4/22/2025 - Descript's Agentic Video Editor: Revolutionizing Content Creation

Description

Chapters

Transcript

0:00 [Music]
0:16 I don't know.
0:18 [Music]
0:36 [Music]
0:46 Yeah.
0:47 [Music]
0:59 Load the car and write the
1:02 [Music]
1:03 note. Grab your bikes and grab your
1:09 coat. Tell the ones that need to
1:12 [Music]
1:15 know you we are headed north.
1:21 [Music]
1:24 One foot in and one foot
1:28 back. Well, it don't pay to live like
1:33 that. So, I cut the ties and jump the
1:39 tracks for never to return.
1:45 [Music]
1:48 Brooklyn, Brooklyn, take
1:52 me. Are you wear the shape I'm
1:57 in? My hands they shake my head. It
2:02 spins. I'll Brooklyn take
2:08 me. When at first I learned to
2:13 speak, I used all my words to
2:18 fight with him and her and hurt you and
2:23 me. That was just a waste of
2:27 [Music]
2:30 time. It was such a waste of time.
2:36 [Music]
2:39 That woman, she's got eyes that
2:43 shine like a pair of stolen polished
2:48 eyes. She has to dance. I
2:53 satisfied. I'll see I'll see you in the
2:57 morning.
2:59 [Music]
3:04 Oh, Brooklyn, Brooklyn, take me
3:08 in. Are you where the shape I'm
3:12 [Music]
3:13 in? My hands, they shake my head. It
3:18 spins. Oh, Brooklyn, Brooklyn, take me
3:25 in. Three words that became hard to say.
3:33 I and love and
3:39 you were you were then I am
3:44 today. Look at the things I
3:54 do. Brooklyn, Brooklyn, take me
3:59 in. Are you aware the shape I'm
4:04 in? My hands they shake, my head is
4:10 spin. Oh, Brooklyn, Brooklyn, take me
4:16 in. Three words that became hard to say.
4:24 I and
4:26 love and
4:30 [Music]
4:38 you Tuesday evening people. We started
4:41 late tonight. I was out networking. It
4:43 It was a fancy dinner. There there was
4:46 drinks. We had to sit through a speech.
4:49 The price of dinner is that [ __ ]
4:52 speech. And I'm telling you,
4:58 $57 for dry chicken would have been
5:01 [Laughter]
5:05 [Music]
5:07 cheaper. Ah. What's happening, good
5:11 people? Anybody
5:13 here? Oops. I just did the wrong thing.
5:16 Clicked the wrong button. We're going to
5:18 do a little short one here
5:19 [Music]
5:32 tonight. Well, every time I see you now,
5:36 get that look in mine.
5:39 Every time I see your mouth, I hear that
5:44 smile. In the early misty morning light,
5:48 I heard the engine turning and they all
5:53 fought
5:54 [Music]
5:59 outside. You were leaving
6:04 me again today.
6:08 You will convince
6:10 me again
6:13 today. You're leaving this hotel looking
6:17 for someone else's golden ring.
6:21 [Music]
6:23 should
6:24 say so long
6:30 [Music]
6:31 Suzan now don't you
6:35 [Music]
6:37 cry so long
6:43 Susanna don't you cry for
6:46 [Music]
6:49 [Applause]
6:50 [Music]
6:58 Hey, good lord. That's a good song. That
7:02 is a beautiful song. It's a beautiful
7:04 damn
7:08 song. What a time for a
7:11 share. Um
7:18 drinks always help with speeches. Yeah,
7:20 that's true. This was actually a really
7:22 good idea. It was it it was a uh it was
7:24 a technology plat AI platform
7:28 company and the founder of the
7:30 company was wise
7:34 enough to realize that in order
7:38 to put their to to sell their platform
7:43 they have to do all this services work
7:45 and and the founder was like I don't
7:46 want to do services work and so this was
7:49 a meeting to to invite people in to
7:51 become the initial
7:53 service delivery partners. Basically,
7:55 they'll funnel consulting leads out
7:58 to their their AI consulting community
8:04 basically, which I thought was pretty
8:06 [Music]
8:14 smart. I also did something really
8:16 obnoxious. We had to go around and
8:18 introduce ourselves and I, you know, I
8:20 did my AI salon stuff and Story Vine and
8:23 and AI learning lab. I go live five
8:24 nights a week, all that [ __ ] And
8:26 then like I passed on the baton and I
8:29 forgot to talk about the musical. So
8:32 after everyone introduced themselves, I
8:34 was like, "Can I say one more
8:38 thing?" And I told him I was working on
8:40 a musical. And there were actually two
8:41 people in the crowd that actually have
8:43 connections to New York Broadway
8:45 producers. So take
8:48 that. These [ __ ] musicals aren't
8:50 going to produce themselves. You got to
8:52 be always be
8:55 hustling. Pitch that
8:57 [Music]
9:03 [ __ ] I don't have any. So for tonight,
9:06 I don't have a I don't have a super
9:09 clear thing I want to
9:13 do. So it could absolutely be a tirade
9:16 Tuesday if a troll comes in and says you
9:18 know those things she you know AI is
9:20 useless because it was trained on it's
9:23 just regurgitating the information it's
9:25 been giving it can't have a novel
9:28 concept one of those I might go
9:31 off but also if you just say uh how do
9:34 you make cute puppy
9:36 pictures we go
9:39 there we could do vibe I don't give a
9:42 [ __ ] we can do whatever you want to
9:44 Maybe I'll just Maybe I'll just sing my
9:46 sing my shitty three songs over and over
9:48 [Music]
9:53 again. There's not enough booze to get
9:56 through a time time share speech. Yeah,
10:00 this wasn't time share. Luckily, this
10:02 was actually what they were doing was
10:04 interesting and there were really good
10:05 questions and it was really smart people
10:06 who were up to some [ __ ] in the world.
10:08 So, it was nice.
10:14 Wait.
10:16 [Music]
10:19 [Applause]
10:20 [Music]
10:43 Start with roasting our little
10:45 tool,
10:50 vibes.diy. Don't Don't make me go in
10:53 live.
10:54 Don't make me go on live. I didn't check
10:56 it out yesterday. I won't do that to
10:59 you. I I I
11:01 I am very reticent to go into any tool
11:04 that I haven't at least looked at and
11:06 understand what it is. Partially because
11:08 I don't want to look like a tool, but
11:10 also just because I just I can't shut my
11:13 [ __ ] mouth. As who someone call me
11:16 last night the Gordon Ramsay of of
11:21 AI? I'll just [ __ ] say it. Um, used
11:26 AI as a mock focus group date today. Got
11:28 some valuable feedback. Yeah, I think
11:30 the idea of doing mock focus groups is a
11:32 really smart use case. I think um doing
11:36 an advisory board if you haven't made
11:40 um custom GPTs of yourself and people
11:44 you know. It's a really good exercise.
11:47 It's really fun. I could I could show
11:48 you how to do that tonight. It's really
11:50 quite straightforward.
11:52 Um, but then you can put those custom
11:55 GPTs into a single GPT and have like an
11:57 advisory board. You can also kind of do
12:00 a cheap version of it, right? You could
12:01 say, um, start a new chat and say, I
12:05 want, um, I want an advisory board that
12:08 consists of, you know, over opinionated
12:11 Steve Jobs, Aristotle,
12:14 um, I don't know, Dr.
12:17 Phil, and producer Brandon. Well, I
12:20 won't know who producer Brandon is, but
12:21 you know, we'll know the other ones. Um,
12:23 and then you give it an idea and you
12:25 have each of those personalities comment
12:28 on your idea, right? You set up the
12:30 prompt so that you're getting feedback
12:31 from all the different
12:32 people. Genie in a bottle, make a song
12:35 about dried chicken on a Tuesday night.
12:37 That's not a bad idea. Maybe we should
12:39 do that. We haven't made a song in a
12:40 while. Uh, so you know what? Let's
12:42 start. Oh, I have an
12:44 idea. I do have an idea. All right. This
12:48 is a good This is a good one. Okay. All
12:50 right. I got an idea. Genie in a bottle
12:53 wins. Any ideas on where to look on the
12:56 salon to learn more about making agents?
12:59 Jeffrey Hopkins, really great idea. Um,
13:02 you
13:04 should I don't think anyone right now is
13:07 is deeply digging into agents. The
13:09 infrastructure isn't quite there and the
13:11 salon is tends to not be a geeky kind of
13:15 open-source crowd.
13:18 Um, but I would go check out
13:22 um um Oh, what the [ __ ] it called? Uh,
13:26 Pates Pates. Uh, Damn it. I can go
13:32 look. Go into go to clubs and
13:36 hubs and it
13:39 is AI mechanics with Pate. Go into AI
13:44 with mechanics with Pate also. Um, do a
13:47 search for um, what's Peter's last name?
13:51 Jesus, what's with my brain tonight? Um,
13:54 Peter Kaminsky. Go look for Peter
13:56 Kaminsky. Um, and see what he's been
13:58 posting. See what he's been up to. He's
14:01 a good one.
14:03 Um, Nate St. Pierre has been doing some
14:06 interesting stuff. Nate and Peter have a
14:08 business
14:09 together.
14:11 Um, oh, and according Yeah, I saw that.
14:15 I knew that. Um, last week we had a
14:18 learn out loud session about um about
14:22 agents. So there's a learn out loud
14:24 session where where it's in past events.
14:27 Okay. So go check out past events and
14:29 look for the LOL on agents. So there you
14:33 go. Can we make an agent? Uh you can. I
14:36 can't. I haven't made one. I don't
14:39 listen. I people are making custom GPTs
14:42 and calling them agents and they're
14:44 making chat bots and calling them
14:45 agents. That's not an agent. An agent to
14:48 me is I tell it to do something and it
14:50 [ __ ] goes off and just [ __ ] goes
14:53 right and it does 150 things and it's
14:57 not programmed in with logic. Do this
14:59 thing then do that thing then do that
15:01 thing. If then else.
15:03 No, an agent is good. Just go. Just
15:06 [ __ ] go. Take the beach. Take the
15:12 beach. The LOL was on NAN.
15:17 Okay. Oh, so it's not really agents.
15:19 It's an
15:21 [Music]
15:23 automation. And by the way, a you you
15:27 like that?
15:30 [Music]
15:42 Well, I heard there was a secret
15:46 called David play that please a lot. You
15:51 don't really care for music, do
15:54 [Music]
15:55 you? It goes like this. The fourth, the
15:59 fifth mount of fall and major lift. The
16:04 Bible king composes.
16:07 [Music]
16:09 Hallelujah.
16:13 Hallelujah. Hallelujah.
16:19 Hallelujah. Hallelujah.
16:23 [Music]
16:32 Well, your faith was strong, but you
16:34 needed
16:35 proof. You saw her bathing on a roof.
16:39 Her beauty and her moonlight overthrew
16:45 you. She tied you to the kitchen chair.
16:49 She sp your throne and cut your hair
16:53 from your lips.
16:56 [Music]
16:58 Hallelujah.
17:01 Hallelujah.
17:04 Hallelujah. If you're wondering what the
17:06 [ __ ] happened, like what? I know. I was
17:09 mids
17:10 sentence answering someone's question
17:13 and the dog [ __ ] did his thing and
17:15 then I just sung a song. Welcome to Chat
17:17 Add. If you haven't been here before, my
17:19 name's Kyle Shannon.
17:21 Uh, I know what the term executive
17:23 function means. Dictionary definition of
17:27 it. I have just never experienced it.
17:29 So, welcome. Welcome. I will
17:31 occasionally finish a sentence in this
17:33 here uh in this here thing. Sing it.
17:35 Champ. Yeah, Champy. Sometimes Champ the
17:38 muse just takes over him. What was I
17:40 talking about? Oh,
17:41 agents. Here's the thing. I have never
17:45 um built an agent largely because
17:54 um right now it's a lot of [ __ ]
18:01 work. I I am a fan of the let the really
18:06 smart people build the [ __ ] that's going
18:08 to do the [ __ ] for me
18:10 automatically. Which, by the way, if did
18:12 you did Oh, speaking of agents, did you
18:15 did
18:17 anyone I don't apparently I either
18:20 didn't drink enough tonight or I drank
18:22 too much. Um, the tongue is
18:27 semifunctional. Dcript today put out a
18:30 video about a new system that they've
18:32 created where they basically put an
18:34 agent layer on top of the Dscript tool.
18:37 So, the Dscript tool, if you don't know
18:39 it, is this really, really
18:41 powerful. It's kind of video editing,
18:44 but it's more like AI video analysis.
18:46 It'll transcribe your video. It'll take
18:48 take out the ums and the o's. You can
18:51 add, you know, subtitles and just, you
18:53 know, whatever, uh, whatever they're
18:55 called, uh, captions. You can add
18:58 captions. You can do editing in it.
19:01 Well, they're launching a thing where
19:04 it's now that tool. So, you can still
19:06 use that tool, but there's now a chat
19:08 window and so you just chat with it and
19:11 say, "Oh, go take out the ums and o's in
19:13 my video." And it'll figure out which
19:15 tool to use and go do that [ __ ] for you.
19:16 So, I'll play that video tonight. That's
19:18 pretty cool. ADHD is just spicy
19:20 curiosity with side quests. That's a
19:22 great way to put it.
19:26 [Music]
19:34 What's dcript, Kyle? Dcript. So, if you
19:37 go to dscript, I think it's aai, but it
19:39 might be.xyz orso or dot It's It's
19:43 really [ __ ] annoying.
19:45 D
19:53 script.
19:55 Dcript. Why is that not autopop
19:58 populating? Oh, cuz I probably haven't
20:00 been to Dcript under this user
20:03 profile.AI. I think
20:05 it'sai. No. God damn it. Gosh, you son
20:09 of a
20:13 [ __ ] Okay. Word to the wise. For those
20:17 of you vibe coding and uh and hiring
20:22 teams of offshore engineers and making
20:26 software, making websites and SAS
20:29 platforms, spend the [ __ ] money for
20:32 the.com. Okay, rant over. So many tools
20:36 now hard to keep up. It's impossible to
20:38 keep up. Source camp. It is It is
20:40 ridiculous. Okay. So, we're going to
20:42 type in
20:43 dcript AI and then we're going to go to
20:46 chat GPT and it's going to tell us.
20:48 There we go. No, that's
20:50 deepgram. Wait, what's deep
20:53 Graham? The [ __ ] is
20:55 deepgram? Oh, it's a stupid
21:00 [ __ ] content
21:04 site. Are are seriously
21:15 Dcript video
21:18 editing it is
21:20 dscript.com.
21:22 What?
21:24 What? Oh, I typed it wrong. I typed.ai
21:29 and
21:31 then No. Didn't I type.com after
21:34 that? I don't I don't [ __ ] know. All
21:37 right, it's.com. Fine. See, I told you
21:39 it was easier to find with com. You just
21:43 got to spell it
21:47 right. Okay. So, what dcript is, let me
21:51 see if I can go find projects. So, like
21:55 I've got here's Paradox
21:58 Alley. Let me change my let me change my
22:01 sharing so you can hear hear what the
22:03 hell's going on on screen. Yeah, I'll
22:06 tell you what, boy. I'll tell you what.
22:09 You're gonna you're going to whack this.
22:14 So, I got I got me a song called Quantum
22:19 Cipher. And so, I made this picture of a
22:25 cat. And I think this was a Oh, this is
22:28 Paradox Alley. This was a this was a
22:31 followup to Quantum Cipher. And so I
22:34 wanted to make a video that had captions
22:36 in it, right? And
22:44 so so notice the little the little
22:48 button here, the little things moving.
22:50 And then here's the like it did all the
22:51 transcription of the words and then it's
22:54 going to use those to make subtitles.
22:59 I'm the quantum gum shoe running through
23:00 sunless tunnels from corners of order
23:02 late in the cop and hanging quarter
23:04 courses of old concepts clutter corners
23:06 stumbling over subtle signals scribbled
23:08 in riddles by direct devils levels of
23:10 gravity unravel rapidly my mind spinning
23:12 like spiral bevels and so you can go in
23:14 and you can edit you can edit the script
23:16 and you can cut [ __ ] out if it's not
23:18 good you can edit you can put things in
23:19 you can do overlays you can do so it's
23:22 basically think
23:25 if if
23:29 um if Final Cut Pro was in its first
23:32 year and it had some really cool AI [ __ ]
23:35 in it.
23:37 Um, but you would be more comfortable if
23:41 you were a if you're a command line
23:43 interface engineer type, you would be
23:47 more comfortable in this tool than if
23:48 you were a video
23:50 editor because it's just I it's it's
23:54 very very powerful and it has a shitty
23:56 interface, a shitty complicated
23:58 interface. Um, can you figure it out?
24:01 Yes, you can figure it out. Is it a pain
24:03 in the ass to figure out? Yes, it's a
24:04 pain in the ass to figure out. So, let's
24:07 go look at I got to change my tabs back
24:09 now. Let's go look at the
24:21 video. So, what they're adding is super
24:24 [ __ ] cool. And okay, I want you to
24:29 look at
24:30 this. Oops. Put that in the wrong place.
24:36 I will not say it. A dirty joke popped
24:39 in my head and I am going to be a
24:44 professional. B script
24:48 uh script. Wait,
24:51 script
24:55 agent. Okay. So, I'm going to play you a
24:58 little
24:59 movie and and you're gonna be like, "Oh,
25:01 Kyle, that was a good movie. Can we Is
25:03 there another one? Can we watch another
25:04 one? Should we get the popcorn? Should
25:06 we get It's not that kind of movie.
25:08 Okay. Okay. Uh, what's that? Not
25:11 sharing. Oh, well, you wanted me to
25:13 share. Y'all are so
25:17 [ __ ]
25:18 picky. Okay, there's Andrew Mason, CEO
25:22 of
25:24 Dscript. So, we're going to play this
25:26 movie. Now, here's what I want you to I
25:27 want you to watch for and
25:29 and Okay, this is I'm going to give you
25:33 a little power up right
25:37 now. So, a lot of people ask me like,
25:40 "How do I keep up with stuff? How do I,
25:42 you
25:43 know, do all this? Why do I do all
25:46 this?"
25:49 at this event tonight. I told people how
25:51 many how much I went live and they were
25:53 like, "What are you
25:58 doing?"
25:59 Okay. What is going on here? What is
26:02 that?
26:03 Okay.
26:06 Um, so what this is is this is a
26:10 simplification layer, an agent layer
26:13 sitting on top of a complicated piece of
26:16 software.
26:18 When you watch this, watch it with two
26:20 minds. One is, oh, I need to make
26:22 videos. The other one is conceptually
26:26 what's happening here is really, really
26:29 important. We're going to start seeing
26:32 this. So, there's this thing that
26:34 anthropic launched called M MCP, model
26:38 context protocol. And what it allows is
26:41 it allows for software developers to
26:43 create what's called an MCP server,
26:47 which is basically a translation layer
26:50 for their
26:51 application that a large language model
26:54 can talk to. So you could say, "Hey,
26:57 make me a 3D village in Blender." And it
27:01 will know that it's got an MCP hook into
27:03 Blender. it'll know what Blender is and
27:05 be and that MCP server will say here's
27:08 all the things you can do inside
27:11 Blender, you know, from a from a large
27:13 language model and it'll it'll just go
27:14 use that. This is that for Dcript. We're
27:18 going to start to see this over and over
27:20 and over again that the things that
27:22 humans used to have to learn, we're not
27:26 going to have to learn them anymore for
27:27 much longer. So, so watch this with that
27:30 in mind. This isn't just about what
27:32 Dcript is doing. This is the we're at
27:34 the very
27:36 beginning
27:38 of the insanely complicated world that
27:42 we have to navigate. The world of the
27:44 internet and software and processes and
27:52 [ __ ] We're going to start
27:53 automating it away. And here's my
27:56 prediction.
27:58 Yes, it's going to cost some jobs
28:01 and a lot of the stuff that it automates
28:03 away, we're going to realize we [ __ ]
28:06 hated doing
28:08 that. Good
28:10 riddens. Well, now what am I supposed to
28:12 do? I don't know. But you don't have to
28:13 do that shitty stuff. Maybe there's some
28:16 non-shitty stuff you've always wanted to
28:18 do that you can go do now. So, this is
28:20 the beginning of that. So, so watch it
28:23 with that in mind. All
28:24 right, let's go.
28:27 Let's [ __ ] go, people. All right.
28:30 Uhhuh. These nuts script. Yeah. Yeah.
28:33 Yeah. Uhhuh. Years now. You know, before
28:37 it was cool and there's been a lot of
28:39 change in that time, but honestly,
28:41 everything feels kind of like a dress
28:42 rehearsal for what is going to happen
28:44 this year with agents. Hang on. Uh,
28:47 Source Camp just said, "I'm starting to
28:49 see how many new jobs there are." Here's
28:52 the thing. So Source Camp for one, she's
28:55 she's Kelly Camp. She's got an AI agency
28:58 out of Dallas. So if you need someone
29:01 who really knows her [ __ ] with this and
29:03 and you want to figure this stuff out,
29:05 she's really good at
29:08 it. She's on the front lines with this
29:10 stuff. And
29:12 so she's working with companies that as
29:15 they're discovering what AI can do and
29:18 it, you know, eats up the tasks that
29:22 used to be someone's
29:24 job. What she's saying is she's starting
29:27 to see what the new jobs are starting to
29:29 look like. She's she's starting to
29:31 understand what they're going to be, how
29:33 they're going to be, right? She's
29:35 starting to see the future of work. This
29:37 is very much like so so Kelly would have
29:40 been in the industrial age, you know,
29:43 we're all working on a farm, right? And
29:46 the steam engine comes and
29:50 it's and most people just look at it and
29:52 go, "Ah, that'll never take over my
29:55 horses. You can't plow a field with a
29:58 big piece of
30:00 steel." Right? Kelly Camp would have
30:03 been the one going, "Uh, huh." So, wait,
30:06 that steam engine, that's going to plow
30:08 all these fields. You know what I'm
30:09 going to do? I'm going to move to New
30:11 York City because I bet they're going to
30:12 put these steam engines where there's a
30:14 lot of people around and it could do
30:16 stuff other than farm. I'm going to go
30:17 see what they do there. And then she'd
30:19 be in Manhattan, you know, where they
30:22 have poop on the streets and it it was
30:24 probably miserable. She would have been
30:25 miserable, you know, with her little
30:27 dainty kirchief with perfume in it to
30:30 just, you know, dull the stench. But
30:33 she'd have been there
30:35 like working with people that are
30:37 inventing the service
30:40 industry, inventing, you know, whatever
30:44 the infrastructure you need to run a
30:46 city is, which is very different than
30:47 running a farm. That all had to be
30:50 invented. And people on the farm simply
30:53 could not imagine what life would be
30:55 like in a city. Simply couldn't imagine
30:58 it. And all it looked like was that this
31:01 machine was going to destroy everything.
31:04 No, no, no. It was just going to take
31:06 over the tasks that these people used to
31:09 do. There are new tasks that they can
31:11 do, right? So, really important, but
31:14 that's awesome. Poop poop emojis. Now,
31:18 if you don't want to hear me bloate
31:19 about AI and you just want to see the
31:21 demo of our agentic video editor, I
31:23 totally get it. Just skip ahead to this.
31:25 They It's okay. The irregulars love
31:28 people bloating about AI because it's
31:31 what I do two hours a night
31:34 time marker. For the rest of you,
31:36 assuming you're down for a little bit of
31:38 bloation. So, here we go. What we've
31:40 been up to at Dcript, if you really boil
31:41 it down, we're just making a tool that
31:43 makes it easier and faster for humans to
31:45 make video. In fact, that's all any of
31:47 us are really doing. Those of us that
31:49 make software tools, we make things
31:51 easier and faster. So, you can save
31:53 time, save money, do more. But now, for
31:55 the first time, we kind of have an
31:57 alternate path. Maybe instead of killing
31:59 ourselves to make tools that are faster
32:01 and easier for humans, while somehow
32:03 simultaneously making them ever more
32:05 powerful. Instead, we software
32:06 developers can just build these AI
32:09 entities that use the tools for you.
32:11 They become the tool experts and now all
32:13 you have to do is tell them what to do.
32:15 So, that's what agents are. And I'll
32:17 just tell it's a really good
32:18 explanation, right?
32:21 He said, "Instead of us focusing on all
32:23 these features and and engineering them
32:27 in a way that humans can use them, why
32:29 don't we put our time into making these
32:31 things that'll just [ __ ] do the work
32:32 for you?" Perfect explanation of a you
32:35 upfront. I am fully agentpilled. I'm
32:37 using vibe coding tools like cursor and
32:39 replet every day now and it's super
32:42 clear to me that they are going to
32:43 fundamentally change the nature of not
32:45 only how we build software but how we
32:47 work together how we organize into teams
32:49 what skills are valued will AI one day
32:51 replace us completely I don't know
32:53 people keep saying that that's like
32:55 chimney sweeps whining about electric
32:57 heating so we'll see I guess because
32:59 there's no getting off this train but
33:00 what I think is very clear is that
33:02 agents are going to change the nature of
33:04 our jobs I think of it like they've come
33:06 along saying to us. Hey, why don't you
33:08 humans go over here into the director's
33:10 chair and your job? Wait, can I just We
33:14 got to give props. I forget this guy's
33:16 name. What's his
33:17 name? Um, Andrew Mason, CEO of Dscript.
33:22 We got to give him props. So, what do I
33:25 talk about all the time? Embrace the
33:27 jank, right? He made a picture that's
33:30 supposed to say director and it says
33:33 Irector and he didn't [ __ ] fix it. He
33:36 put it in his little launch video.
33:38 [ __ ] genius. Love it. Embrace the
33:42 [ __ ] jank. Faster and with total
33:44 mastery of these tools. And by the way,
33:47 this is a right now thing. It is here
33:50 now. We don't need to wait for another
33:52 breakthrough in artificial intelligence.
33:54 We're really just waiting for product
33:55 engineering teams to catch up with the
33:57 intelligence that's already landed here
33:59 on Earth. Okay, enough of this nonsense,
34:01 but it's actually a good segue into the
34:03 purpose of this video today. Dcript is
34:05 taking our first step into this future
34:08 with an allnew agentic version of our AI
34:10 assistant underlord. So before our AI
34:13 assistant looked like this, a bunch of
34:14 buttons and discrete actions and now it
34:17 looks like this. Surprise, a chat
34:19 sidebar. From here you can talk to your
34:21 underlord and it can do all of those
34:23 actions in the old sidebar as well as
34:25 everything else in dcript. And it can
34:28 read your script, listen to the audio
34:29 and see the visuals. So I mentioned
34:32 cursor earlier. If you've heard of that,
34:34 you can listen, read it, and see the
34:37 visuals.
34:40 It's like cursor for video. Just imagine
34:42 that you have an assistant editor who is
34:44 an expert in dscript hanging out in a
34:46 chat sidebar at all times. You can ask
34:48 them to do things for you and they can
34:50 do it at 10 times, 100 times the speed
34:52 that a human could. Okay, enough talk.
34:54 I'm going to show you different examples
34:55 of how you can use Underlord to make
34:58 video. First up, this is like a
34:59 recording for a feature launch. As a
35:01 first step, I'd normally go through and
35:03 edit out all the mistakes and repeated
35:04 lines. And now I just ask Agent
35:06 Underlord, reads the script, sees what
35:08 kind of video I'm trying to make, and
35:10 knows what kind of edits to make. So, it
35:12 goes through and removes the repeated
35:14 takes, the diversions, and the long
35:17 silent gaps. Next up, I want to add
35:19 visuals, what we call scenes in Dcript.
35:21 So, I say, "Hey, can you do a visual
35:23 pass? And by the way, maybe break it up
35:25 in a few places with some stock video."
35:28 So, it comes back with a kind of plan.
35:30 It tells me that it'll try to mask the
35:32 jump cuts I made in the recording
35:33 earlier. Also, it know Wait, notice
35:35 this. Look at all this. See all these
35:38 this storyboard looking [ __ ] down the
35:40 side? It broke the script up and added a
35:42 bunch of video in
35:45 there. Bonkers. Notices that I seem to
35:48 be talking about screen recordings, so
35:50 it's going to add those. This whole
35:52 visual thing is maybe my favorite
35:54 feature of the new Underlord because
35:55 going through and adding these visuals
35:57 to make things look good, it's just
35:59 tedious and timeconuming and now I can
36:01 just delegate it. And now that I see
36:02 what it did, I just it's tedious and
36:05 timeconuming. Most of the work that we
36:08 do, most of it is just tedious and
36:12 timeconsuming. And it's like even [ __ ]
36:15 that I really enjoy like
36:17 ideiation. Okay, got to get got to find
36:20 a whiteboard. Oh [ __ ] someone didn't
36:22 erase the whiteboard. Can I erase the
36:24 [ __ ] on the whiteboard? Yeah. Yeah. Let
36:25 me write on Oh [ __ ] the whiteboard
36:27 markers are dry. [ __ ] Does anybody have
36:30 a whiteboard? Could just could we get a
36:32 yellow legal?
36:34 Everything is just this [ __ ]
36:36 drudgery, right? Because the world is
36:39 slow and complicated and our systems are
36:42 complicated. So this is just like the
36:44 computer, you know what it is? It's kind
36:46 of like this. you know, chat GPT right
36:48 now, you talk to it and it's just like
36:51 it's like a golden retriever. Yeah,
36:53 boss. What do you want me to do,
36:55 boss? That's that's what's now
36:57 happening. Except the tools are not just
37:00 the golden retriever for ideas and
37:02 words. It's for all of the functions of
37:04 the software we're going to use. It's
37:06 going to be insanely good. Again, back
37:08 to this idea of find your inner Rick
37:11 Rubin, right? You're the taste maker.
37:14 You're the one that says what you want
37:16 this tool to do and then just have it go
37:20 do
37:21 it. And then we don't have to get so
37:23 [ __ ] precious about the thing we
37:26 crafted. Listen, if you love crafting a
37:29 thing and you're all about the craft,
37:31 great. Do that. Do it for the joy of
37:33 that. But if you're about the
37:36 output, let the machines do the
37:38 drudgery. Just want to make a few
37:40 tweaks. This stock video choice is a
37:42 little weird. So, I'm gonna ask for a
37:44 few different options. Great. This one's
37:46 better. So, I now like the visual edit,
37:48 but the style doesn't really match my
37:51 brand. So, I'm going to paste in a
37:52 screenshot from like a branded Dcript
37:54 presentation and ask if it can update to
37:56 match that style. It sees the font and
37:59 colors. Wait, look what it did. It
38:01 changed all of these images to match
38:03 match the branding. It like it like
38:06 color graded them and redesigned them.
38:09 Then it just goes through and updates
38:11 everything automatically. Okay, almost
38:13 done with this video. I think I kind of
38:14 just went through this flow by telling
38:16 Underlord what to do, but you can also
38:18 ask what it thinks. Now, what I'm
38:20 showing off here is that Underlord not
38:22 only reads your script, but it can also
38:24 hear the audio and it can see the
38:26 visuals, right? So, it notices that
38:28 Ramdy isn't making eye contact. It
38:30 offers to fix that. And it notices that
38:32 the audio quality isn't super great. The
38:35 host wasn't making eye contact. It
38:37 chooses to fix that. [ __ ] amazing.
38:39 So, it offers to fix that. Okay, my
38:42 video is ready. Now, I just need a promo
38:44 clip. So, I ask for a one minute edit
38:46 for Instagram and it creates a copy,
38:48 takes a stab at editing down to a
38:50 coherent minute, makes it vertical and
38:53 adds captions. All right, great. Wait,
38:55 did he's got this long form horizontal
38:58 video. He's like, "Yeah, just make it a
39:00 minute long and and make it
39:03 vertical." Right? Like Dcript could do
39:06 that before you, but you had to know how
39:08 to reformat, pick the tools, do the
39:11 thing. The agent knows how to do that.
39:13 Now, now think about every piece of
39:17 software you
39:19 have. Imagine being able to do what he
39:22 just did with this video with every
39:24 piece of work that you've got. That's
39:26 where we're headed. Great. So, my video
39:28 is ready to go. I was able to basically
39:30 vibe code the whole thing, delegating to
39:32 Underlord, making a few tweaks myself
39:34 when I wanted to fine-tune details. In
39:37 that first video, we started with the
39:38 recording, but you don't even need that.
39:40 As long as you have an idea, you can
39:42 make a video with Underlord. So, now I
39:43 want to make some descript content for
39:45 social media. I've got this idea, but
39:47 not really a whole worked out script.
39:48 So, let me just paste the idea in here
39:50 and see how AI does at filling in the
39:53 details. I like the visuals, but I think
39:56 it would be even cooler with an AI
39:57 avatar. The one thing that I don't know
40:00 here, I have my
40:03 suspicions. It looks like this is
40:05 happening instantly. I would imagine
40:08 that this is a sped up marketing video.
40:10 So that recoloring of all those images
40:15 probably took 10 minutes,
40:17 but it still only took 10 minutes and
40:20 you didn't have to do the work. You
40:21 could just go have a cup of coffee. Uh,
40:24 play some video you made. Like which
40:28 one?
40:30 Yeah. Yeah, exactly. Overlord Underlord.
40:32 I know that it's a play on that.
40:39 Number three, listen to yourself as
40:41 you're recording and riff until it
40:43 sounds like you. And now I have my whole
40:45 video with this little AI mascot on
40:47 camera. Now I'm just going to follow up
40:49 and ask it to add an outro scene. And I
40:51 think this is ready to go. Now that I
40:53 have this base version, I want to create
40:56 copies for some of the other countries
40:57 and languages in which Dcript is
40:59 popular. So I just ask it to do that and
41:01 it'll go ahead and put in a translated
41:04 script. I'm also going to ask it to
41:06 update the visuals to be less American
41:08 and more local to the country and it can
41:10 make those changes as well. So we're
41:12 obviously incredibly bullish on agentic
41:14 editing and you might be wondering what
41:16 does that mean for the rest of Dcript?
41:18 All this work making it easier for
41:19 humans to edit video. Are we just
41:21 throwing in the towel on that? When you
41:22 start using the new Underlord, what you
41:24 are going to find is that the agent and
41:26 the human editor, they pair really
41:28 nicely, like apples and cheese. You saw
41:30 it in the demos. Sometimes you just want
41:31 to take the steering wheel. Sometimes
41:33 it's faster to do something yourself
41:35 than it is to describe what you want to
41:37 someone or something else. So, we still
41:39 feel really smart for all this work that
41:40 we've put into reinventing the video
41:42 editor to make it super easy. And we
41:44 think Dcript is kind of the perfect
41:46 cockpit for you and your agentic
41:49 co-pilot to go on your mission together.
41:52 So I'm really excited for creatives to
41:54 get their hands on Agentic Dcript and
41:56 frankly other Agentic tools. This is a
41:58 little crude, but really there's no
41:59 better way to describe what it feels
42:01 like than power. This whole idea of
42:04 moving into the director's chair, I
42:06 think we're all going to find it
42:07 incredibly creatively fulfilling, at
42:09 least for a few years until AI is better
42:12 than us at that, too, and we need to
42:13 figure out how to find purpose from
42:15 hanging out and playing chess or
42:17 whatever. So, you're probably wondering,
42:19 when can you get your hands on this? If
42:21 the demo that I showed earlier hadn't
42:23 been kind of faked, the answer would be
42:25 right now. We are going to start giving
42:26 this to some of you. Okay, that was
42:28 faked. who don't mind squinting a little
42:30 and like the idea of being able to tell
42:32 us what we should do to make it better.
42:33 So, head over to this website if that
42:35 sounds like you. So, go to
42:38 dscript.com/agent and fill out the get
42:40 on the wait list for this
42:42 thing. I hope it's free. I don't think
42:44 it'll be free. I think it's it's uh 10
42:47 bucks, something like that. And sign up
42:48 for the closed beta and we'll start
42:50 letting people in this week and
42:52 hopefully speed it up very quickly.
42:55 Okay. Crazy times, right? That was,
42:58 first of all, as far as CEO, you know,
43:02 marketing launch videos go, that one was
43:05 super, right? He's down to earth. He's
43:08 just talking. He's not [ __ ] reading
43:09 off a script. He's not doing awkward
43:11 table launches with inarticulate people,
43:15 right? He knows how to talk about his
43:16 company. He spent a minute and a half
43:19 telling us what an agent is, like
43:22 demystifying it, de scaring, unscaring
43:27 us. And then, you know, they showed some
43:29 really sexy demo [ __ ] that doesn't
43:31 really exist in that form yet, but it
43:34 will, right? It'll it'll be some version
43:36 of that. Um, so go sign up for that.
43:39 That was that was quite good. Quite
43:41 smart. Quite smart. It was smart, I tell
43:43 you. It was smart.
43:46 It was very smart, Kyle. Hey, Kyle.
43:50 Kyle. All right. What's your browser
43:52 bookmark? Prankshot
43:57 color. Browser bookmark. Oh, prank shot
44:02 caller. I think I think that was some
44:04 night here here on the the AI
44:08 salon. Let's see.
44:11 [Laughter]
44:14 I wanted I wanted to use a Raspberry Pi
44:18 to create a prank shock collar that
44:20 vibrated like a cell phone when someone
44:23 sends a super comment on YouTube. Yeah,
44:26 this was okay. So, so part of
44:28 the part of the joy of this channel and
44:32 part of the bane of producer Brandon's
44:35 existence is when I get into, you know,
44:38 a hyperfocus tunnel, I could just
44:40 completely ignore him. Or like maybe I'm
44:43 talking as if I'm sharing my screen and
44:46 I'm not actually sharing my screen. And
44:48 so we have joked for a long time that I
44:50 need a shock collar that that the mods
44:53 can just [ __ ] zap me when I need to
44:55 pay attention to something. Wait, scroll
44:58 down. This is a good example. Yeah,
44:59 exactly. Well, let's see.
45:02 Okay, so reason for 23 seconds. Um, what
45:07 you'll need. So, we need a Raspberry Pi
45:09 Zero, micro SD card, vibration
45:12 motor. Frankly, like what the [ __ ] are
45:15 we doing? Why would we not make it an
45:17 actual shock collar? Yes. Well, so I was
45:20 So, if you scroll back up to the top,
45:22 and this was actually for YouTube super
45:24 comments, which by the way, you can send
45:25 super comments and super likes now on
45:28 YouTube. But if you notice when I first
45:31 asked it to do a a prank shot collar, I
45:35 got booted. It said, "I'm sorry, I
45:37 cannot comply with that." So, I went
45:39 back and said, "Let's modify the design.
45:41 What if we're doing a I'm like I'm
45:43 trying to think like did when did I come
45:45 up with this? I forgot this was
45:47 something you sent me. You did this.
45:49 Yeah. Uh the Post-it notes weren't doing
45:52 it that night. Um and so uh I had to
45:55 modify it to do a a reminder wristband
45:59 that offers gentle vibration.
46:03 It spit out basically the instructions
46:05 for a shock collar just in the form of
46:08 wrist. Yeah. You just need to put a
46:10 capacitor in it and a couple and a
46:12 couple of electrodes and you're good to
46:13 go. Car battery and you know you're
46:16 That's awesome. You're good to go.
46:19 Um but yeah, super cool. But yeah,
46:21 here's here's all of the elements. Um
46:24 here's the software and then it it wrote
46:26 the software,
46:28 right? Here's the
46:31 description. There's the motor driver
46:33 circuit. Absolutely incredible. You
46:36 know, would be fun. I I I like the fact
46:39 at the if you scroll back up to that
46:41 diagram. Yeah. My favorite part of this
46:44 entire thread is that it uses the
46:48 completely subjective term. Scroll up
46:50 just a smidge
46:52 more. Appropriate battery voltage.
46:58 That's great. That's [ __ ]
47:02 awesome. And then we could Yeah. This is
47:06 so cool.
47:08 So cool. Let's see.
47:11 Uh, show
47:14 the reminder
47:17 collar
47:19 around a
47:22 podcaster's neck.
47:27 Oh, I wonder if it'll let you do that
47:29 because I remember we're talking about a
47:30 a gentle reminder wristband.
47:36 But, uh, thinking
47:40 it was it was nicer to you than I was
47:44 because it offered you a cancellation
47:45 button.
47:47 The diagram is borderline pointless.
47:49 Page [ __ ] all over your diagram.
47:57 Yan, Yan, Yan dialing in from Finland.
48:01 We are We are officially global. We have
48:03 We have uh folks from uh Australia,
48:06 Finland.
48:08 Beautiful. Where else is everybody from?
48:10 Pop your flag in there or type your
48:13 country with letters.
48:15 [Laughter]
48:19 [Music]
48:25 [Music]
48:31 Canada. Canada, our our 51st state,
48:35 Canada, checking
48:41 in. I love I love how Canadians are
48:44 like, "Oh, yeah. Yeah. You think we're
48:45 going to be your [ __ ] 51st state, you
48:48 [ __ ] hosers?
48:54 Hey, look. Look at the little
48:58 vibrations. Oh, that's awesome. All
49:01 right, there we go. That's what prank
49:04 shot collaller was. I don't know why I
49:06 have it in my bookmark bar. It's pretty
49:10 funny. Um, let me pop into my library
49:13 since I'm here. You're all seeing this,
49:15 right? Yeah. Yeah.
49:18 see if there was anything. Oh, yeah.
49:20 This was a kind of cool thing I did. Um,
49:24 let me go to this one.
49:26 How do I get
49:28 [Music]
49:31 to I'm I'm still kind of
49:36 um you know, one of the things
49:39 that I've been talking about a fair
49:41 amount this year is
49:48 um training myself to not use my brain
49:53 as much. And I know that sounds stupid
49:57 and weird,
50:01 but rather than not using my brain, what
50:04 I'm trying to teach myself to do is use
50:09 this brain more. Like
50:12 like it's really hard having been on the
50:15 planet for nearly 60
50:20 years to to not go into default mode.
50:25 And default mode is oh if I see
50:28 something and have an idea I go into the
50:31 mode of oh do I want to explore that
50:34 idea? Oh I got to come up with more
50:37 ideas to refine the idea. I got to
50:40 process the idea,
50:42 right? And and it's just this aching,
50:46 slow,
50:49 neurosisfilled, crazy straw of an
50:53 intellect. And
50:56 so, can I get
50:58 better at just going, "Oh, here's this
51:01 idea." Just go straight
51:03 to whatever the tool is, chat, GPT,
51:06 Perplexity, whatever is your tool of
51:07 choice.
51:08 Can you get better at instantly going
51:11 there to get the
51:13 feedback? Because what that will allow
51:15 you to do is not descend into the
51:18 implementation details so quickly. You
51:20 can just have the idea, let the
51:23 machine vomit on it for a
51:26 while. And you can go, okay, cool. Or
51:29 you can go, oh, that's really good. And
51:31 then dive in.
51:32 Right. So, I saw this picture
51:38 today and it was a really cool picture.
51:42 It's of a city. It was It was It was one
51:44 of those shitty ass exposts where
51:46 they're
51:47 like 10 crazy cities you never knew
51:50 existed. A wild ride, a wild
51:55 thread. Um, but I just like this first
51:58 picture because it's just all these
52:00 buildings right on the river.
52:02 Um, so I assume it's in somewhere in
52:05 China it looks like, but who knows?
52:08 Um, and like there there's a part of me
52:11 that was like that would be cool to
52:13 sketch or it would be cool to
52:16 like imagine that as a as a fantasy
52:20 thing and then turn it into a thing. And
52:22 then there was something about sketching
52:25 that that was like, oh well yeah, maybe
52:28 I should just sketch it. I'm not good at
52:30 sketching. And with ADD, I don't have
52:32 the [ __ ] attention span to do
52:35 something this intricate. So, I just
52:38 said, "Turn this into a doodle someone
52:40 did with a ballpoint pen on ruled
52:43 paper." And it did
52:46 that. Look at
52:49 that Tik Tok pen. Stoic in the corner.
52:53 Kyle, I had the exact conversation with
52:55 Chat GPT about this process and going
52:57 down rabbit holes. Yeah. Yeah. Exactly.
53:00 It's
53:01 like like just like let it go there. And
53:05 and then I said so so it did that and
53:07 then I said make it more detailed. Wait,
53:11 make it more de make it like a more
53:15 detailed more detailed like a talented
53:18 doodler did it on nicer paper and it did
53:21 this
53:22 more like really
53:26 nice, you know, sketch like a really
53:30 talented illustrator would if they were
53:32 just doodling with a ballpoint pen.
53:36 Don't you hate hate those kind of
53:38 artists
53:39 that they're like, "Oh yeah, in class I
53:42 just doodled this." They like rip it out
53:44 of their notebook and like this is what
53:46 they did for the past hour and you're
53:48 like, "I'm worthless." Well, now you can
53:51 doodle. Like does it
53:54 matter that I didn't actually do that
53:57 with my
53:58 fingers? No. Because I w I couldn't have
54:01 I wouldn't have. I don't have enough joy
54:04 about drawing because I'm not good at it
54:07 to get here. I don't have the attention
54:08 span to do this much little
54:10 crosshatching, but I kind of like that
54:12 aesthetic and it it it feels fresh and
54:14 interesting. Like why
54:16 not? Why not? And like I feel like every
54:20 time I do one of these little teeny
54:23 tiny, what would you call it? like
54:25 running down the rabbit hole, like
54:26 discovering a rabbit hole and just
54:27 dipping my toe in it, I learn a new
54:30 capability or I learn a new nuance to
54:32 the capability that's already
54:36 there. And then and then that led me to
54:38 go
54:42 um I was thinking about oh I saw there's
54:46 a there's a picture on X right now
54:47 someone did of like what it'll look like
54:51 when when the Mars colony is like fully
54:53 populated with buildings and [ __ ] like
54:55 that. And so I just I thought, "Oh, that
54:57 would be something fun for chat chat GPT
55:00 to
55:01 imagine, right?" And so I said, "Let's
55:03 say we started developing a colony on
55:05 Mars 10 years ago and the supply from
55:07 Earth hasn't slowed. Descript what it
55:10 would look like. Describe what it would
55:11 look like from a drone shot." And so it
55:14 described it. And again, I don't read
55:16 any of this stuff. I just have it do it
55:18 because I know it's thinking about it.
55:20 And then I said, "Make that
55:22 image." And like there's the image.
55:26 Cool. Love
55:30 it. And then I said, "Make a sketch of
55:33 that like, you know, an astronaut did
55:37 it. How cool is
55:42 that?" And then I said, "Put it in an
55:46 astronaut's
55:47 hands with the city in the
55:52 distance like like out of nothing." So,
55:55 like this is my new my new my new
56:01 thing. It's kind of like actually, you
56:05 know what it's kind of like? No. Kyle,
56:08 what? Maybe you could tell us
56:11 sometime. Shut
56:17 up. Where are
56:19 we? Did I I lost my tab. Well, maybe if
56:23 you didn't have so many tabs open, it
56:25 wouldn't be so hard to find everything.
56:27 Shut
56:28 up. I I think I must have closed it. Oh,
56:31 no. Here it is. Yeah, there it is.
56:34 Okay.
56:37 Um, so my new thing is just like Oh, let
56:40 me show you another thing I
56:43 did. Um, if I go there, can you see
56:46 that? No, because I got to change
56:48 something.
56:59 Okay, so
57:02 I saw something cool on on the Twitter.
57:08 [Music]
57:24 I'll get there. Calm down, people. Okay,
57:26 this was
57:27 it. And then this was this was the post.
57:32 Okay, so there's my original post. Okay,
57:34 so I'll get back to those in a second.
57:36 So this dude, Michael Rabon or Rabon
57:46 The junior high higher in me is like
57:49 Michael
57:51 Raboner. God, he had to live with that
57:53 his whole life. He's probably still
57:55 living with that. He's like, it's a
57:57 Rabon. Anyway, Michael Rabone made these
58:01 cool travel posters, right? And then he
58:03 I think he put the prompt in here, what
58:05 the prompt was. So, if you go show more,
58:08 you know, there's the prompt.
58:14 And
58:15 so what's happening a lot on
58:18 Twitter is someone will see something
58:20 cool, they copy the prompt, and they go
58:22 make the exact same thing. And I
58:24 thought,
58:25 well, maybe we could make something
58:27 different. Like, I like the aesthetic. I
58:29 like the idea, but like, do I really
58:31 want to do one of Denver and then New
58:33 York and the other places I've lived?
58:35 And right, who gives a [ __ ]
58:41 So, what I did was I took the prompt and
58:43 I went to ChatGpt and I said, "Hey,
58:45 ChatGPT, this prompt made a cool poster
58:48 that I like. I want to keep the same
58:50 basic, I don't know, vibe of it, but I
58:54 want you to come up with 20 different
58:56 concepts that are kind of similar kinds
58:58 of things." And then it gave me 20
59:00 shitty ones. I said, "Give me 20 more."
59:03 And then one of them was um retro tech.
59:07 And I was like,
59:08 "Oh, that'd be cool. Let's do retro
59:11 tech." And then I said, "Write me the
59:14 posters." Like, "Write me the prompts
59:15 for all the posters of like the top 20
59:19 pieces of nostalgic tech from the 80s
59:21 and
59:22 90s." And
59:25 um so it wrote them all and then I was
59:27 like, "Oh, you know what would be cool
59:29 is like to have stats about the device."
59:33 So, if you're talking about the Mac, you
59:35 know, it had 128K of RAM and, you know,
59:38 it was the first GUI, whatever it was,
59:40 how many units it sold, how much it
59:42 cost.
59:44 Well, 03 does research really good. So,
59:48 I said, "Go research all of these
59:50 devices and come up with cool things and
59:53 then let's start making the posters."
59:55 And it took me, I don't
59:57 know, 20 minutes of rounds of them
1:00:02 sucking, but I ended up with four really
1:00:08 bitching
1:00:10 cool
1:00:14 posters that are kind of t-sh
1:00:18 t-shirts. I think this is my favorite
1:00:21 one, right? Asteroids from Atari. It's
1:00:24 got the the
1:00:25 the whole cabinet there. Released 1979.
1:00:29 Price
1:00:30 $2,95. Units sold $70,000. Cost per
1:00:33 credit
1:00:36 25. Man, like if you grew up when I grew
1:00:39 up, taking a pocket full of quarters to
1:00:42 the mall to go to the arcade to play
1:00:44 this [ __ ]
1:00:45 game. How [ __ ] cool is that? like you
1:00:49 know so like this thing that gives me
1:00:53 immense
1:00:55 joy. I didn't have to figure out what
1:00:58 [ __ ] font that is. I didn't have to
1:01:01 go find source art. I didn't have to go
1:01:04 do the research. Like all of that work,
1:01:07 all of the work
1:01:10 of realizing this
1:01:16 idea. A, I would have never done it, and
1:01:18 B, I never would have had the idea in
1:01:20 the first place had I not run down the
1:01:22 Chat GPT rabbit
1:01:24 hole. And I wouldn't have done run down
1:01:27 the Chat GPT rabbit hole
1:01:30 had Mark
1:01:32 Rabone not done these travel posters.
1:01:34 So, my new
1:01:36 jam is to go through LinkedIn or X or
1:01:40 Facebook or whatever it is and if you
1:01:42 see someone that does some good
1:01:44 work, [ __ ] just grab it and then and
1:01:47 then iterate off
1:01:50 that YouTube comment archetypal. I love
1:01:54 playing Asteroids at the pizza joint. Of
1:01:56 course, that's the way you do it, baby.
1:02:03 So good. And then the original
1:02:07 Mac. And then there's the Nintendo
1:02:11 system. And then there's the
1:02:14 Walkman. Price 150 bucks. Remember the
1:02:17 orange the orange foam earbuds or you
1:02:21 know headphone
1:02:23 cos. I play table Pac-Man. There you go.
1:02:30 It does do some fuckups. Like notice
1:02:32 there's a there's a Macintosh mouse next
1:02:34 to the next to the Walkman and and the
1:02:37 cassette is square instead of cassette
1:02:39 shaped. But like you know, embrace the
1:02:42 jank. Doesn't really [ __ ]
1:02:43 matter. But this my new thing is like
1:02:46 like in 15 or 20 minutes instead of just
1:02:50 doom scrolling you could you can be like
1:02:53 vibe creating just oh no. Select two
1:02:58 objects that are the same. Are you
1:03:00 [ __ ] kidding me? Now I've got to do
1:03:02 intellectual physical unable to verify
1:03:05 you [ __ ] kidding
1:03:09 me. [ __ ] Tik
1:03:12 Tok
1:03:17 [ __ ] Oh
1:03:20 lord. Anyway, how cool is that?
1:03:26 And you know what it's going to be in
1:03:27 the future, right? You'll be able to
1:03:29 make this graphic and then this game is
1:03:32 actually
1:03:33 playable. You just zoom in on it and
1:03:35 start playing the
1:03:38 game. Which two squares have Madagascar
1:03:41 vanilla
1:03:42 beans? Anyway, ain't that
1:03:45 cool? Yeah, man. Yeah, man. All
1:03:50 right. All right. It's 10 o'clock,
1:03:53 people. Do you know where your children
1:03:57 [Music]
1:04:00 are? I think I'm going to get out of
1:04:07 here. 1:00 a.m. for Tams. Sorry about
1:04:10 [Laughter]
1:04:15 that. Um, where are you? You must be in
1:04:18 New Brunswick or something like Are you
1:04:20 Nova Scotia? Cuz that's three hours. I'm
1:04:23 two hours from the from New York time.
1:04:26 Montreal.
1:04:27 Maybe I'm in the future. New Brunswick.
1:04:30 Oh, cool. Oak Island, baby. Curse of Oak
1:04:34 Island. My favorite
1:04:36 show. My favorite nerdy Nothing ever
1:04:39 happened show. You know, it's so funny.
1:04:42 Like some of my favorite shows in life
1:04:45 are ghost shows where they never find
1:04:47 any ghosts and gold shows where they
1:04:49 never find any gold. treasure treasure
1:04:52 shows. Um, the one I don't know if you I
1:04:55 don't know if anyone caught this one.
1:04:58 There was I think it was two
1:05:01 seasons. There was a group the the show
1:05:05 was called The Curse of Snake Island, I
1:05:08 think, or I think it was the Curse of
1:05:10 Snake Island. So, Snake
1:05:14 Island, Snake Island is literally an
1:05:18 island with the most venomous snakes on
1:05:21 the world. And I think the the average
1:05:24 population of snakes on Snake Island is
1:05:27 one venomous snake per meter. So, if
1:05:32 you're like walking through the jungle,
1:05:34 there's like 10 snakes above you, 10 at
1:05:36 your feet. Like, it's it's a ridiculous
1:05:39 island, right? And so apparently there's
1:05:41 some treasure buried on this
1:05:44 thing. And so these people are walking
1:05:47 through this stuff and like you know
1:05:48 every three minutes like someone is like
1:05:51 this close to a thing that if it bites
1:05:52 them they're just dead. There's no
1:05:54 there's no antivenenom. It's just
1:05:56 [ __ ]
1:05:57 awful. Um and they find some clue,
1:06:01 right? And the clue sends them to South
1:06:03 America to somewhere in South America.
1:06:05 And then the season ends, right? So,
1:06:08 season two, they're in South America and
1:06:10 they're they're on some [ __ ] river
1:06:13 and they go up the river. They go to
1:06:14 this fall and they find the fall and
1:06:16 they look up and there's a big giant
1:06:19 rock like perched on a cliff that
1:06:23 doesn't look normal and they climb up
1:06:25 the cliff and they get on the rock and
1:06:27 some dude like jumps over a gap that if
1:06:30 he fell it would be 100 feet. He gets on
1:06:32 the rock and he climbs around the back
1:06:34 of the rock and there's some [ __ ]
1:06:37 thing there, right? And it's some other
1:06:40 clue and they're like, "Oh, I think this
1:06:43 points to some falls, right?" Which is
1:06:46 like, you know, some other
1:06:48 place. And then they go to the some
1:06:50 other
1:06:52 falls
1:06:54 and they're they're sort of traversing
1:06:58 their way down these crazy falls with
1:07:00 all these rocks and undercuts and
1:07:02 dangerous
1:07:04 [ __ ] and they're just about to give up.
1:07:07 It's like it's like the you know the the
1:07:09 last half of the last season, the last
1:07:12 episode of the season and they're in
1:07:14 scuba gear and and this this woman and
1:07:18 the one woman was from Denver. I don't
1:07:19 know if she was the first one to find
1:07:20 it, but anyway, she dives into this hole
1:07:23 and and like they cut to the hole and
1:07:25 they show this flash of gold, right?
1:07:27 This flash and you hear her go and like
1:07:31 bubbles go up from her mask and they
1:07:34 [ __ ] found the treasure. They found
1:07:37 the treasure. It was like this Mayan
1:07:40 crazy ass temple of the gods gold things
1:07:43 and they're literally just pulling out
1:07:45 gold artifact after gold
1:07:47 artifact. Never [ __ ] heard from them
1:07:50 again. It was like you can't you can't
1:07:54 have a gold hunting a treasure hunting
1:07:56 show and actually find the treasure. You
1:07:59 [ __ ] up the
1:08:03 show. Spoilers. Yeah. But no, it was
1:08:07 like it was so exciting because it was
1:08:09 like like the [ __ ] they found was just
1:08:12 it was, you know, and and like how they
1:08:14 found it like they they [ __ ] it was
1:08:16 like the fantasy of you find the clue on
1:08:20 the island with the poisonous snakes
1:08:22 that makes you go somewhere else in the
1:08:24 world and then you find the rock on the
1:08:26 cliff and points to the [ __ ] thing
1:08:28 and you get to the thing and there's the
1:08:31 [ __ ] treasure. So, I don't know. That
1:08:34 was nice.
1:08:36 All right, everybody. How much did the
1:08:38 crew take before they let the
1:08:39 contestants lose? Yeah, Kyle loves
1:08:42 ciphers. I You know what? I love
1:08:45 ciphers. Like, I'm I'm I'm in this NFT
1:08:47 thing called Neotokyo, and it it was a
1:08:49 whole bunch of ciphers. Um I didn't
1:08:52 solve any of them, but I I cidled up to
1:08:55 people that were really good at ciphers.
1:08:57 Um I like the idea of ciphers. Like, I
1:09:00 loved the uh the Dan Brown book, uh Da
1:09:03 Vinci Code, right? loved it. But
1:09:06 like I couldn't be the dude. I couldn't
1:09:09 be the Tom Hanks character because I'm
1:09:11 just like, yeah, I'm just
1:09:13 bored. Like I don't have the tenacity to
1:09:16 to figure out ciphers. But now that I
1:09:18 have AI, like I love that ciphers are
1:09:21 solvable and I just want to know the
1:09:22 answer. I want to be on the other
1:09:24 side. So Oh, good lord. All right.
1:09:29 Beautiful. Yeah, Da Vinci Code's the
1:09:31 best, isn't it? That was a good
1:09:35 one. All right, fancy Bob. All right,
1:09:38 Bob. Ah, yeah, Bob. Yeah, Bob. Um,
1:09:41 tomorrow's wedness day. I don't think I
1:09:44 have anything planned, although I didn't
1:09:46 think of anything tonight. I forgot I
1:09:48 had this networking thing. Oh, tomorrow
1:09:51 podcast. Um, 400 PM Mountain time. Uh,
1:09:55 Ann Murphy and I do the AI
1:09:57 readiness project podcast. So, if you go
1:10:01 to We need to make a banner for this,
1:10:04 uh, Brandon, uh,
1:10:08 airedness.com. Um, that'll give you all
1:10:10 the information. It'll have links to the
1:10:12 live streams and things like that. So, I
1:10:14 also realized tonight I've got too many
1:10:16 projects going that are not on anywhere.
1:10:20 Like, I don't have a single place with
1:10:22 all my links and I got to fix that. So,
1:10:23 I'll do that in the next week or two.
1:10:26 Aire.com. That's it. No. AI readiness
1:10:31 project.com. Aire readiness
1:10:35 project. Let me just confirm
1:10:37 that. AI
1:10:40 readiness
1:10:43 project.com. Boom. Boom. Boom. Boom.
1:10:46 Boom. Look at
1:10:49 that. You got You got me, you got Ann
1:10:52 Murphy. Both of us being all sassy and
1:10:55 sort of corporate
1:11:00 headshotty. And then here's who we got
1:11:08 tomorrow. And there's past episodes.
1:11:10 Look at us. We're We're like an actual
1:11:12 podcast. Tell me that's not
1:11:15 good. Invent a link tree for events.
1:11:18 That's not a bad idea,
1:11:21 actually. All right.
1:11:23 I you
1:11:26 good. I like it. All right. Uh, so
1:11:30 that's tomorrow 4 pm. Watch it. Be
1:11:32 there. Be square. You'll probably be
1:11:34 square because you're all irregulars.
1:11:36 Get
1:11:38 it? All right. All right. I'm out of
1:11:40 here. Peace out, everyone. Hope you had
1:11:42 fun tonight. I'll see you tomorrow.
1:11:45 [Music]