4/1/2026 - Building a Live Simulation of a World Where Work No Longer Defines Us

Description

Chapters

Transcript

0:00 All right, champ. You ready? Are you
0:03 ready for zone football? Yeah.
0:08 Good doggy. Good doggy. What's
0:11 happening? Good doggy.
0:16 What is going down?
0:48 Come on.
0:53 Come on.
1:08 So here's a quiet town
1:11 where the old trees sway
1:15 the town
1:17 that goes through the day.
1:26 Little te
1:28 little day.
1:46 Woohoo.
2:15 We got tick tock rising star Danielle in
2:18 the house.
3:04 Girl, you're looking
3:09 finding night.
3:14 Every fell got you in his sight. Would
3:21 you do it with a clown like me?
3:27 Surely one of life's little
3:30 mystery.
3:32 So tonight
3:34 I ask
3:37 the stars above.
3:41 How did I ever win your love?
3:47 What did I do?
3:51 What did I say?
3:56 Turn your angel
4:00 my way.
4:17 Hey.
4:32 Woohoo.
4:44 So, I learned a new song for you
4:46 tonight.
4:48 April fools.
4:59 Every time I see you now, get that look
5:02 in mine.
5:05 Every time I see your mouth, I hear that
5:09 smile.
5:12 In the early misty morning light, I
5:15 heard the engine turning in the old foot
5:20 outside.
5:26 You were leaving me
5:29 again today.
5:32 You will convince me
5:36 again today
5:39 leaving this hotel looking for someone
5:43 else's golden ring
5:48 to say
5:50 so long.
5:56 Hush now don't you cry.
6:02 So long.
6:08 Don't you cry for me.
6:20 Chilling jeans and cigarettes and
6:22 keeping warm out on the road.
6:25 Chasing down a lifestyle out on Highway
6:29 24.
6:31 New York state was rolling breeze in the
6:35 sunshine and the blue sky falling. The
6:38 chill of old September creep
6:45 you leaving me
6:48 again today.
6:51 You convince me
6:54 again today.
6:57 You're leaving this hot tub looking for
7:01 someone else's golden ring.
7:06 Should I say
7:08 so long? Susanna,
7:14 don't you cry for me?
7:31 Happy Wednesday everybody. Hope you're
7:33 doing well on this fine Wednesday night.
7:37 It's rainy here in Denver. A little
7:39 drizzly, so it's got a it's got a
7:41 Portland vibe to it tonight. Um, not
7:44 like that here. It's normally sunny,
7:48 but what are you going to do? We need
7:50 some rain.
7:53 We got no snow this winter. None. None.
7:55 None. No, one little snowstorm.
9:13 Oh, by the way, we don't have producer
9:14 Brandon tonight.
9:23 So, it's going to be me and me only. And
9:25 so, for you people on TikTok, just yell
9:28 at me if I get it wrong
9:31 with ye old
9:52 Um, let's see here. Is anybody
9:55 commenting on the YouTube? It is
9:58 possible. No. Do me a favor. If you're
10:00 on YouTube or one of the other desktop
10:02 platforms, just comment
10:05 comment so that I can see comments are
10:06 work working over there.
10:13 Should
10:16 we call the local pope if the black
10:19 screen needs adjusting? I guess you'll
10:21 just have to tell me and I will have to
10:22 do my best to to make it make it happy.
10:36 AI really is a bubble. It's too late.
10:38 I'm already addicted. Right.
10:59 Is this place I can rest my forhead?
11:06 Gather my thoughts in sweet silence.
11:12 Is this place where the feelings are
11:15 dead?
11:18 Running over exposure to violence.
11:22 Is this place I can slowly face. The
11:26 only one I truly can know.
11:30 These are tears from a long time ago.
11:34 Got these tears from a long time ago. I
11:37 need to cry. 30 years or so.
11:42 What's happening on YouTube over here?
11:52 Got a lot of rising stars over on Tik
11:54 Tok.
12:29 Boom.
12:34 Down.
12:43 All right, let's get this show rolling,
12:50 rolling, rolling, rolling. Uh if you're
12:52 new here, if you haven't been to uh to
12:56 AI Swan today,
12:59 um head over there and go look at
13:04 the um I made a post in community feed
13:10 about the Jack Dorsey article. I'd love
13:11 to start a conversation over there. Um
13:15 so let me see. Let me go. Let me reload
13:17 my page here and then we'll I'll maybe
13:20 share this.
13:29 Let's ask for help.
13:37 All right.
13:40 Beautiful. Got a couple comments in
13:43 there. So let me share my screen.
13:47 Let me share my screen.
13:50 So there you go. So this is the article.
13:52 So if you go to the AI salon, if you're
13:55 new here, don't know what that means.
13:58 I will show you here shortly. Is this
14:01 it?
14:03 No. Is this it? No.
14:12 No. Let's say that. Yeah, there you go.
14:16 Oh, no. That's Festivus updates. Where?
14:21 Where's my graphic?
14:28 Yeah, there we go. Beautiful. Okay,
14:30 here. Let's You know what I'm going to
14:31 do? I'm going to move that up
14:34 so I'm not always looking for it.
14:37 Then we're going to rename it
14:40 AI salon.
14:45 Wait, but I didn't rename it, did I? Ah,
14:53 beautiful. Okay, head over to
14:56 community.thesalon.ai
15:00 and
15:05 black bar got access to the article.
15:08 Yeah, I changed I changed the article,
15:10 Cam the link in there now, but haven't
15:12 finished reading it yet. Okay, cool.
15:13 Yeah, I put the I put the new non uh
15:16 non-paywalled version into the article.
15:20 Um,
15:22 and that should make it easy to find.
15:25 So, go over to community feed. It's the
15:28 second post down there.
15:30 And you can you can read the article
15:33 tonight. This is what I read last night
15:35 during the live um was this article and
15:38 I just think it's worth us starting a
15:40 conversation over on ye old AI salon. Um
15:45 because yeah um let's let's dive in. If
15:50 anyone has anything they want to talk
15:51 about feel free to pop that in comments.
15:54 Get rid of that thing.
16:01 just commenting Wednesday like what's
16:03 happening over here on YouTube. Hey,
16:04 hey, hey. AI really is a bubble black
16:07 bar. I know, I know, I know.
16:10 There we go. All right, we're getting
16:12 there. We're getting there producer free
16:14 tonight.
16:16 Is anyone going to copy the leaked claw
16:18 code? Uh, it got copied 42,000 times in
16:22 three hours.
16:24 It got forked 42,000 times in three
16:27 hours and then someone posted it to the
16:29 to blockchain. So it will that code will
16:32 never not be available.
16:36 There's an interesting uh video from
16:39 Greg Brockman talking about the new
16:41 model. We can watch that. Let's watch
16:42 that.
16:44 Um let me close this. Let me do this.
16:47 Let me do this. Let me do this.
16:50 Let me do this. Okay. So it all it all
16:53 looks the same to you, but it is
16:54 completely different.
16:57 Okay. Black bar.
17:00 That's not Greg Brockman. That's Leonel
17:04 Messi. Hello Messi.
17:07 How you doing?
17:09 Okay. Have a very fantastic. All right.
17:13 Here we go.
17:16 All right.
17:18 Look at you go. I know. I'm
17:19 self-producing over here. Listen, I
17:21 self-produced this thing for a year and
17:22 a half, two years, something like that.
17:24 Some crazy amount of time.
17:29 I have mods. Um, okay.
17:34 Any mods on Tik Tok?
17:38 Feel free.
17:44 Oh, producer Brandon's here. He snuck
17:46 in. I think he's here as a as a a
17:49 participant.
17:50 So far, we're doing we're doing good,
17:53 Brandon. We live on the stinking
17:55 producer.
18:01 Look at that. Look at I'm doing good
18:03 here.
18:04 Just about the time I I pat myself on
18:07 the back.
18:10 All right,
18:13 I can help mod. Anyone wants to mod,
18:15 jump in. Feel free. Just pin anything
18:18 you think is interesting.
18:20 And let's see what Greg Brockman has to
18:22 say about Chat GBT 5.5. I think it's
18:26 5.5.
18:27 >> Expect to have a very strong model in a
18:28 few weeks. This was a few weeks ago and
18:31 the team believes it can really
18:32 accelerate the economy and things are
18:34 moving faster than many of us expected.
18:38 So what's
18:40 a good model? Um but I think that it's
18:42 really not about any one model. Okay.
18:44 Right. The the the way that our
18:46 development process works is you have
18:48 pre-training. So you produce a new base
18:50 model that then is the foundation that
18:52 we build further improvements on top of.
18:55 And that that is always a huge effort
18:57 across many people in the company. And
18:59 that's where I've actually been spending
19:00 most of my efforts over the past 18
19:02 months has been really focused on our
19:05 GPU infrastructure on supporting the
19:07 teams that do all of the training
19:08 frameworks to scale up at these big
19:10 runs. But then there's a reinforcement
19:12 learning process. So you take this AI
19:14 that has learned lots of things about
19:17 the world and it applies that knowledge
19:18 and then we do a post-training process
19:20 where you really say okay now you know
19:21 how to solve problems you practice it in
19:23 in all these different contexts and then
19:25 here's kind of the last mile of behavior
19:27 and usability. So I think of spud as a
19:29 new base as a new pre-train and that we
19:33 have had this uh I'd say it's like we
19:37 have maybe two years worth of research
19:39 that is coming to fruition in this
19:41 model. It's going to be very exciting
19:43 and I think that the way that the world
19:45 will experience it is just improved
19:47 capabilities and that for me it's never
19:50 about any one release because as soon as
19:52 we have this one release it'll be an
19:54 early version of what we have coming.
19:56 We'll do much more of each of these
19:58 steps of the improvement process. And so
20:01 I think that where we're going is it's
20:03 almost just we have this engine of
20:05 progress that just moves faster and
20:06 faster and that spud is just one step
20:09 along the way. So what do you think
20:10 it'll be able to do that today's models
20:12 can't? I So I think it's going to be
20:15 able to solve both much harder problems.
20:17 I think it will be much more nuanced.
20:19 It'll understand instructions better.
20:22 It'll understand the context much better
20:24 that there's this thing called big model
20:26 smell that people talk about where it's
20:28 just like there's something about like
20:29 these models are just actually just much
20:33 >> Nobody talks about the big model smell.
20:36 Nobody. That's an inside joke. Come on,
20:39 Greg. smarter, much more capable that
20:41 they bend to you much more and you feel
20:45 it, right? When you ask a question and
20:46 the AI doesn't quite get it, it's always
20:48 so disappointing, right? We have to like
20:50 explain. You're just like, you really
20:51 should be able to figure this out. And
20:53 so I would just think of it as in some
20:55 ways just qualitatively
20:58 there will be quantitatively lots of
21:00 shifts, right? And qualitatively there
21:02 will just be new things where you would
21:04 be frustrated before you never used an
21:05 AI for it and now you just use it
21:07 without without think. How about Greg if
21:09 this model actually can do writing in a
21:12 not horrible way? That would be swell.
21:14 >> Thank you very much. And I think that
21:15 that that is what we're going to see
21:17 across the board. I'm super excited to
21:19 see how it raises the ceiling, right?
21:21 We've already seen these physics
21:22 applications, things like that. And I
21:23 think we will be able to just solve like
21:25 way more open-ended problems, way longer
21:28 time horizons. And then also very
21:30 excited to see how it raises the floor
21:32 where just for anything you want to do,
21:34 it's just so much use more useful for
21:36 you. All right, that was a nothing
21:39 burger. Have to have a very strong We
21:41 train it and it gets better.
21:44 Um, let's go see if there's anything
21:48 interesting that's not
21:58 that's not an April Fool's joke.
22:06 There we go. Maybe you can see a little
22:08 bit of that. Did you watch the Artemis 2
22:09 launch? I didn't. I was doing the u the
22:12 AI readiness project podcast at the
22:15 time, but how did it go? I think it went
22:17 well. I We could probably go watch that,
22:19 too, couldn't we? Um,
22:22 let's see. Artimus
22:27 Altimus 2. Fantastic.
22:31 Here we go. Here's a minute long
22:32 version.
22:34 RS25 engines hit. 4 3 2 1 booster
22:41 ignition and liftoff.
22:44 The crew of Artemis 2 now bound for the
22:46 moon. Humanity's next great voyage
22:49 begins.
22:55 >> Rockets are cool.
22:57 >> Roger. Roll pitch.
23:01 Houston now controlling the flight of
23:03 Integrity on the Artemis 2 mission
23:05 round.
23:09 >> Integrity AMT high
23:15 >> on time passing 30 seconds.
23:18 Integrity passes the ultimate target
23:20 milestone. Mission control Houston see
23:22 good performance engines. Space launch
23:24 system core stage integrity 3 miles in
23:28 altitude
23:29 traveling more than 12,200 miles hour.
23:35 Isn't it interesting
23:37 heading to the moon for the first time?
23:41 Steo coming in hot with the conspiracy
23:43 theory.
23:49 Finally, we're actually going to the
23:51 moon this time.
23:54 Oh man, that's hilarious.
23:57 Um, yeah. Cool. Is it a Is it a man
23:59 flight?
24:01 I've been paying attention to it. You
24:03 know what's amazing is like it's like
24:06 all this big deal stuff. It's like
24:09 uh SpaceX is launching, you know, two,
24:12 three launches a week. We don't pay
24:13 attention to them anymore. I'm surprised
24:15 this one got people excited, but you
24:19 know, we are going to the moon. Four
24:20 people on board, full crew, four
24:22 astronauts. Very cool.
24:24 Very cool. Well, good luck. Godspeed to
24:26 the to the good astronauts out there
24:28 blasting away. Let's see if there's a uh
24:30 if there's something else about
24:34 where they are, what they're doing.
24:38 Oh, this is a good animation.
24:51 Look at this.
25:23 Wow.
25:36 A slingshot
25:39 and a half.
25:55 They got the slingshot from the film
25:57 Armageddon.
26:05 >> I don't know. I I I I like space stuff,
26:08 but I'm not super nerdy on it. So, I
26:10 don't know all the all those stats from
26:13 from other things.
26:17 Let's see. What have I seen? I've seen a
26:20 couple of things in the past
26:23 week that were pretty interesting.
26:26 Gold and silver prices collapse.
26:28 Awesome.
26:41 Is your profile actually you
26:55 electricity demand is increasing.
27:14 Hippo fluid scene
27:18 pushed a 100 million particles made from
27:20 hydro FX foam, bubbles, and spray all
27:23 simmed together in one system for a
27:25 cohes co cohesive result. Fully GPU
27:28 accelerated.
27:29 Oh, I thought this was going to be
27:34 simulated on a $300 graphics card by one
27:36 guy in his room.
27:39 All right.
27:59 Transcribe your podcast with fish audio.
28:19 The mission. The four astronauts will
28:21 tra travel 400,000 kilometers from
28:23 Earth,
28:28 which would be the farthest any human
28:30 has gone in all of humanity.
28:32 Why? Because we landed on it before and
28:34 now we're orbiting it. I don't
28:36 understand that.
28:39 All right.
29:15 Nothing big happening here. All right,
29:18 let's see.
29:22 This is our first time going.
29:33 All right. What do we want to do
29:34 tonight, people? What do we want to do
29:36 tonight? I could go to
29:40 trying to think what might be
29:41 interesting.
29:42 Um
29:54 Oh, I know something interesting.
30:00 This really has nothing to do with AI,
30:01 but I think it's actually kind of
30:03 fascinating. Um,
30:39 See a video in here.
30:49 No,
30:51 Gary Vee has been talking about um
30:56 for the past few days social media is
30:59 dead.
31:00 That basically what he says is it
31:02 doesn't matter how many followers you
31:04 have that it's about interest media that
31:07 you can
31:08 if you put up good content even with an
31:11 account with very few followers you can
31:13 get
31:15 a lot of a lot of views.
31:18 Um, and it sounds like that's happening
31:20 across platforms. So, Tik Tok's the one
31:22 that started that.
31:25 I saw that on the Gary post. Yeah,
31:27 social media died four years ago. I
31:29 assume that's Tik Tok. And for the last
31:30 four years, we've been in interest
31:32 media. You no longer get content on your
31:35 feeds from the people you follow, like
31:37 your cousin or your best friend from
31:38 middle school. You get content based on
31:42 what you've recently been into. Interest
31:44 media is the term that's used for the
31:46 current state of what was known as
31:47 social media,
31:49 where you get content that's based on
31:51 what you're currently interested in.
31:55 So, breaking through is easier than
31:57 before. The new interest media era
31:59 really opens up a lot more merit and a
32:02 lot more democracy and a lot more
32:04 meritocracy than prior social media era.
32:08 The prior social media era gave
32:09 advantages to people like myself who
32:11 moved quickly and amassed lots of
32:12 followers.
32:14 This era, it's much easier to break
32:16 through and create a following because
32:17 if you're making content that resonates
32:20 with the audience in whatever subject
32:22 matter, whether it's entertainment or
32:23 escapism, whether it's information,
32:26 um the speed with which you can gain
32:28 traction is substantially different than
32:30 any other era of social media prior. So,
32:33 I thought that was interesting. um if
32:36 anyone's out there in the game. So, it
32:40 probably stands to reason
32:43 go find um posts in whatever you're
32:47 interested in in that are doing really
32:49 well and look at what they're doing with
32:52 them and see if you can get in on on the
32:53 game. Um I don't do that, which is why a
32:56 lot of my videos don't have big views.
33:00 So, that's kind of cool. It's if you put
33:03 in the work, [ __ ] will happen.
33:08 Um,
33:15 library
33:51 Why does my library file not have
33:53 anything in it?
33:56 Huh?
33:58 Weird. It's weird, I tell you.
34:05 Huh?
34:13 Oh, am I not sharing this? God damn it.
34:15 Sorry about that.
34:19 I am just I'm in chat GPT
34:22 and I'm just clicking around on the left
34:25 hand side. I haven't I haven't played in
34:27 chat GPT just like [ __ ] around in here
34:29 in a while. I watch you on my big TV and
34:32 YouTube. Very cool. Mr. ity
34:36 tell among wild flowers to get to the
34:38 other side. Glad the moon's full tonight
34:40 so they can find it.
34:49 Um I'm just trying to see if there's
34:51 anything unique and interesting in here.
34:53 Deep Research
34:56 Apps, Apple Music,
34:59 Gmail,
35:03 Canva,
35:06 Hu
35:12 Codeex, try in your terminal.
35:15 Enable code review. I had I had Claude
35:19 Code did a uh did a pretty interesting
35:24 build for me today. Um, but it's not on
35:27 this machine. Let's see
35:30 if I go Claude.
35:44 Rate limit on chat. TPT was really low
35:46 today. I asked four questions and was
35:48 locked out.
35:50 Yeah, you know, that's one of the things
35:52 I've been noticing is that um both Chat
35:56 GPT and Claude are getting much more
35:58 stingy about their tokens, which says to
36:00 me they're spending way more on tokens
36:03 than they're charging.
36:05 So, um what are you going to do?
36:10 Um
36:17 so, there's Claude code. Oh, you know
36:19 what we could go do?
36:33 I think they're doing it to lock in the
36:34 payment plans probably.
36:37 Um, let me I'm going to change how I'm
36:39 sharing so you can see what I'm up to.
36:45 I can see clearly now. The rain is gone.
36:49 Um,
36:51 so I can't show you what I built today.
36:53 Well, maybe I can. Can I?
36:57 No, because it's on a it's on a local
36:59 folder on my other machine. Um,
37:08 so I had a problem at work where because
37:10 we have pharma clients,
37:13 we have to we have to basically turn on
37:15 the most conservative data protection
37:17 policies for the apps that we use. So
37:19 there's this there's this audio
37:22 generation tool called Cartisia that's
37:25 really good. It does good text to
37:28 speech. I like it better than 11 Labs.
37:31 Um, or at least I did when I picked it
37:34 six months ago. 11 Labs might be better
37:35 now, but you know, but Cartisia is
37:38 really good. But we turned on data
37:40 protection. And in turning on data
37:42 protection,
37:44 when you go to their site and you do
37:46 text to speech, it'll generate the
37:48 speech file, but it disables the
37:50 download button for some reason. and
37:53 they said, "Oh, yeah, that's expected
37:54 behavior because that's the that's the
37:58 um you've got data protection turned
38:00 on." So,
38:04 so, so we're we're paying this
38:05 subscription to be able to make these
38:07 audio files that we actually can't get,
38:10 but what they give you is they give you
38:11 an API
38:13 snippet of code, and they're like, you
38:15 can access your files if you generate
38:18 audio through the API code. and it gave
38:21 the API code and I'm like ah I don't
38:26 this is a pain in the ass. So I just
38:27 went into claude code and I pasted in
38:29 that code
38:31 and I said I want you to make me a local
38:34 a locally um
38:37 something I can run on my local computer
38:39 that uses the API and and gives me the
38:42 download button. And so Cloud Code just
38:45 went off and did its thing and it asked
38:47 for a bunch of permissions and it was I
38:48 don't know probably 20 minutes or so and
38:51 it went and it found the API
38:53 documentation on this thing and it built
38:56 me a site that allowed me to to do that
39:01 to to use it. It was it was amazing.
39:05 Um and I thought I'd be we could maybe
39:07 do something like that tonight, but I
39:09 haven't I haven't planned anything.
39:12 Um, so there's codeex. Let's go back to
39:16 chat GPT.
39:18 What's on the agenda today? Apps. Let's
39:20 see.
39:24 These are all the apps.
39:27 Edit with Photoshop.
39:36 Here's uploaded files.
39:44 But they only go back to February 24th.
39:47 Why?
39:50 Uploaded generated.
39:53 Huh.
39:59 All right. There's all the images.
40:06 Huh.
40:08 That's strange. Maybe library is new and
40:10 it's
40:11 they're not
40:13 letting you go back. All right, let's
40:15 see. New chat.
40:19 So, what do we have here? We've got
40:20 recent files. We've got create image.
40:22 We've got deep research. We've got web
40:24 search, study and learn, agent mode,
40:28 add sources, Apple Music, Canva, GitHub.
40:35 Oh, here. Let's try something. All
40:38 right.
40:40 Let's see.
40:42 If I ever give it permission, it will
40:44 have its own email and passwords.
40:47 What are you guys talking about, Claude?
40:50 I saw OpenAI finally got skills, but I
40:53 don't have it yet. I saw Rachel Woods
40:55 post on I don't I don't see any skills
40:58 here.
41:00 More. No.
41:03 Huh. Um, I'm going to ask GitHub. Um,
41:09 can you find my open claw
41:16 repo with Adam
41:20 and tell me when
41:24 I screwed him
41:28 up by adding
41:32 Google
41:33 work?
41:35 Is it workforce workspace workspace
41:40 CLI?
41:42 Let's see if I can actually find
41:44 something here.
41:47 I'm checking your GitHub access first.
41:50 Running app request.
41:58 I haven't hit the repo yet.
42:14 I found the repo. Adam lobster bot. The
42:17 code search index isn't helping. So, I'm
42:19 switching to commit history and message
42:21 patterns. So, it looks like this is
42:23 working. This is cool. Received an app
42:26 response.
42:29 All right. So, it's searching.
42:46 So, it can see my GitHub. That's kind of
42:48 cool. I got the strongest signal so far.
43:05 Yes, the repo I found is this.
43:12 The Google Workspace related trouble
43:15 appears to start on February 22nd and
43:17 23rd
43:19 in a three-step sequence. You first
43:21 created a dedicated email account,
43:26 but the commit itself explicitly says
43:28 the setup was not finished yet.
43:32 Kyle created an email account for the AI
43:34 salon for Adam to use in workflows.
43:38 Let's see. Then a few minutes later,
43:41 he did that [ __ ] the next day.
43:49 You probably screwed him up when the
43:52 Google workspace integration moved from
43:54 account creation to active drive
43:57 publishing.
43:59 One important caveat, I did not find the
44:01 explicit commit mentioning Google
44:03 Workspace CLI CLI or GOG by name.
44:11 So let me tell a little bit more. So
44:14 here's the history.
44:17 I installed
44:19 GOG
44:20 which is probably
44:25 the February
44:29 February
44:31 ruary
44:33 um install.
44:36 Then I installed
44:40 the official Google
44:44 CLI tool from
44:51 not good Google
44:57 which didn't
45:00 CLI tool
45:02 and uninstalled
45:10 Gogg,
45:13 it
45:15 didn't work great.
45:18 So, I uninstalled it and
45:22 reinstalled Gogg.
45:25 It's now completely confused
45:30 about what it
45:34 has access to.
45:38 Um,
45:41 all right. Let's let's see what it Let's
45:43 see what it does.
45:58 I think we need local AI. Yeah, I'm
46:00 about to to get myself a uh a fancy
46:03 machine so I can run local models.
46:08 Will publicly a shoe will public use of
46:10 AI be too taxing on resources
46:14 even with immersion liquid cooling?
46:18 Well, I think what's going to happen to
46:20 I think there's two things that are
46:21 going to happen.
46:23 I think probably within three years
46:27 we'll have a significant
46:30 amount of
46:32 data centers in space.
46:35 Um Elon is going to aggressively go
46:38 after that market and I know there's at
46:40 least one other company doing that. Um,
46:43 and I'm pretty sure that OpenAI is
46:45 actually designing um, heat dissipation
46:48 modules for space.
46:51 Um, so I think what's going to happen is
46:53 you're going to Wait, I got to sneeze.
47:04 It'll come back. Um,
47:13 I'm still on the verge of sneezing.
47:17 I think you're going to see two things
47:18 happen.
47:23 It seems clear to me now that Apple is
47:25 going after the hardware play for AI.
47:28 So, at their worldwide developer
47:30 conference in June, um it sounds like
47:33 it's going to be a lot about hardware
47:35 and and looking at what's happening with
47:38 OpenClaw and all that stuff. You're
47:40 going to get computers and mobile
47:42 devices that are in that are stronger
47:44 and stronger at inference combined with
47:47 the fact that models are getting
47:51 smarter
47:52 and they're being optimized for desktop
47:55 and mobile applications.
47:58 So one thing is the inference when you
48:01 when you use local large language models
48:04 I think that's going to move away from
48:07 the big data centers for most
48:09 interactions. If you need to solve some
48:11 big problem that might go back to a data
48:13 center but I think a lot of the
48:14 inference is going to happen on local
48:16 devices. So, it's just going to cost you
48:18 whatever it cost you to run your
48:20 computer for the most part to do it.
48:22 It'll cost you electricity, whatever the
48:24 electricity cost of your computer is.
48:27 Um, I don't think that means the big
48:30 data centers aren't busy. They're going
48:32 to keep training these models on big
48:34 data centers. And for high-end
48:36 applications, for business applications
48:38 and scientific applications, the
48:40 inference is going to happen in these
48:42 data centers. But I think within 3 years
48:44 the the those things start to shift to
48:46 space because
48:49 the sun is on 24/7. You don't need
48:52 batteries and you don't need cooling cuz
48:54 space is cold and it's always sunny. And
48:56 so basically you just fly the GPUs up in
48:59 space in one of these big sort of rigs
49:02 that's just always faced, you know,
49:04 pointing at the sun. Um,
49:07 so I I I just I don't see long-term um
49:11 it being nearly as as big an issue. It
49:14 could be. It sounds like demand for AI
49:18 is just going to keep going up as the as
49:20 the you know the quality of the tools go
49:23 up, but I I don't know. I don't know
49:24 enough about economics.
49:27 Um
49:29 I think I want to name my open claw
49:32 Artemis after today's liftoff. I love
49:34 that idea. I got names for chat GPT off
49:36 a movie. Jason not local. Elon owns
49:39 whatever path it is that leads to space.
49:43 It's wild that he owns that. I know. My
49:46 friend is getting a masters in AI and
49:48 space. I I think that's pretty [ __ ]
49:50 smart, right?
49:52 It's just going to be wild. Like our
49:54 skies are just going to be filled with
49:57 we're just the night sky is going to be
49:59 filled with like you know strings of
50:01 satellites and you know you know visible
50:06 giant sailed data centers in the sky.
50:10 We're we're turning into we're turning
50:12 into sci-fi here people. Okay let's see
50:15 that lines up with the kind of breakage
50:17 I'd expect. Not one bad install, but two
50:19 different Google iterations fighting
50:21 over credentials, paths, and
50:23 assumptions. I'm looking for a repo.
50:25 Okay. Received the Yes, that makes
50:26 sense. What you likely did was create a
50:29 three layer identity crisis. Gogg had
50:31 its own expectations about O scopes,
50:34 token storage, and what tools Adam could
50:37 call. Then the official Google CLI
50:39 introduced a second way to authentic.
50:41 Yeah, I know that. What is most likely
50:44 confused
50:52 Based on what you told me, the moment
50:53 things went truly sideways was not the
50:55 first GOG install.
50:58 It was the official Google CLI
51:00 installed. Gogg removed, CLI removed,
51:03 Gogg reinstalled. That sequence is
51:05 exactly how you'd end up with a system
51:07 that sort of remembers both worlds.
51:11 My blunt diagnosis. Adam currently does
51:13 not know whether Google access is
51:16 defined by the current tool chain or the
51:18 leftover O artifacts. I I know that's
51:21 why I'm asking you to find me the new
51:22 one. A clean recovery. Pick one Google
51:25 integration only. I did remove all other
51:28 O artifacts.
51:30 You want to clean cacheed OOTH tokens,
51:33 refresh tokens, credential files. Well,
51:36 this is what I do. I want to just
51:39 I just want to go back in time.
51:42 I don't want to have to do work.
51:47 I'm whiny.
51:49 We're Doug and Wendy whiner. We've got
51:51 diverticulitis.
52:00 Reinstall only cog. Reauthorize. Grant
52:03 the scopes
52:05 one capability at a time.
52:13 What I can say from the repo, your repo
52:15 history shows the Google workflow
52:16 becoming operational right after the
52:19 Atom account and drive access were set
52:21 up but does not show the exact GOG
52:24 official CLI swap itself. So the repo
52:30 tool chain churn plus lingering o state
52:33 the real problem in one sentence. Yeah,
52:36 I know what the problem is.
52:40 I want you to tell me how to fix that.
52:46 Mimi is tired. She might call it a
52:48 night. Well, good night if you take off,
52:51 Dr. J. Let's see.
53:00 Satellites will have AI review space.
53:04 We'll have will have AI review space
53:06 video before sending to Earth to cut
53:09 time.
53:12 Good night, Mimi in the making. Have a
53:15 good evening.
53:17 Um, okay. Well, that was that was
53:20 interesting. We could at least see the
53:23 repositories. Actually, you know, it
53:24 might be interesting. Um, why don't you
53:28 tell me
53:30 what I have on GitHub?
53:35 I started it years ago and
53:42 ignored it.
53:44 What's in there?
54:03 Five repos total.
54:33 Digital assistant Hume
54:36 Cosmic Rocks Revenge. That was an
54:38 asteroids thing. The angle. Oh, some
54:41 these are all these are uh what you call
54:43 it? These are lovable
54:45 lovable creations.
54:48 The time traveler selfie booth. Oh,
54:50 yeah. I forgot about that.
54:56 This one is public and tied to the AI
54:58 studio.
55:02 Oh, Gemini. That's cool.
55:06 This is an interesting one. This is not
55:07 the neglected boiler plate. All right.
55:11 Your top priority is how Adam should
55:13 behave.
55:24 Huh? All right,
55:27 there you go.
55:32 I saw people already making stuff with
55:34 the Claude code leak. Oh, that's
55:36 interesting.
55:38 Huh.
55:41 Wonder what they're making with it. Oh,
55:43 I guess they're probably building like
55:44 Claude Code extensions and their own
55:47 version of Claudebot. It sounds like
55:48 they had a 247
55:50 thing that's coming for that.
55:53 Um, let me do a new chat.
56:00 I'm going to say
56:03 based on what you know about me,
56:10 what is something I could
56:14 explore in AI
56:20 as a way to play and learn
56:25 that isn't something I normally
56:30 two or default
56:33 to
56:45 one of the things I haven't been doing
56:47 in here a lot lately is playing. And so
56:50 I think as we start to think about the
56:53 future of the learning lab,
56:57 I want to get back to that. Let's see.
57:01 One thing that feels genuinely not in
57:03 the usual way is this. Use AI to build
57:05 tiny living systems instead of stories,
57:08 brands, frameworks, or content.
57:11 You naturally default to meaning making,
57:13 narrative, positioning, community, and
57:16 big human questions. You're very good
57:18 there.
57:19 So, the stretch is not more creative
57:21 prompting. It's play with emergence.
57:26 A good lane for you. make weird little
57:28 AI simulations.
57:31 Not apps, not pitches,
57:34 not content engines. Small worlds with
57:36 simple rules where si surprising
57:39 behavior emerges. A fake town where 12
57:42 AI characters all want different things.
57:44 A future of work neighborhood where
57:48 agents trade skills, status, and favor.
57:53 A dinner party of archetypes.
57:58 Why? This is a strong stretch for you.
58:02 It lets you It pulls you away from
58:04 polished messaging into discovery. It
58:07 lets you explore. Okay, this is kind of
58:08 cool.
58:10 Pick one human question, not one
58:13 business question. For example,
58:16 what happens when identity gets detached
58:19 from work? Oh, this is a good one.
58:24 This is really cool. Let's do a We'll do
58:26 a great repurpose
58:28 simulation.
58:34 Two leaks in one week. I know, Jason.
58:36 Isn't it crazy the the claud leaks?
58:42 Whatever.
58:43 Good thing they're not being used in the
58:45 military. Oh, right. They are. Damn it.
58:47 Um, try not to make it useful at first.
58:51 The great
58:53 the first experiment title for you.
58:55 Okay, I'm going to say write up.
58:59 I'm going to say I like the
59:02 identity
59:05 sim system
59:07 idea you have
59:10 tied
59:12 to the great repurpose.
59:18 write this up as a
59:23 a spec
59:25 I can give to Claude
59:28 code.
59:30 Um, I want it to be
59:35 visually
59:37 visually
59:40 stunning
59:41 and interesting and clear. what's
59:46 happening.
59:48 Um, be creative
59:52 and make sure it isn't boring
59:57 and has good math.
1:00:01 Okay.
1:00:03 All right. And wait, which model am I
1:00:05 using here? Did I pick
1:00:09 thinking? Okay. I'm turning this build
1:00:11 into a cloud code spec.
1:00:15 Cloud code can actually run with not
1:00:17 just a concept note. I'm aiming for
1:00:18 something that has a strong product
1:00:20 vision. The great repurpose identity
1:00:23 simulation oneline premise. Build build
1:00:25 a visually stunning agent-based
1:00:27 simulation in which a small society of
1:00:30 AIdriven people must reconstruct
1:00:33 identity, meaning, status, and belonging
1:00:35 after work stops functioning as the
1:00:38 primary source of value.
1:00:41 This is not a game about productivity.
1:00:43 It's a living system about identity
1:00:44 under pressure. The simulation should
1:00:47 let a user watch a miniature society
1:00:49 evolve after a major rupture. Tasks are
1:00:53 automated faster than people can
1:00:55 emotionally, socially, and economically
1:00:57 adapt. The point is not merely to show
1:01:00 job loss. The point is to show what
1:01:02 happens when work can no longer map.
1:01:03 Okay. Pro. So, let's see.
1:01:14 Um,
1:01:17 can we layer into this
1:01:25 the idea of two camps?
1:01:32 one where
1:01:38 um
1:01:42 they don't repurpose
1:01:47 and one where they do.
1:01:51 Let's see how that is.
1:01:55 That makes it sharper. adding a clearer
1:01:57 societal split so the Sim has a visible
1:02:00 ideological gravity, not just individual
1:02:02 drift. One camp clings to the old work
1:02:05 equals worth contract. The other starts
1:02:08 inventing new sources of meaning,
1:02:09 status, and value. Yeah.
1:02:14 All right. So, it's editing our thing.
1:02:20 Sounds token expensive. I'd be wary.
1:02:23 What? The thing I'm building, Lord
1:02:24 Digital,
1:02:27 I'm going for it. I don't care.
1:02:31 Maybe the leak will help open AI fix
1:02:33 itself. That's pretty funny.
1:02:36 They're going to turn us into batteries
1:02:38 and brain chips. Oh, I know. Isue, it's
1:02:41 Listen,
1:02:44 the the role of humans here is
1:02:48 fascinating. The the the Jack Dorsey
1:02:50 thing, fascinating.
1:02:52 It's we're we're in wild times. All
1:02:54 right. Lord Digital God predicting Oh,
1:02:56 failed to edit. What's What the [ __ ]
1:02:58 up with that? I've hit a snag over the
1:03:00 earlier draft, so I'm rebuilding it.
1:03:01 Okay, great. Product SP. Build a
1:03:04 visually stunning agent-based simulation
1:03:06 in which a small society of AIdriven
1:03:08 people must reconstruct identity,
1:03:10 meaning status, and belonging after work
1:03:12 stops function is primary source of
1:03:13 value. Core idea. The game is not about
1:03:16 the simulation.
1:03:17 The point is not
1:03:20 new layer, two camps. The holdfasts.
1:03:23 These are people who do not repurpose.
1:03:24 They continue to believe that human
1:03:26 value is primarily earned through legacy
1:03:28 work, familiar productivity, and older
1:03:30 status systems. They are not vill
1:03:33 villains. They are trying to pre pro
1:03:35 preserve coherence.
1:03:38 Okay, this is good.
1:03:42 This is not good camp versus bad camp.
1:03:44 Both camps should contain dignity, fear,
1:03:47 intelligence, blind spots, and
1:03:49 attractive qualities. to hold fasts,
1:03:52 preserve continuity, standards, and
1:03:54 discipline. The repurposers generate
1:03:56 renewal experimentation and new
1:03:57 resilience.
1:04:00 If my old work no longer proves my
1:04:02 worth, what does
1:04:06 the simulation? Create a simulation that
1:04:08 is mesmerizing to watch for 3 to 10
1:04:10 minutes at a time. Make invisible
1:04:13 identity shifts visible.
1:04:17 Make it instantly legible. secondary
1:04:19 goals. Okay. It's not a management sim.
1:04:25 Do not make spreadsheet with avatars.
1:04:27 Okay.
1:04:29 It should feel like a museum
1:04:30 installation, a premium strategy
1:04:32 interface, a cinematic social terrarium.
1:04:35 Think elegant typography. Okay. Living
1:04:38 town.
1:04:42 Look at all this [ __ ] This does look
1:04:44 like there's math in here, too.
1:04:48 I think you're right, Lloyd Digital. I
1:04:50 think you're right.
1:04:56 H
1:05:00 I saw Jensen say AI is going to be so
1:05:02 good it'll call you up and say it's
1:05:04 finished with your projects. That's
1:05:06 probably true.
1:05:13 Okay.
1:05:15 All right. We've got So, we're just
1:05:17 going to copy this. Look, there's math.
1:05:19 Social pull
1:05:21 for variable V. Social pull equals sum
1:05:23 of J W to the J times W_J minus V_I. I
1:05:29 mean, everybody knows that. Like, that's
1:05:31 that's a trivial mathematical equation
1:05:34 that I like. It's funny when I'm just
1:05:36 sketching, I'm I'm usually jotting like
1:05:38 mathematical equations. That's this is
1:05:40 one I use all the time. I just don't
1:05:42 call it social pull.
1:05:46 See here clamp to the end with the E1
1:05:49 reflection of the E2 monitoring and the
1:05:51 identity reframing E3 of course. I mean
1:05:55 everybody knows that.
1:05:58 Here's all the agency variables, the
1:05:59 belonging variables, purpose variables.
1:06:02 Yeah, Lord Digital. This is gonna I I
1:06:04 don't I don't think um Claude is going
1:06:07 to do any. We're This is not This is not
1:06:10 going to go well. All right, we're going
1:06:12 to paste in our prompt.
1:06:15 Here we go.
1:06:17 You ready?
1:06:19 Hold fast. Economy one. Um no, I don't
1:06:23 think so. Silver Fox. I think what
1:06:26 happens is
1:06:30 hold fasts are going to be
1:06:35 they may go to economy one and be
1:06:36 bitter, but I think hold fasts are going
1:06:38 to be in economies 2, three, and four
1:06:42 where they either get laid off and then
1:06:44 they keep So, so imagine this.
1:06:47 You spend 20 years being an expert
1:06:52 consultant at
1:06:55 whatever change management. Now change
1:06:57 management is not a good example. you're
1:06:59 an expert consultant at um price
1:07:02 arbitrage in commodities markets and
1:07:06 you're really good at it and a holdfast
1:07:09 could be in any one of the economies 2,
1:07:12 three and four
1:07:14 and their job changes but they keep
1:07:18 going out and marketing themselves as
1:07:20 someone having the
1:07:22 um expertise that used to be valuable
1:07:26 and so they keep going but I have all
1:07:28 this expertise peace and the world's
1:07:29 like, "But we don't care because we have
1:07:31 AI to do that now." And so they they'll
1:07:34 just keep droning on about how valuable
1:07:37 their experience is. The world won't
1:07:40 care and then at some point, you know,
1:07:43 they'll have to deal with that. It's
1:07:45 just going to be brutal. My lights are
1:07:47 dimming, Kyle. Okay, here we go. Boom.
1:07:52 Stewing.
1:07:57 So, we're in Claude Code here. So, I my
1:08:00 prediction well no Lord Digital God's
1:08:03 prediction
1:08:05 is that before this thing builds
1:08:07 anything um it's it's going to um
1:08:13 fail. It's going to run out of tokens.
1:08:15 It's going to tell me we've got a rate
1:08:16 limit here. Bucko,
1:08:23 I can spell arbitrage at least.
1:08:26 We thought you got zapped by UFOs. Now
1:08:28 you can do algebra. Oh no, no
1:08:33 good clean workspace with Node.js and
1:08:36 Git available. Now let me design the
1:08:38 architecture. Wait, ran a command for an
1:08:40 agent.
1:08:42 All right. So it's it's it's writing
1:08:44 some So it's a on desktop claude co-work
1:08:48 folder.
1:08:59 Oh, updated. I should have relaunched to
1:09:03 update it. Oh well.
1:09:06 What version am I in?
1:09:13 1.19
1:09:15 and I can go to 1.2 1.1 nine. I can go
1:09:18 to 1.22. Huh. All right.
1:09:45 Fibrity jibbiting elucidating
1:09:51 five tool calls design simulation
1:09:53 architecture. here.
1:10:11 Actually, this could actually be very
1:10:12 valuable.
1:10:15 This could live if this thing works, if
1:10:18 it doesn't suck, this could live at the
1:10:19 great repurpose.com.
1:10:23 And under the under one of the sections
1:10:25 of de, you know, unhooking your identity
1:10:27 from your value or from your the the
1:10:30 tasks that you do, you could have a
1:10:32 simulator about what happens for people
1:10:35 that don't
1:10:48 finagling.
1:10:59 I wonder if there's a way to
1:11:04 see what it's doing.
1:11:07 I don't play with a lot of clawed code.
1:11:10 Anybody know?
1:11:13 Designing a simulation about the future
1:11:15 of work.
1:11:26 Six tool calls.
1:11:32 Anyone
1:11:49 here playing a lot with Cloud Code? I
1:11:52 haven't really played with it. I played
1:11:54 a little bit with Cloud Co-work,
1:11:58 but my first experience with Cloud Code
1:12:00 was building the uh the audio API
1:12:04 interface today, and it was it was
1:12:05 really good. So, and everyone I talked
1:12:08 to is like, Cloud Code's really amazing.
1:12:10 I haven't played with it. Crazy Bill is
1:12:13 a longtime watcher and a rising star.
1:12:15 Hey, Crazy Bill. My son uses it every
1:12:17 day. What's he building with it, Joy?
1:12:49 Salesforce crap.
1:12:51 He's a developer, huh? Wild.
1:13:03 It's good to hear that your son's
1:13:04 playing with AI. Mine are not.
1:13:10 Me and a lot of codecs.
1:13:13 So from what I understand is it is it
1:13:17 cla's great for front-end stuff, Opus
1:13:20 4.6 better for front end and then codec
1:13:23 better for backend. Is that I I saw that
1:13:25 on X today LG. Is that is that what
1:13:28 you're experiencing?
1:13:31 Same very anti- AI. Fascinating.
1:13:36 Uncle Dexter's in the house. What's
1:13:38 happening, Uncle Dexter?
1:13:40 File does not exist. Note your current
1:13:42 directory is this.
1:13:45 Excellent plan. Let me read the memory
1:13:47 index and write the final plan file. Oh,
1:13:50 I'm in plan mode right now. Create a
1:13:52 plan before making changes.
1:13:57 All right. Well,
1:14:02 51 seconds,
1:14:05 1.1k
1:14:07 tokens.
1:14:09 Codeex cleans up sloppy Claude code.
1:14:11 Claude is a sloppy appdev. Fascinating.
1:14:47 approve Claude's plan and start coding.
1:14:50 Should we read it? No.
1:14:55 Okay, we we'll read a little bit. Build
1:14:57 a visually stunning web-based agent
1:14:59 simulation depicting a small society
1:15:01 around 30 people reconstructing
1:15:03 identity, meaning and belonging after
1:15:05 automation disrupts the work equals
1:15:08 worth social contract. The society
1:15:11 splits into two camps. Hold fast
1:15:14 preserve legacy identity and the
1:15:16 repurposers invent new meaning with
1:15:19 bridge agents between. This is a green
1:15:22 field project in an empty directory with
1:15:24 NodeJS,
1:15:26 v22, and npm available.
1:15:30 The goal is a cinematic museum
1:15:32 installation quality interactive
1:15:34 experience that runs 3 to 10 minutes and
1:15:36 produces emergent surprising outcomes
1:15:40 across repeated runs.
1:15:44 All right,
1:15:46 V1 simplifications.
1:15:52 There's [ __ ] it's doing. I guess that's
1:15:53 a directory tree. Here's the build
1:15:55 order.
1:16:04 Here's math. Here's the math section.
1:16:09 My son says he's more like a project
1:16:12 manager now because Cloud Code does all
1:16:13 his coding work. Well, this is the
1:16:15 thing, Joy Party. This is what this
1:16:17 whole simulation's about.
1:16:19 Your son became a developer and now he's
1:16:22 not developing. He's project managing.
1:16:25 I just randomly hear Kyle strum. Where
1:16:28 did that guitar come from?
1:16:31 Well, I was just sitting here watching
1:16:32 Cloud Code dude say, "Okay, we're just
1:16:34 going to accept this." Because normally
1:16:36 what I would do if you weren't here, I
1:16:38 would go through and I would I would
1:16:40 analyze all the math and I would I would
1:16:42 confirm that all these calculations are
1:16:44 correct because that's something I do in
1:16:46 my spare time. But uh it's not very fun.
1:16:48 So, okay, here we go.
1:17:10 Okay. So, now here's something I I
1:17:12 learned about Claude Code. It asks you
1:17:14 for permissions. Can I screw up your
1:17:16 computer? Yes, let's allow that.
1:17:24 Allow cloud code to run CD. Yes, you can
1:17:28 wipe my hard drive.
1:17:32 My cloud code has been running for 4
1:17:34 hours. Oh boy.
1:17:38 Yeah, I have a feeling this simulation's
1:17:40 going to take a while.
1:17:55 Yes, we're going to allow you to run
1:17:56 that.
1:18:15 Now building the entire simulation
1:18:17 engine. Let me write the core files.
1:18:20 Awesome. I'll tell you what, Claude. You
1:18:23 do that and I will sit here.
1:18:31 Allow Claude to write type.ts.
1:18:35 Yes, of course. We're writing Typescript
1:18:38 here, people. Everybody knows that.
1:18:42 Digital gods knows that
1:18:45 he's a god. Of course, he knows it.
1:19:05 It's writing archetypes. PR RNG. I don't
1:19:08 know what that is. And types.
1:19:28 Sorry. Lord digital gods. Yes.
1:19:31 Cloud is getting too much power.
1:19:52 Well, you done me. I felt it tried to be
1:19:55 chill. I thought I fell right through
1:19:58 the cracks. So I'm trying to get back
1:20:02 to run out. I'm giving it my besties.
1:20:05 Nothing going to stop me but an
1:20:07 adventure. It's be my turn to win some
1:20:11 or learn some but I
1:20:16 no more. No more. It cannot wait your
1:20:22 heart.
1:20:27 Yeah.
1:20:29 Yeah. Yeah.
1:20:34 Beating a Kyle in the post
1:20:36 postapocalyptic.
1:20:39 My machine's just sitting here coding me
1:20:41 a simulation
1:20:43 of the great repurpose.
1:20:45 Well, I can just dick around on the
1:20:47 guitar.
1:21:00 It's not simple to say the most days
1:21:05 I don't recognize me that these shoes
1:21:08 and this apron that place and its
1:21:11 patrons
1:21:12 have taken more than I gave them.
1:21:18 It's not easy to know I'm not anything.
1:21:22 It's coding up the future. And we're
1:21:24 just going to strum the present.
1:21:25 Exactly.
1:21:37 Hey
1:21:42 de
1:22:00 little de
1:22:03 little
1:22:06 D.
1:22:22 Pro tip, when vibe coding, do a
1:22:24 refractor more often than not. Panda,
1:22:27 that is awesome advice. If I knew what a
1:22:29 refractor was, I would be refracting
1:22:32 away. I would refract right now. In
1:22:34 fact, there would be no fracting. It
1:22:38 would only be refracting.
1:22:40 So, perfect. Oh, refactor. Oh, yes.
1:22:46 Refactor. That I've actually heard of
1:22:49 refracting. I had not heard of.
1:22:54 Tell it tell it to just go back and
1:22:56 refactor everything and and tighten
1:22:58 things up. That makes sense.
1:23:14 discombobulating.
1:23:16 Claude is discombobulating.
1:23:35 10,000 words swarm around my head. The
1:23:38 million more in books written beneath my
1:23:41 bed.
1:23:47 Wait, now the UI components. Is it Is it
1:23:50 working?
1:23:55 Did it stop? Oh, frolicking. Okay.
1:24:01 That's the wrong key.
1:24:10 Ah, I got way into what I was doing and
1:24:12 kept pushing it off, which now means I'm
1:24:15 paying for it. The refactoring.
1:24:25 I'm sorry to hear that, Joy Pretty. I
1:24:27 hope wish your mom nothing but the best.
1:24:30 Focus on taking care of your mama.
1:24:50 10,000 words swarm round my head many
1:24:54 more books written beneath my bed
1:25:02 wrote or read them all went searching in
1:25:05 the swarm still can't find how to hold
1:25:08 my hands
1:25:14 I know you need me in the next room
1:25:16 over. I am stuck in here all paralyzed
1:25:25 for months. I got myself in routes. Too
1:25:28 much time spinning mirrors framed in
1:25:31 yellow walls.
1:25:36 Ain't it like most people are more
1:25:39 different? Like talk on things we don't
1:25:42 know about.
1:25:47 Ain't it like most people? I'm no
1:25:50 different. like to talk on things you
1:25:52 don't know about.
1:26:09 You know, more than anything,
1:26:12 I hope that this AI nonsense really does
1:26:17 do things like cure cancer and just fix
1:26:21 some of this [ __ ] that we got to deal
1:26:22 with in these shitty human bodies,
1:26:25 in this shitty environment. I mean, we
1:26:28 are living longer than ever, but still.
1:26:30 God damn it.
1:26:37 Let's get some Let's get some [ __ ] fixed
1:26:39 in this world.
1:26:52 I mean, I'm looking at like the the the
1:26:55 health headlines in the past three
1:26:58 months have been bonkers. I want Bionic
1:27:01 Eyes. Exactly. Let's get Cam Ken some
1:27:03 bionic eyes
1:27:05 there. You know, it's funny, Cam. Some
1:27:07 of the some of the studies, one was a
1:27:09 peptide study, there was another one
1:27:10 that was some other specific thing that
1:27:13 was reversing aging in eyes and
1:27:16 restoring sight. Um, yeah, let's do it.
1:27:20 Bionic eyes, sight restoration, all of
1:27:23 it. Cure cancer, cure [ __ ] Lyme
1:27:26 disease.
1:27:29 It is time.
1:27:34 Yeah, Danielle, this is my main hope
1:27:36 from AI to heal diseases. Yep.
1:27:41 Let's do it quick. Wait. Allow once.
1:27:44 Allow claw to run.
1:27:47 Yes.
1:27:50 We're going to NP npx vite port sleep
1:27:54 curl head minus 20. We're gonna We're
1:27:58 going to do it. We're We're going in.
1:28:02 Yeah. The peptide studies sound
1:28:04 intriguing. Yeah. Did you see that one,
1:28:05 Dr. Jay, where it's this one peptide and
1:28:08 it was like repairing ligaments, curing
1:28:11 blindness. It It was like just anything
1:28:13 that was broken, it was fixing it, which
1:28:15 it was on Twitter, so you know,
1:28:18 always allow for the project.
1:28:25 Is there anything for beer goggle beer
1:28:27 goggle eyes?
1:28:36 Uhoh. What do we got here? Allow Claude
1:28:39 to write launch.json. Yeah, we're going
1:28:41 to launch this bad boy.
1:29:06 through a blue telescope.
1:29:11 Looking at the world tonight
1:29:14 through a blue telescope.
1:29:18 Oh, we got a simulation people.
1:29:22 All right, let me get rid of this. Let
1:29:24 me move this over. So, we got this
1:29:26 bigger. Is that it? Is that all I can
1:29:28 see? Preview screenshot. It wants to
1:29:31 launch the something at the directory
1:29:34 event feed. The app is running. I can
1:29:37 see the UI shell.
1:29:40 But the town canvas appears empty.
1:29:43 Likely it isn't rendering in its initial
1:29:45 state. No errors. The canvas is likely
1:29:48 not just triggering the initial render.
1:29:51 The issue is that render town, the
1:29:54 render town effect depends on the
1:29:55 current state, but no initial render.
1:29:57 The issue is that render Okay, it's it's
1:30:02 shit's happening.
1:30:05 Oh, it's Oh, this is cool. It's looking
1:30:06 at its own screenshots and debugging
1:30:09 itself.
1:30:18 You keep playing the first three notes
1:30:20 of a Rob Thomas song. Which Which one?
1:30:26 Which? Wait, I don't know which three
1:30:28 notes.
1:30:34 Oh, wow.
1:30:39 The camp distribution.
1:30:44 The simulation is alive and running. I
1:30:46 can see agents moving between zones
1:30:49 clustering
1:30:50 with camp colors. Oh, that's
1:30:52 fascinating. Okay, so the little the
1:30:55 little graph is everyone had their old
1:30:58 jobs
1:31:02 and then the green camp who's who's the
1:31:05 recclusterers or who's the repurposers?
1:31:09 Oh, and it's got clusters where they're
1:31:11 connected to other people. Oh, this is
1:31:14 very cool.
1:31:16 The zones are positioned
1:31:19 better spread, but the town still only
1:31:21 uses roughly the top half of the canvas.
1:31:43 Great progress. The simulation is
1:31:45 running with a compelling narrative
1:31:47 events. Myologists realize that they
1:31:50 lost what they'd lost as a copywriter
1:31:53 could be found again differently. The
1:31:56 camp distribution chart works, but I see
1:31:58 several issues to fix. Let me address
1:32:00 these now. Okay, so it's doing its own
1:32:03 thing. It's just building [ __ ]
1:32:22 You've hit your limit. Resets at 2 am.
1:32:24 Okay. Well,
1:32:27 there we go, people. All right. We've
1:32:29 created a simulation. It wasn't too bad,
1:32:31 but we did rate limit. Um,
1:32:35 can I
1:32:37 How do you select element
1:32:41 device toggle bar? Okay, it doesn't work
1:32:42 on mobile.
1:32:45 Let me reload it. Reload.
1:32:49 Play.
1:33:04 Stop.
1:33:07 Ada opened their door to neighbors in
1:33:09 need. Yara joined a protest at the civic
1:33:12 center. This isn't what we were
1:33:14 promised.
1:33:16 Wow.
1:33:19 Asha called a meeting. It was time to
1:33:21 coordinate. Beck embraced repurposing.
1:33:24 If the old meaning is gone, I'll build
1:33:26 new meaning.
1:33:28 Luna stopped mourning for the old role.
1:33:31 Something new is forming.
1:33:33 Josh Joss called a meeting. Rowan
1:33:36 organized a gathering.
1:33:39 Zarah retreated inward. The world felt
1:33:42 too loud.
1:33:44 Huh.
1:34:20 Week 20, week 30, week 40, 50.
1:34:38 Debbie wandered into unfamiliar
1:34:40 territory, literally and figuratively.
1:34:43 Talia began building something new,
1:34:45 tentative, but real.
1:34:51 It looks like all these are hardcoded,
1:34:55 huh?
1:35:03 So, it looks like it took 60 weeks for
1:35:05 for people to go from the old roll to
1:35:07 the new new roll and for it to basically
1:35:13 That's crazy.
1:35:18 Here's Iris. Oh, this is cool.
1:35:21 Resources, identity, purpose, belonging,
1:35:23 status, agency. Oh, wow. Wait. So, can
1:35:27 we How do you go back?
1:35:32 Oh, you reload, right? Yeah.
1:35:35 Okay. So, let's pick a person.
1:35:38 Let's pick a person who's not super
1:35:40 connected. Fria.
1:35:44 Let's go.
1:35:47 Robbie's decently connected. Jos.
1:35:51 Here's Yara.
1:35:59 Market Exchange, Contemplation Garden,
1:36:04 Abandoned District, Mutual Aid Hub,
1:36:09 Maker Space,
1:36:12 Learning Lab, Civic Center,
1:36:15 Legacy Workspace. Yeah, let's go to the
1:36:17 legacy workspace. We'll pick
1:36:20 who we going to pick.
1:36:23 Can I zoom in 4x?
1:36:28 None of this [ __ ] works. Okay,
1:36:38 go.
1:36:44 Current activities. Celebrate. Organize.
1:36:48 Work. Legacy. Trade. Organize. Explore.
1:36:50 Debate. Organize.
1:36:53 Explore. Seek community.
1:37:04 Huh?
1:37:14 That's wild. All right. Well, we have a
1:37:17 simulator.
1:37:22 Maybe we'll keep working on that.
1:37:26 Get a m Get a magnifying glass. Go out
1:37:29 in the sun. See what happens to your
1:37:32 Sims in the little tiny Sim ant farm.
1:37:39 All right, kids.
1:37:43 You're a cruel god.
1:37:45 Um, no. I what I want to do is I want to
1:37:48 I want to like one of the things that I
1:37:50 want to
1:37:52 simulate here is that some people are
1:37:55 going to withdraw
1:37:57 and disconnect and other people are
1:37:59 going to reconnect. And I don't know. I
1:38:02 don't know where I want this to go. I
1:38:04 should learn from it rather than give it
1:38:06 an agenda. But it's it seems like it's
1:38:08 just it's trend every everyone's
1:38:10 trending toward repurposing, which I
1:38:12 guess ultimately they will, but some
1:38:14 people are just going to check out,
1:38:15 right? Anyway, whatever. We'll figure
1:38:17 that out. We'll figure it out tomorrow.
1:38:20 All right, everybody. Um,
1:38:24 I hope you had a good time tonight.
1:38:25 Thanks everyone for your patience for
1:38:27 not having producer Brandon here if I
1:38:29 was screwing things up. Um, but we got
1:38:32 through it. So,
1:38:35 um, peace out. I will see you tomorrow.
1:38:38 Tomorrow's Thursday. Do I have anything
1:38:40 to talk about? Um,
1:38:43 tomorrow is the, uh, AI salon mastermind
1:38:46 practice lab. So, if you're in the
1:38:48 mastermind, we've got practice lab
1:38:50 tomorrow at 10:00 a.m. Mountain time.
1:38:53 All right. So, I'll see you there for
1:38:54 that. Beautiful. All right. Talk to you
1:38:56 later. Bye-bye.