ ] new theorem and then we prove
1:04:58 it? And so I asked 03 I said I said you
1:05:02 know research and give me a history of
1:05:05 major mathematical proofs and what
1:05:09 technologies came out of them. Um and
1:05:12 then and then you know hypothesize what
1:05:16 might be the next 10 major mathematical
1:05:19 breakthroughs and what will be the
1:05:20 things that come out of them. Um, it was
1:05:23 pretty cool. You know, I'm still not
1:05:25 into math because I just don't It's like
1:05:27 when I look at a chessboard, I know
1:05:29 where the pieces go, but I do not see
1:05:31 patterns, right? Because I don't know it
1:05:33 well enough, and I'm like that with
1:05:34 math. So,
1:05:35 anyway, all right. The models are stupid
1:05:38 in high level mathematics. Have you
1:05:40 played with 03 and 04 mini high? Because
1:05:43 apparently it's quite good. Um,
1:05:46 paradoxes and chaos of math. Yes, I that
1:05:50 sounds like it means something. cyber
1:05:56 tech. Could you scroll to the top of the
1:05:59 LinkedIn article again? I want to take a
1:06:01 screenshot. Yeah,
1:06:05 [Music]
1:06:08 absolutely. Oops. Here. And let me get
1:06:11 rid of I'm going to get rid of this
1:06:12 tracking thing so that the URL is just
1:06:15 the pure URL.
1:06:22 And I wish I could get rid of this
1:06:23 thing. Can I get rid of this thing? This
1:06:25 two column
1:06:29 [ __ ] All right, there you go.
1:06:32 Screenshot
1:06:34 away. Where is the article, please? The
1:06:36 article is in uh on on YouTube in the
1:06:41 comments or it's on LinkedIn if if you
1:06:44 search for how I used a reasoning engine
1:06:47 to use a reasoning engine. That's the
1:06:49 name of the article and I'm Kyle
1:06:51 Shannon. So I asked Chat GPT why Brits
1:06:54 say maths instead of math and basically
1:06:57 said same reason we say mathematics and
1:07:00 not
1:07:01 mathematic. Oh that's fascinating Jeff.
1:07:03 I've thought that too. I feel like math
1:07:06 maths is just pretentious, but it
1:07:08 actually makes sense that we're the
1:07:11 idiots. Once again, America are the
1:07:14 idiots. We're wrong. The world is right.
1:07:18 Maths is the way is the way you should
1:07:19 say it. But but in our defense,
1:07:24 math, there's there's like two soft
1:07:31 consonant sounds together. Math.
1:07:36 Math.
1:07:38 Math's. It's hard. Someone just math.
1:07:41 Math is
1:07:42 easier. We're lazy. We're
1:07:45 lazy and we're arrogant and we think
1:07:49 we're the
1:07:51 best. Welcome to
1:07:55 America. It's always
1:07:58 maths. Um like moths.
1:08:01 Kyle. I I only say
1:08:05 moth. Do you know what you call um a a
1:08:08 uh a a swarm of moths?
1:08:14 Moth. Um
1:08:17 anyway, go play with this. Go, you know,
1:08:20 go play with the prompts here.
1:08:21 Especially if you're not a maths person
1:08:24 or a physics person or you're not
1:08:26 solving PhD level science
1:08:30 problems. Here's 30 different prompts.
1:08:33 It's actually more than that because
1:08:34 there's a basic prompt and an
1:08:36 intermediate prompt and an advanced
1:08:37 prompt. So, there's 90 different prompts
1:08:40 in this
1:08:42 article. Some are about images, some are
1:08:45 about other [ __ ] Wait,
1:08:52 what? This isn't the complete
1:08:55 article. Uh
1:08:59 uh wait, this isn't the complete
1:09:01 article. This one was from January. I
1:09:03 did I did a sooner one. Hang on. Hang
1:09:05 on, people. There's a newer one with
1:09:08 with more more things. Oh, you know
1:09:11 where I did it? I did it on
1:09:13 [Laughter]
1:09:21 X. All right, hang on. People, stop
1:09:24 judging me. Oh, look. My little my
1:09:27 little character says I'm
1:09:30 live. All right.
1:09:32 Articles.
1:09:34 Okay. Go to go to Twitter
1:09:37 X and go to my profile, Kyle Shannon,
1:09:41 and then click on
1:09:43 articles. This is the article.
1:09:46 Okay. On X. Unlock 03's sup hidden
1:09:51 superpower. A playbook for non- STEM
1:09:54 prompts for everyday innovators. That's
1:09:56 why that article wasn't feeling familiar
1:09:57 to me. This is the one. This is the non-
1:10:00 STEM prompts for everyday
1:10:05 innovators. And this is the one that's
1:10:07 got 30
1:10:10 different use cases in three different
1:10:13 categories.
1:10:15 Okay.
1:10:19 Good. Soon we'll have a Millennium Prize
1:10:21 winner. We will. I tested Claude and
1:10:24 Grock and
1:10:25 Deepseek. No one could do do advanced
1:10:28 mal matrix calculations correctly. Well,
1:10:31 hang on. Let's go do it. Can Can you So,
1:10:35 Sigma, did you did you try 03 and 04
1:10:38 mini high advanced matrix calculations?
1:10:42 Okay, we're in 03 here. Um, hey
1:10:46 um,
1:10:48 you can you explain to me what an
1:10:55 advanced what is it? Matrix
1:10:58 calculation.
1:11:00 Um, advanced matrix. Can you explain
1:11:03 what an advanced
1:11:05 matrix calculation is?
1:11:14 And again, I'm going by David Shapiro
1:11:16 here who who is good at maths. Um, and
1:11:19 he says math is basically solved with
1:11:21 03. Think of matrix as a spreadsheet
1:11:24 full of numbers. Okay, great. Whatever.
1:11:27 So, give me an example of something I
1:11:29 should do. Sigma. Yes, I did test
1:11:31 mathematics 03. Can't solve the 4x4
1:11:36 matrix. Een vacators and values.
1:11:45 God, did you eg the lap the the laplace
1:11:50 matrix of two matrices? Yes, it's what
1:11:54 TP TPUs for calculations. Okay, listen.
1:11:58 From where I sit, this shit's good at
1:12:00 math. From where you sit, sounds like
1:12:03 it's not
1:12:04 yet. That's a solution for matrix.
1:12:11 Okay. Um
1:12:15 um design a 4x4
1:12:20 matrix. What did you say? A
1:12:25 4x4
1:12:27 matrix.
1:12:30 Een
1:12:31 een vocators. Look, I'm learning new
1:12:34 words tonight.
1:12:39 and value
1:12:42 [Music]
1:12:44 um
1:12:47 exercise for you to solve. So we'll have
1:12:51 it design the
1:12:52 exercise
1:12:56 thinking this is this is really stupid.
1:13:00 This is like putting a math PhD in a
1:13:03 Broadway theater and saying, "Direct the
1:13:05 the actors. Go block the
1:13:11 actors." Oh
1:13:14 man. Egen vectors. Oh, Egen vectors and
1:13:17 Egen values. Wait. Analyzing
1:13:21 import. Oh
1:13:25 god. Oh yeah. Egen vectors here. Let me
1:13:28 stop it. Let's do een
1:13:32 vectors and
1:13:35 een
1:13:38 values should be in the 4x4
1:13:44 uh
1:13:50 design. Like again, my job here is to
1:13:54 act like the idiot so you don't have to.
1:14:00 Oh, it's Egen. It's It's got an I in it.
1:14:02 Well, whatever. Um, uh, Chat GPT knows
1:14:06 how it's supposed to be spelled. It's
1:14:09 asking if you like 03 as a
1:14:14 personality.
1:14:16 Um, I
1:14:18 don't I like I like 40 as a personality.
1:14:23 Although so, so they did the thing where
1:14:26 it got really sick ofantic and then they
1:14:28 rolled that one back. Since they've
1:14:31 rolled it back, I'm not super happy with
1:14:33 it. The exercise, take this symmetric
1:14:36 4x4 real matrix.
1:14:39 Okay. Find the een values of a for each
1:14:45 een value.
1:14:48 give a nonzero EGEN vector and if you
1:14:51 like a unitlength version. You know,
1:14:53 it's funny. I was thinking about Egen
1:14:56 vectors and Egen values the other day at
1:14:58 lunch and and I was like, but wait a
1:15:01 minute. It it's like I need a unit
1:15:04 length version. I can't just have
1:15:06 vectors and values without a unit length
1:15:08 ver length version, right? And I'm
1:15:11 thinking this as I'm making my peanut
1:15:13 butter and
1:15:14 jelly. And uh and then I just stopped
1:15:18 the brain the brain shut down at that
1:15:20 point. All right. Show the four EG
1:15:23 vectors from an orthonormal basis. Hey
1:15:27 Brandon, didn't you say you were or
1:15:28 ortho or
1:15:37 orthonormal? Funny. I need gravy.
1:15:44 Oh my god. As a sanity check. Well, so
1:15:47 so listen.
1:15:50 Um, we obviously need to form things on
1:15:53 an orthonormal basis, but as a sanity
1:15:56 check, we need to verify that that P to
1:16:00 the minus1 AP does still equal A, where
1:16:03 P is the matrix of columns where the
1:16:06 Egen vectors and the I don't know what
1:16:08 the Oh, it's not A. It's a upside down
1:16:11 V. What's an upside down
1:16:14 V? First syllable is
1:16:19 I. Listen, if you smarty pants think I'm
1:16:22 making fun of you, oh contr, I like this
1:16:25 is literally Greek to me. And I'm pretty
1:16:28 sure this is Greek, right? Didn't the
1:16:30 Greeks come up with all these
1:16:36 symbols? Oh my
1:16:38 god, Kyle. When someone wants you to
1:16:42 start Egen, you need to make a 1970s
1:16:44 retro mod. Exactly. I think we're going
1:16:47 to we're going to go back and do a uh
1:16:50 we're going to go do a defender
1:16:53 clone. Okay. So So there it's
1:16:55 understood. Okay. I'm going to go solve
1:17:06 this. Uh, capital lambda is an upside
1:17:09 down V. Oh, yeah. Yeah, capital lambda.
1:17:12 I knew
1:17:18 that. It's importing Simpai as SP maths.
1:17:22 Uhoh, it wrote something in
1:17:26 red. One, write the matrix and spot the
1:17:29 block structure.
1:17:32 The upper left and lower right 2 by two
1:17:35 blocks are independent. So we can solve
1:17:37 each of
1:17:39 them. Look,
1:17:41 maths. It put maths on
1:17:46 screen. Have you guys seen the uh the
1:17:49 show Prime? It's not Prime Suspect. It's
1:17:53 Prime. What's the What's the name of
1:17:55 that show? I think it's on Netflix.
1:18:00 Prime. No, it's not prime numbers. It's
1:18:03 like prime suspect or prime not
1:18:06 magics. It's about a math genius and how
1:18:10 all these people around him keep getting
1:18:12 killed because he's working on prime
1:18:15 numbers. And basically prime target.
1:18:18 Yeah, that's what it
1:18:19 is. It's called Prime Target. It's
1:18:21 really good. And it's basically
1:18:25 um he's super genius working on
1:18:28 primes and if he
1:18:32 solves the thing that he solves
1:18:35 ultimately um cryptography is dead and
1:18:38 so that's why everyone's killing him
1:18:40 trying to kill
1:18:42 him. Orthog orthonogality
1:18:47 check
1:18:49 diagonalization the EN vectors as
1:18:51 columns of
1:18:54 P. Yeah, this is the cool thing too if
1:18:56 you haven't seen this like so within
1:18:59 within its reasoning. So this is all the
1:19:00 reasoning it's doing. So it's writing
1:19:03 you know it's writing Python and
1:19:04 executing it as part of the process. The
1:19:07 result EGEN values one, two, three, and
1:19:09 eight corresponding
1:19:11 orthonormal EGEN vectors listed above. A
1:19:17 diagonalis cleanly via
1:19:20 P lambda P to the T as
1:19:24 required. Is this an
1:19:32 answer? Pete's going to lose his mind
1:19:34 when he watches this back.
1:19:40 Did does is is this right at all? Does
1:19:43 it matter? Does should we just move
1:19:50 on? Oh my god. New image and
1:19:54 irregulars. It's making fun of me trying
1:19:56 to do math.
1:19:57 [Laughter]
1:20:01 Maths. Dude, what the [ __ ] Are you
1:20:07 okay? Actually, you know what? This is
1:20:10 going on
1:20:12 my copy image or no, I think I got to
1:20:15 save it. Um, save image as this is this
1:20:20 is going on my uh on my LinkedIn profile
1:20:23 right
1:20:29 now. Start a post.
1:20:34 Last
1:20:36 week, someone saw one of my nightly AI
1:20:44 learning
1:20:45 lab live
1:20:49 sessions and messaged
1:20:54 me
1:20:57 with
1:21:01 this dude.
1:21:05 Dude,
1:21:14 WTF. Are
1:21:16 you
1:21:19 okay? The answer is a
1:21:23 resounding
1:21:27 no. Oh, yeah.
1:21:31 [Music]
1:21:37 Why can I Oh, because it's AVIF. Of
1:21:39 course. Of course it's the wrong [ __ ]
1:21:42 file format. Of course. Of course it is.
1:21:45 Of course it is. Why would we be able to
1:21:48 take something that was produced on a
1:21:51 machine and import that into a different
1:21:54 piece of software on that machine? cuz
1:21:56 someone at some point decided that the
1:21:59 AVIF format was important
1:22:03 and all the other people said, "No, it's
1:22:06 not." And then some of the other people
1:22:08 were like, "Well, we're going to put it
1:22:09 in our software anyway." So that when
1:22:11 Kyle tries to do his unhinged [ __ ]
1:22:15 Tik Tok channel where he talks about AI
1:22:18 in a semicompetent way, none of the [ __ ]
1:22:21 will
1:22:23 work. We hope that works out for him.
1:22:33 [Music]
1:22:37 The hair the hair is
1:22:40 perfect. It's pausing so you can read
1:22:43 your post again. Did I have a typo in
1:22:45 there? I'm sure I did.
1:22:47 Last week someone
1:22:51 say saw fine silver fox fine saw one of
1:22:57 my nightly AI learning lab live
1:23:00 sessions fine and messaged me message me
1:23:03 with this dude what the [ __ ] are you
1:23:05 okay the answer is a resounding no um
1:23:10 thank you
1:23:12 to Serena wait how do you spell your
1:23:14 name Serena Serena for um
1:23:21 memorializing the
1:23:30 concern. Oh, good
1:23:32 lord. OCD. I know. I know. Okay. Did Did
1:23:39 Where? There it is. Okay. We're Dude,
1:23:43 what the [ __ ] Jeffy KS, are you
1:23:48 okay? Wait. S I R E N A. S I Oh, good
1:23:53 lord. S I R E N A.
1:23:57 [Laughter]
1:24:06 [Music]
1:24:08 Okay. All right. We good to go? Are are
1:24:11 all of Are all of my drunken white
1:24:15 junkies are Are we
1:24:18 okay? Can I hit post,
1:24:27 please? They've got little Post-it notes
1:24:30 sticking out of there white. It's all It
1:24:33 won't stay flat anymore because they
1:24:35 keep breaking the spine to lay it flat.
1:24:41 [ __ ]
1:24:42 nerds. You good? Okay,
1:24:49 good. Oh, good lord.
1:24:52 [Laughter]
1:24:57 Okay, my postit budget for this show.
1:25:01 Oh, that was fun. All
1:25:04 right, we did it timewise. Have we
1:25:07 talked about anything tonight? I don't
1:25:09 think we
1:25:10 [Laughter]
1:25:13 have. Okay. So, um actually, yeah. You
1:25:18 know what, Brandon? Flip flip back over
1:25:20 to
1:25:23 uh to the AI
1:25:30 salon. So, if you're not a member of the
1:25:32 AI salon, get the to that a the
1:25:36 salon.ai.
1:25:37 AI. And then um Tuesday night, so
1:25:42 tomorrow night at 5:00 PM Mountain,
1:25:45 we're doing an AI salon presents and
1:25:48 we've got Telicia White who's the
1:25:50 community manager for Lovable. She's
1:25:53 going to be joining us and and showing
1:25:56 us some vibe coding stuff. And Vicki,
1:25:58 are you going to be I don't know if
1:25:59 Vickiy's in here tonight. I know she was
1:26:02 tra traveling. I think she might be
1:26:04 Vicki might be presenting as as well,
1:26:06 but I'm not
1:26:07 sure.
1:26:09 Um, she's here. Okay,
1:26:12 cool. Um, so, so Vicki, I don't know.
1:26:15 Are you are you doing something as well?
1:26:17 So, so please spread the word about
1:26:20 this. This is going to be a good one. I
1:26:21 think lovable of the vibe coding apps
1:26:23 right now is the one if you are not uh
1:26:27 if you are not particularly techy, it is
1:26:29 the most forgiving. it it's the one that
1:26:31 requires the least of you. I'm here not
1:26:34 presenting just Telicia. Okay. So,
1:26:35 Telicia from Lovable will be presenting
1:26:38 and uh yeah, so come to that. So, when
1:26:41 you go to the AI
1:26:43 salon uh up in the upper leftand corner
1:26:47 at the very top, just go to events and
1:26:50 then you'll see salon presents right at
1:26:52 the top and just RSVP for that.
1:26:58 Okay. Yeah. And there's actually a big
1:27:00 announcement tomorrow. We've got some
1:27:02 some exciting new stuff we're doing in
1:27:04 the salon. So, um, so that's going to
1:27:06 happen as well tomorrow. So, get your
1:27:08 butt to the meeting. All right. You are
1:27:11 presenting very important news. Yeah,
1:27:13 exactly. Yeah, we we've got a a big
1:27:14 announcement tomorrow. So, I'll let it
1:27:16 be a surprise for
1:27:17 tomorrow. Um, all right. So, tomorrow
1:27:22 Salon presents
1:27:25 uh uh
1:27:31 uh I'm trying to
1:27:33 think. I don't want to start it now
1:27:36 because it's going to take too long. But
1:27:38 I think what I might do, maybe we'll do
1:27:40 it Wednesday because tomorrow I'm going
1:27:42 to be I love Mondays. Tomorrow I'm going
1:27:45 to be crispy because of the salon. I
1:27:47 like I'm always optimistic that I'm like
1:27:49 gonna get back from the salon and I'll
1:27:51 be like I'll be all peppy and then I'm
1:27:53 like I'm exhausted. So tomorrow's not a
1:27:56 good night either. So I I think tomorrow
1:27:58 I'll just whine. So today was meltdown
1:28:02 Monday and then tomorrow will be whiny
1:28:11 Tuesday. You you people with with word
1:28:14 skills.
1:28:16 You people with word skills are going to
1:28:18 have to come up with something for
1:28:19 Tuesday. Tirade Tuesday. Um I'll do
1:28:22 tirade Tuesday and then Wednesday I
1:28:23 think what I'm going to do is I want to
1:28:26 do a compare and
1:28:28 contrast to runway
1:28:31 um what are they called?
1:28:34 References. And then Midjourney also has
1:28:38 this new thing called omnireence where
1:28:40 you can like put an image in this
1:28:42 omniresence thing and it'll add it to
1:28:44 whatever image you're generating.
1:28:47 Um, I think runways is
1:28:50 better. And then I also think there's
1:28:53 another one. It must be Korea or Clling
1:28:55 I think also has references. Or is it
1:28:57 Pika? I don't I can't keep track
1:29:00 anymore. But I feel like we should do So
1:29:02 what these things are
1:29:05 is they allow you much higher level of
1:29:08 control. So like you can take a
1:29:09 character that looks like a specific
1:29:11 character and add it into a scene.
1:29:18 Troubleshooting Tuesday. Yeah, maybe
1:29:19 we'll do troubleshooting Tuesday. Come
1:29:21 with a problem and we'll we'll figure it
1:29:23 out. I like
1:29:25 that. Or tirade. I I might go off. Maybe
1:29:30 in troubleshooting something. I could go
1:29:32 off. You bring the problems. I'll bring
1:29:35 the
1:29:39 frustration. Just joined. Did you
1:29:42 already talk about the open AI decision
1:29:45 uh for nonprofit control and wind surf
1:29:47 by? Um I didn't Austin.
1:29:50 Um so there was an announcement today.
1:29:53 So as as you likely know if you've been
1:29:55 on this channel um OpenAI has said
1:29:58 they're going to move to a for-profit
1:30:00 model and you know Elon Musk lost his
1:30:04 [ __ ] over that. It's got to be
1:30:05 nonprofit. Um, they announced today that
1:30:07 that they are the for-profit entity is
1:30:11 going to be a BC Corp, right, which is
1:30:14 essentially a for-profit entity for
1:30:16 good. I don't know if that's what it is
1:30:18 now. I don't think it is, but I'm not
1:30:20 sure. Um, but it's going to be it's
1:30:23 going to be a BC Corp and then it's
1:30:25 still going to be managed by the
1:30:26 nonprofit board. Um, which I find
1:30:29 fascinating. So, I don't know why they
1:30:31 did that. Um, I don't know if it's got
1:30:33 to do with shareholder lawsuits or if
1:30:37 they just decided it was in their better
1:30:39 interest not to go full public, but that
1:30:42 happened. And then Yeah. And I didn't
1:30:43 see Did they make an announcement today
1:30:46 officially that they're buying wind
1:30:47 surf? Because I know it's been rumored
1:30:49 for about three weeks, Tourette's
1:30:51 Tuesday. Every day is Tourette's day at
1:30:53 the AI learning lab.
1:30:55 Um, could you take can you take a
1:30:58 specific product and make that part of a
1:31:00 video image scene? Yes, you
1:31:04 can. Deal closed. Oh, the deal closed
1:31:07 with Windsurf today. Holy [ __ ] No, I
1:31:10 didn't see that at all.
1:31:12 $3
1:31:14 billion. A 500 500 Well, that's $500
1:31:18 million nonprofit. Probably $500
1:31:22 billion. what they're going to be $3
1:31:25 billion for Windsurf. So these vibe
1:31:29 coding
1:31:31 apps, I don't know if you know this, do
1:31:33 you remember when when Chat GPT first
1:31:36 came out, everybody was bad mouthing
1:31:39 rapper apps? So rapper apps were
1:31:41 anything that people put [ __ ] together
1:31:43 that were just built on top of large
1:31:46 language models. And all the developers
1:31:48 were like, "Ah, those things are just
1:31:50 [ __ ] They're never going to work
1:31:51 out." Well, a rapper app called Windsurf
1:31:54 just sold for three billion
1:31:58 dollars. So, so if you have an idea for
1:32:03 something, [ __ ] build it because the
1:32:06 money out there is stupid right
1:32:10 now. Technically, an investment was for
1:32:13 the profit
1:32:14 branch. Wow, that's
1:32:17 crazy. All right. Well, cool. Love it.
1:32:23 Um, let me show you. I'll I'll do a
1:32:25 quick uh midjourney thing. I'll just
1:32:28 I'll just do a quick uh thing. Here's
1:32:31 here's one that I did. Here's some
1:32:33 images that I
1:32:39 did. So, so I've got this picture. Let
1:32:42 me see if I've I've got it in
1:32:44 here. I don't think I put it in here.
1:32:47 Did I? Hang on.
1:32:51 Yeah, this
1:32:54 one. So, this image, this was an image
1:32:58 that I created um in
1:33:01 2022 as part of my Kyle Shannon Dreams
1:33:04 art project. Um, and it's it just this
1:33:09 is one that just gives me joy because
1:33:11 it's so [ __ ]
1:33:13 stupid. It was like, you know, stable
1:33:16 diffusion would either make me like this
1:33:17 big tubby 400 pound dude or make me like
1:33:20 this chiseled model. There was rarely
1:33:22 anything in between. So, I just thought
1:33:24 this and I thought the hair was perfect.
1:33:26 And so, I took that dude and I put it in
1:33:29 the in this new omni reference thing.
1:33:32 So, um, the way it works is you grab an
1:33:37 image or you can upload an image, but
1:33:40 you grab an image and you drag it into
1:33:42 Omni
1:33:43 Reference. So, you've got image prompt
1:33:47 here. You've got style reference here.
1:33:50 So, this is keep the composition of the
1:33:52 image. Style reference is keep the style
1:33:55 of the image. And then this one is add
1:33:57 this thing to whatever I describe. So,
1:34:00 what we could do is we could say, let's
1:34:02 take a let's go find a fun
1:34:07 style.
1:34:11 Um, this one's kind of
1:34:15 cool. So, we'll take that as a style
1:34:18 reference. Kind of weird. And then we'll
1:34:22 say man and lion
1:34:26 walking down the middle of
1:34:31 a city
1:34:33 street. And so because I've got this
1:34:37 omni reference the the dude and the lion
1:34:40 just like in these images should be
1:34:42 there. You can click on this waiting
1:34:44 thing. Um it defaults to 100 and
1:34:47 apparently if you raise this really high
1:34:49 it's bad. So, I don't know. I've been
1:34:51 just keeping it around
1:34:53 200. And then we've got this weird style
1:34:56 reference. And let's try. We'll see what
1:34:58 it
1:35:00 does. The thing about midjourney is you
1:35:03 got to just experiment, experiment,
1:35:06 experiment, experiment, experiment.
1:35:11 [Music]
1:35:25 So, it doesn't seem like it kept the uh
1:35:28 the style. Oh, maybe here. Here it is. I
1:35:31 lowered my style waiting. Yeah, now it's
1:35:33 doing it as a weird drawing.
1:35:37 [Laughter]
1:35:45 The appropriate response for this image
1:35:48 is, "Dude, what the [ __ ] Are you
1:36:02 okay?" So, yes, you can take an
1:36:07 [Laughter]
1:36:16 Okay, we're gonna we're gonna upscale
1:36:18 that
1:36:23 one. Oh my
1:36:32 god. Lordy. We're going to upscale this
1:36:35 one, too.
1:36:40 Oh man. Dude, what the [ __ ] Are you
1:36:43 okay? No. No, I'm
1:36:50 not. Oh, man.
1:37:00 [Laughter]
1:37:08 Did you get that certificate in how to
1:37:09 use midjourney? I don't want to talk
1:37:12 about
1:37:19 it. Oh my
1:37:22 god. [ __ ]
1:37:26 awesome. All
1:37:28 right. 94% complete. Come on. Come on.
1:37:32 Midjourney. Finish it up. Let's
1:37:36 go. All right. That looks a little too
1:37:48 plasticky. I think the original was
1:37:55 better. Maybe not.
1:38:04 [Laughter]
1:38:22 You you you all do know this is just
1:38:26 really about me entertaining myself,
1:38:28 right? You know, that's what the channel
1:38:33 is. I think we need to turn this one
1:38:35 into a uh into a a
1:38:39 video. I think I think we all understand
1:38:42 that. So, we're going to download that
1:38:50 one. Let's see.
1:38:55 [Laughter]
1:39:04 And then we're going to go to Luma Labs.
1:39:07 Is anyone still here? You've all left,
1:39:10 right? There's 45 people here. Sorry.
1:39:22 It's [ __ ] It's [ __ ] midnight on
1:39:24 the East Coast and there's 45 people
1:39:27 here watching me laugh at stupid [ __ ]
1:39:30 Thank you for hanging
1:39:35 out, but we're going to make these two
1:39:37 into um into
1:39:40 videos. If you haven't seen this, this
1:39:43 is how you do it. It's okay, Kyle. We're
1:39:45 talking amongst ourselves. I know.
1:39:48 Please, please
1:39:51 do. Good
1:39:53 lord. Okay.
1:39:58 Downloads. All right. So, man and lion
1:40:03 walk down the
1:40:07 street and then we're going to
1:40:10 do camera
1:40:12 moves. Oh,
1:40:16 effects. Oh, transition
1:40:20 format angle.
1:40:23 Nice. So, this is going to
1:40:26 be crane down, crane up, truck
1:40:31 left, dolly
1:40:34 zoom. Where? Zoom in. Handheld. Zoom
1:40:38 out.
1:40:41 Pan. Tilt. Oh, push pull
1:40:45 out. All right, we're gonna do a pull
1:40:49 out. And then we're going to do the
1:40:51 other
1:40:53 image. And we're not going to do any
1:40:56 prompt other than No, maybe we'll
1:41:02 do I think we'll
1:41:05 do Bolt
1:41:08 Cam. If you haven't seen Bolt Cam, it's
1:41:10 pretty cool.
1:41:13 Oh
1:41:14 man, it's really got Stan alive from the
1:41:17 V begs. Yeah, it definitely
1:41:24 does. There starting to
1:41:28 walk. Diffusion models are
1:41:33 wild. So the diffusion is when you see
1:41:36 it all blurry and then it just gets less
1:41:38 and less blurry. It's basically where it
1:41:40 starts out is this completely fully
1:41:43 noised image that
1:41:45 it recombobulates, you know, thousands
1:41:49 of noisy images into a single average
1:41:51 noisy image and then just dnoises it
1:41:55 over 40 or 50 or 60 or 70 steps.
1:42:06 [Music]
1:42:10 This is going to be cool actually. And
1:42:12 then we'll add we'll add audio to
1:42:24 this. The Pink Panther theme. Well,
1:42:27 we'll just add sound effects to them for
1:42:29 now and then we'll maybe we'll put these
1:42:30 together as some sort of There it is.
1:42:35 So, I'm going to add audio.
1:42:39 Um, New York City
1:42:42 traffic
1:42:45 sounds. Man and lion
1:42:50 walking.
1:42:55 Lion
1:42:57 [Laughter]
1:43:03 purring. Hang on. I got to change my
1:43:05 audio.
1:43:08 [Music]
1:43:27 [Music]
1:43:45 [Music]
1:43:47 It's going to be traffic
1:43:56 [Music]
1:44:09 and
1:44:11 ah yes
1:44:15 archetypal architect using 03 to create
1:44:18 prompts to take to lovable holy crap.
1:44:20 Yeah, that's a perfect use case for it.
1:44:23 Love it. That's great. Cityscape with
1:44:26 purring under. Yeah, it's not doing it
1:44:28 great. But let's see. Let's see our bolt
1:44:32 [Laughter]
1:44:50 cam. Oh man. And all right, let's make
1:44:53 audio here. We'll go um let's see New
1:44:56 York City
1:44:59 um urban
1:45:05 um
1:45:07 noise and then we'll
1:45:10 say swooshes with camera moves.
1:45:19 I'm looking for a man in
1:45:21 finance. 65, blue
1:45:26 [Applause]
1:45:29 [Music]
1:45:34 [Music]
1:45:37 eyes. Try it again.
1:45:45 [Music]
1:45:59 Okay, perfect. So, we download
1:46:05 this. Why does it sound like you did
1:46:08 those yourself? Exactly. So now what you
1:46:11 do
1:46:13 is is you go so so listen again I've
1:46:19 been accused of being not
1:46:21 professional and I think if you look at
1:46:23 my hair the the hair tells a very
1:46:26 different
1:46:27 story
1:46:29 right
1:46:31 so so follow my
1:46:34 advice go make stupid [ __ ] and then post
1:46:38 it for other people to
1:46:39 [Laughter]
1:46:52 Hey, and then you put something like
1:46:56 you're
1:47:02 welcome. Oh no. Is this not going to let
1:47:05 me upload it? Please upload. Please
1:47:07 don't error out. Please,
1:47:11 please,
1:47:14 Elon,
1:47:16 please.
1:47:18 Please. Oh, it's going to error
1:47:22 out. 50%. Come on, you can do it. 60.
1:47:26 You can do
1:47:31 it. Come
1:47:33 on. No.
1:47:36 100
1:47:38 post.
1:47:43 Yes. All right. Go to X. Go find that
1:47:47 post and boost it.
1:48:06 [Music]
1:48:12 [Music]
1:48:13 [Laughter]
1:48:17 They say I'm not professional. Come
1:48:22 on. You know how much money it would
1:48:24 have taken to find a dude that fat? No
1:48:28 comments. to find a dude that fat with
1:48:31 that cool hair. Source those jackets.
1:48:35 Find an actual lion head or a real lion
1:48:38 that you teach to stand on its hind
1:48:40 legs. This is a minimum $10,000
1:48:47 shoot. Gareth laughing out loud. What
1:48:50 did I just come into? You came into a
1:48:52 mental breakdown.
1:49:00 All right, I'm leaving. Y'all are on
1:49:03 your own, but come to the salon
1:49:05 tomorrow. Okay. All right. If you come
1:49:08 to the salon tomorrow, I'll tell you who
1:49:10 my hairdresser is. I'll do it. I'll do
1:49:13 it. All right. Get a lot of requests for
1:49:18 that. A lot of comments about my
1:49:21 professionalism. A lot of requests for
1:49:24 who does your
1:49:26 quaffure. All right, peace out, man.
1:49:30 Hope you had fun tonight. That was so
1:49:32 freaky
1:49:34 sounding. All right, I'll see you
1:49:36 tomorrow. Yeah, archetype. Uh tomorrow,
1:49:39 maybe tomorrow um you can tell tell us
1:49:42 how well your uh 03 written prompts are
1:49:45 doing and lovable. That's kind of a cool
1:49:47 idea. All right, everyone on YouTube and
1:49:49 LinkedIn and X. I will see you later.