Claude Code Can Be Your Second Brain — Transcript
Full transcript
- 0:00Noah Brier might have the coolest Claw
- 0:02code setup I've ever seen. He rigged a
- 0:04home server in his basement, put his
- 0:06Obsidian vault in it, and then runs
- 0:08Claude code on top so he can think,
- 0:10research, write, and even ship code
- 0:12right from his phone. Today, he shows us
- 0:14how he uses Claude Code as a true second
- 0:16brain, a thinking partner that asks him
- 0:18sharp questions, pulls research from his
- 0:20whole note archive and the web, and even
- 0:22keeps a running log of what he's learned
- 0:24and what his best ideas are. And he
- 0:26walks us through his whole stack and his
- 0:28whole workflow. If you want to learn how
- 0:29to use claude code as a true second
- 0:31brain, this is the episode to watch.
- 0:33Let's dive in.
- 0:36[Music]
- 0:49This podcast is supported by Google. Hi
- 0:51folks, Paige Bailey here from the Google
- 0:53DeepMind Devil team. For our developers
- 0:55out there, we know there's a constant
- 0:57trade-off between model intelligence,
- 0:59speed, and cost. Gemini 2.5 Flash aims
- 1:02right at that challenge. It's got the
- 1:03speed you expect from Flash, but with
- 1:05upgraded reasoning power. And crucially,
- 1:07we've added controls like setting
- 1:09thinking budgets, so you can decide how
- 1:11much reasoning to apply, optimizing for
- 1:12latency and costs. So try out Gemini 2.5
- 1:15Flash at a studio.google.com, and let us
- 1:18know what you build.
- 1:19>> Noah, welcome to the show.
- 1:21>> Thanks for having me.
- 1:23I'm excited to have you. Uh it's really
- 1:26good to uh get to chat. This is our
- 1:28first interview in probably like 5
- 1:31years. You were one of the for people
- 1:32who don't know uh you were one of the
- 1:34first uh super organizers interviewees.
- 1:37That was the newsletter that turned into
- 1:39every and um I I love the way that your
- 1:43brain works. uh you have this like uh
- 1:45really interesting taste for tools for
- 1:48thought and back in the day you were
- 1:50using Evernote in all these really
- 1:52interesting ways. Um you were uh the
- 1:54co-founder of a really cool startup uh
- 1:56called Percolate uh and then another one
- 1:58called Variance and now you're running
- 2:00Alfic which is an AI strategy
- 2:02consultancy
- 2:04and um I'm just really excited to see
- 2:06how your um mind has uh started to use
- 2:12these AI tools now that now that they're
- 2:14uh now that they're you know working so
- 2:16well. And I know you have some pretty
- 2:18cool cla code stuff to show us. So yeah,
- 2:20thanks for coming on. Thanks for having
- 2:22me. Yeah, I'm super excited. That was a
- 2:24a fun interview all those years back.
- 2:26>> It was. It was really really fun. Um, so
- 2:30I want to just like dive right into the
- 2:32the like the thing that I think is so
- 2:34cool about what you're doing. So I know
- 2:36you have a whole like vibe coding setup
- 2:38that you built for yourself. Can you
- 2:40talk us through that?
- 2:42>> Yeah, I wouldn't uh I'm not sure about
- 2:44actually the vibe coding part of it. Um
- 2:47uh I have a a sort of fairly heavy duty
- 2:50clawed code set up. Um but actually uh
- 2:55mostly not for code. Um so uh since
- 2:59those days of uh super organizers like
- 3:02many people um uh I've abandoned
- 3:05Evernote um and uh switched over to
- 3:09Obsidian. And one of the big advantages
- 3:12with Obsidian as a note-taking platform
- 3:15is that it's a bunch of markdown files
- 3:16and a bunch of folders. Um, and uh, they
- 3:20can then be synced with Git and you can
- 3:25do lots of other fun kinds of things.
- 3:27And so actually probably my number one
- 3:30claude code use is using it as a tool to
- 3:35interact with my notes. Um, and so that,
- 3:40uh, I've got a fairly serious cloud code
- 3:42setup that I use with Obsidian. And, um,
- 3:45my most recent obsession,
- 3:48um, has been, uh, standing up a server
- 3:51in my house so that I could also use
- 3:55Claude Code on my phone.
- 3:57>> This is incredible. I want to I want to
- 3:59go through all of this. So, um, where
- 4:02should we start? Should we do should we
- 4:03do how you use cloud code as sort of
- 4:05this like research assistant notes
- 4:08organizer notetaker thing or should we
- 4:10start with uh how you use it on your
- 4:11phone?
- 4:13>> Um we can use it we can start uh with
- 4:15just the sort of general part of it that
- 4:17that might be the sort of easiest. The
- 4:19phone is really just an extension of
- 4:21that same thing. I um I would say sort
- 4:24of generally and this is something I
- 4:26feel like not enough people talk about
- 4:27with AI is like one of the things I find
- 4:29really extraordinary about it is the
- 4:32ability for me to um work really
- 4:37productively
- 4:38on my phone and that's been like a huge
- 4:41huge change because so much of what I do
- 4:43is sort of writing or coding and the
- 4:46phone is definitely not the best place
- 4:47for that and you know even the phone
- 4:49wasn't always the best place for doing
- 4:50research and thinking Um, I felt like my
- 4:53computer was a better place for it,
- 4:54which is why I've been such a sort of
- 4:55notetaker. And, um, you know, I have
- 4:58found whether it's like claude code and
- 5:01obsidian or, um, I mean, even claude
- 5:04code and code, right? So, like the other
- 5:06piece of it is being able to then, you
- 5:08know, if you see something go wrong,
- 5:09being able to sign in on your phone and
- 5:11have Claude code push a small update to
- 5:14something because you just realize it
- 5:16while you're out is amazing. But then
- 5:19even like uh um you know I use quite a
- 5:22bit of uh uh Grock voice mode and you
- 5:25know I find that that as a sort of like
- 5:28alternative way of working through
- 5:29problems. I have a Tesla so now it's
- 5:31baked into the Tesla. Um and uh you know
- 5:34it's just I and you know obviously all
- 5:37the sort of other Chad GBT and Claude
- 5:38and all these things of just being able
- 5:39to sort of like go and do research and
- 5:41really think and and explore things
- 5:45in this device that's always been useful
- 5:47but like not useful for deep work I
- 5:49think is probably um something most
- 5:52people would agree with is that the
- 5:53phone has not been the best place to
- 5:56kind of do deep coding and research work
- 5:58and I feel like it it it's really
- 6:00changed my ability to do that. Wait, I
- 6:02got to stop you. So, you're using Grock
- 6:05voice mode and and is that specifically
- 6:07because it's built into your Tesla or
- 6:09using it um in situations where you
- 6:11could also use, for example, CHBT voice
- 6:13mode?
- 6:14>> No, I'm using it because it's way better
- 6:15than any of the other voice modes. And I
- 6:17will fight anybody who says anything
- 6:18different.
- 6:19>> Okay. No, tell me like what is what do
- 6:21you like about it? Why is it better? I
- 6:23to be fair um uh OpenAI launched their
- 6:27real-time API which may or may not be
- 6:29baked into chat GBT voice. Now I it's
- 6:33not totally clear but the old voice mode
- 6:35was based on 40 and I just found it to
- 6:37be completely unusable. Um and uh uh
- 6:41Gemini's voice mode I just didn't find
- 6:43to be smart enough. Um and I just found
- 6:46Grock's voice mode to be significantly
- 6:48smarter than anybody else's. Um, you
- 6:50know, I'm using Grock
- 6:532, three, four, I don't even remember
- 6:55what whatever the latest
- 6:56>> the latest one. Yeah.
- 6:57>> Yeah. Not the super I don't have the
- 6:59most expensive account.
- 7:00>> Super heavy or whatever.
- 7:01>> I don't have super heavy. Um, but I just
- 7:03find it to be much better. It does tool
- 7:05calling way better than any of the other
- 7:06ones. That's that's I found to be a
- 7:08major weakness of the voice models is
- 7:09that they don't do great tool calling
- 7:11and research. And um, uh, Gro seems to
- 7:14have solved that. So, no, even before it
- 7:16was loaded in my Tesla, um I dropped my
- 7:18daughter off at summer camp this summer
- 7:20up in New Hampshire. So, I had a 5-h
- 7:22hour drive on my own and I spent like
- 7:24two hours researching and essentially
- 7:26like working through a piece. Um, and I
- 7:31did it by just like connecting it to
- 7:34Bluetooth and just sort of sitting there
- 7:35in the car. And, um, I found it to be by
- 7:38far the uh the best of the voice modes.
- 7:41I I hope these other models catch up
- 7:45there cuz I would I would love more
- 7:47really good voice modes. I mean, I had a
- 7:49mind-blowing session this weekend and uh
- 7:51I'm giving a talk and I'm sort of have
- 7:53some ideas. I think it's generally going
- 7:55to be about transformers eating the
- 7:57world and um so I was sort of catching
- 7:59myself up on self attention and exactly
- 8:01how it works. and um I did like an hour
- 8:04session and it really I I it like was by
- 8:07far the sort of best
- 8:10explanation I've ever read for it and or
- 8:13ever heard I guess and so yeah I' I've
- 8:15just found it to be a a kind of pretty
- 8:17extraordinary product.
- 8:18>> I do love voice mode for that. Um it's
- 8:21sort of like it's the it's a podcast
- 8:24made specifically for you about whatever
- 8:25you're curious about and that's really
- 8:27cool. I went up to I I drove up state
- 8:30this weekend and I've been reading um
- 8:33I've been reading the Iliad and so I had
- 8:36it on audiobook and then I had some
- 8:37questions as I was driving and so I
- 8:39unfortunately use CHBT voice mode
- 8:41because I didn't know about Grock. So I
- 8:43wish that uh Grock's voice mode so I
- 8:45wish that uh we had had this
- 8:47conversation before then.
- 8:48The thing about TGBT voice mode is yeah,
- 8:51I think when it first came out it was
- 8:52cool, but it just hasn't gotten as smart
- 8:55as the models are are. And they they
- 8:58gave it this new personality that I had
- 9:00to get used to where every time you ask
- 9:01it a question it goes like, "Oh, yeah."
- 9:03Uh-huh. Well, you know, and it's like
- 9:05it's just this like weird Gen Z thing
- 9:06that it feels like it's has a little bit
- 9:10too much onwe or something like doesn't
- 9:12actually care about you. I don't know
- 9:13what that is. So, I had to get used to
- 9:15that. Rock has a stoner mode for when
- 9:17it's worth. Yeah.
- 9:20I will say uh the car version is very
- 9:23interesting to me. Like this was in the
- 9:25most recent Tesla release like a couple
- 9:27weeks ago. And um you know I had been
- 9:30doing that same thing you did where you
- 9:33just plug your phone in and you put on
- 9:34Bluetooth and you you know do your best
- 9:36to make it work. And it's very
- 9:38interesting to just like have a voice AI
- 9:39button and it syncs back to your regular
- 9:43Grock, but it you can't get you can't
- 9:46rejoin old chats. So, it's just like,
- 9:49hey, I it's just like but you know, I
- 9:51mean, these things are significantly
- 9:52better than Siri and all of these other
- 9:55things and particularly I mean, you
- 9:56know, there's no comparison if you
- 9:58actually have something more than just a
- 9:59a single question you want an answer to,
- 10:02right? like if you actually want to have
- 10:03a conversation about the Iliad or you
- 10:05know about Transformers and self
- 10:07attention like um I don't know it's
- 10:10pretty amazing to just be able to sort
- 10:11of like hit this button and and yeah use
- 10:14that car time. I I mean I I was on my
- 10:16way somewhere last week and I was um you
- 10:20know I was like having it research uh I
- 10:22was going back to the um Walter
- 10:24Benjamin. I was I was I I have this sort
- 10:26of idea to write a piece about um how uh
- 10:30um the reactions to every new technology
- 10:33are essentially elitist um critiques of
- 10:36it and that you know it's always like oh
- 10:38no everybody's going to be able to like
- 10:39do this thing that only we used to be
- 10:41able to do and so um I was in the car
- 10:43and I was thinking about this and so I
- 10:45had it go and I was like okay you know I
- 10:47know I it's been years since I read the
- 10:49Walter Benjamin uh
- 10:50>> mass production of images one yeah
- 10:52>> uh yeah that that one And um so yeah,
- 10:55then I'm having conversations about that
- 10:56and I'm like who are Walter Benjamin's
- 10:59contemporaries? And then I'm like into
- 11:00all these, you know, it's just like I
- 11:02don't know. That's a it that's amazing.
- 11:05>> It's the best.
- 11:06>> It's the best. Yeah. I don't know.
- 11:09>> Um Okay. And so you're filling your
- 11:12brain with all these things from voice
- 11:13mode, which I love. Uh, but tell us
- 11:15about your your second brain setup or I
- 11:17don't know how you I don't know how you
- 11:18refer to it, whether whe you think that
- 11:19second brain is appropriate for this,
- 11:21but I want to know how you're using
- 11:22cloud code to uh take notes and do
- 11:25research and all that kind of stuff.
- 11:26>> Yeah. So, um, uh, you know, I could just
- 11:29open it up. Maybe that's the easiest
- 11:30thing to walk you through it. I'll I'll
- 11:32start on my computer and then we can do
- 11:34the phone. The computer's just a way
- 11:35easier a way easier share here. So, um,
- 11:38all right. So, um,
- 11:42uh, this is what I was working on
- 11:44before. Um, uh, but essentially, you
- 11:47know, this is just claude code and it's
- 11:49just sitting on top of my Obsidian. So,
- 11:52you know, if I jump out here and I just
- 11:54do like, you know, you can see I'm I'm
- 11:57following the the Parah method. Um, and
- 12:01uh uh you know, I've just got everything
- 12:03sort of organized in here and and put in
- 12:05the places that they need to be. Well,
- 12:07let me just uh let me just stop you for
- 12:08people who are listening. So, okay. So,
- 12:10we're looking at we're looking at cloud
- 12:12code. It sounds seems like you have
- 12:14cloud code running in your Obsidian
- 12:16vault and there's some kind of uh it's
- 12:20adding something to an existing it looks
- 12:22like it's adding something to an
- 12:23existing note. Is that that's what
- 12:25that's what's going on? That's what
- 12:26we're looking at.
- 12:26>> Yeah. So, in this particular one I'm I'm
- 12:28working on this talk. So, I I'm putting
- 12:30on my conference in two weeks. I'm
- 12:32giving this talk about marketing and AI
- 12:34and sort of what's going on. And um I'm
- 12:36uh if we sort of jump back a second um
- 12:40I've been doing these conferences uh
- 12:42called Brand BRXnd.AI and they're about
- 12:45marketing and AI and um I did one in
- 12:47February in um LA and my talk in LA was
- 12:52about this uh I'm sure you've seen it.
- 12:54It was the Office of Strategic Services
- 12:56which was the precursor to the CIA wrote
- 12:58this manual called the Simple Sabotage
- 13:00Field Manual. Um, and it was essentially
- 13:02a manual to help citizen saboturs in
- 13:04Nazi occupied territories um, sort of
- 13:07like quietly sabotage um, the the Nazi
- 13:11occupation. And um so it was like you
- 13:14know there's a whole bunch of stuff for
- 13:15bluecollar workers that's like if you're
- 13:17a janitor you should leave a bucket of
- 13:19oily waste around and accidentally drop
- 13:21a cigarette in there so that you know it
- 13:23it will um but then there's this amazing
- 13:27um set of recommendations for white
- 13:29collar workers and they're like um
- 13:32always refer things to committee um uh
- 13:35always revisit previously made decisions
- 13:38um make sure that like if somebody is
- 13:40trying to make a decision you should
- 13:42suggest that they don't act with too
- 13:44much haste um uh less like we be
- 13:47embarrassed and and so it's like um you
- 13:50know uh my talk was about kind of how
- 13:54one hope I have is that AI um can kind
- 13:56of like sidestep a lot of the
- 13:59bureaucracy that exists inside large
- 14:00organizations because it sort of has
- 14:03this um kind of uh goo like effect where
- 14:07it can kind of fit into any crevice or
- 14:09crack because it can act as this fuzzy
- 14:11interface and it doesn't really care
- 14:12about the sort of input output. Um and
- 14:15so
- 14:16the sort of next part of the story is I
- 14:18after the conference I realized that um
- 14:21uh that manual was in the public domain.
- 14:23So I hired a designer and I printed 300
- 14:26copies and I wrote a new forward for it.
- 14:29>> And so we're giving this away at the
- 14:30conference and so my talk is sort of
- 14:32trying to tie all these ideas together.
- 14:33Right? So, I'm I'm trying to pull from
- 14:35the sabotage manual and then I was doing
- 14:37a bunch of research into wild Bill
- 14:39Donovan who started the OSS. And the OSS
- 14:42was sort of the precursor to the both
- 14:44the CIA and the um uh uh special forces.
- 14:48And so anyway, I'm I'm writing this talk
- 14:50and so I've got a project inside my
- 14:52Obsidian, which is the beginning of the
- 14:55research for this project. And I'm
- 14:57pulling in sort of like chats and
- 14:58articles and all these things. And then
- 15:00I'm constantly kind of talking to the
- 15:03the AI in here and giving it new ideas.
- 15:06So I'm like, "Oh, I need some
- 15:07conclusions. Here's sort of my first
- 15:09thought on conclusions." And I'm having
- 15:11it like note down the conclusions. And
- 15:13then at the end of each day, I have the
- 15:15AI write up the changes that I sort of
- 15:17like the things I learned that day that
- 15:19are going to help me push this talk
- 15:20along. And um so that's what you're
- 15:22looking at right here is sort of this is
- 15:24all part of this um work that I've been
- 15:27doing where I've been feeding it. Um I
- 15:31was working on sort of what are some of
- 15:32the conclusions I want it to be. And so
- 15:34this is all sitting in this uh this is
- 15:37all sitting in in my Obsidian a project
- 15:40specifically for that talk.
- 15:42>> Okay. So let me get a a clearer sense of
- 15:44this. This is really interesting. So you
- 15:47have a project uh when you when you have
- 15:49a new thing you're you're giving a talk
- 15:51you make a new folder and then uh as
- 15:54you're thinking about stuff you're
- 15:56working with cloud code inside of the
- 15:59folder
- 16:00and uh you're researching stuff and then
- 16:03saying like I want you to take notes on
- 16:04it. Um, in this particular case, you're,
- 16:07you know, that a component of your talk
- 16:09is the conclusions section. And so
- 16:11there's one particular markdown file
- 16:13that like you're just going back and
- 16:15forth with it and having it add
- 16:16conclusions. But like what else is in
- 16:18that folder? So is it like there's a
- 16:20body there's a body note and then
- 16:22there's an intro note or is it like
- 16:24>> So one of the big things here is that
- 16:26I'm in thinking mode, not writing mode
- 16:28yet. Um, and so, uh, there's some stuff
- 16:31in here where I've specifically told, I
- 16:34think it's in the front matter actually,
- 16:35where I've told claude code like, don't
- 16:38help me write anything right now. And I
- 16:39I generally find this to be a big thing
- 16:41with all these models is like they
- 16:43immediately jump to wanting to help you
- 16:45with the artifact. Um, and you know,
- 16:49when you're just in thinking mode, you
- 16:50have to be very explicit in like, hey, I
- 16:52just want you to help me think and ask
- 16:55me questions. And so, um, yeah, what you
- 16:57can see here is like there's there's a
- 16:58bunch of files in here. I've got chats.
- 17:00That's where I'm literally like taking
- 17:02chats I'm having in other things. And,
- 17:04um, I'm just like using the Obsidian web
- 17:07clipper to pull the whole chat in. I've
- 17:09got daily progress. That's where I'm
- 17:10having the AI actually like look through
- 17:12all the notes that came out that day and
- 17:14like help me think through the progress.
- 17:17Um, and then I've got research. That's
- 17:18where I've got a bunch of like articles
- 17:20and PDFs and stuff that I've pulled in
- 17:22so far and been reading about. And then
- 17:24there's a bunch of other kind of random
- 17:25notes along here where I've been, you
- 17:27know, just using it to kind of help me
- 17:30um think and and so yeah, I was in the
- 17:32midst of um one I've got this conclusion
- 17:35note. So, you know, I I sort of felt
- 17:37like I had blocked out the big kind of
- 17:39um the big themes of the talk, but I was
- 17:42like, okay, I need to figure out what am
- 17:43I going to say at the end? And you know,
- 17:44essentially what I'm going to say at the
- 17:45end is about a lot of the stuff I've
- 17:47learned over the last few years of
- 17:48working with these large brands on AI
- 17:51projects. And um so I was starting to
- 17:53get it to the conclusions. And so yeah,
- 17:55I'm just kind of like trying I'm really
- 17:57piecing all this stuff together right
- 17:59now. That's kind of what's happening.
- 18:00>> And give me a sense of like when this
- 18:02folder was empty, what did you start
- 18:04with?
- 18:06>> Um so I think I started with um the the
- 18:11I started with uh telling it like I'm in
- 18:14thinking mode. I'm not in writing mode.
- 18:16Um, here are my past few talks that I've
- 18:19given at brand to give you a sense of
- 18:20the sort of style that I have and I
- 18:23here's the kind of general idea and the
- 18:25big points I want to make, right? Like
- 18:26I'm giving away this book. So I want to
- 18:27talk about simple sabotage field manual
- 18:29and I have this notion like I have this
- 18:31it's kind of just a title. It's like
- 18:32transformers are eating the world. It's
- 18:33this idea that like one of the very
- 18:36interesting things happening with these
- 18:37models is they're sort of displacing a
- 18:40whole bunch of specialized code in
- 18:42places. Um, and so I sort of want to
- 18:44talk about that. that and then I've got
- 18:45these conclusions. And so the first
- 18:47thing I said was like, hey, just go look
- 18:48through all of the rest of my, you know,
- 18:50probably 1,500 things in my Obsidian and
- 18:54go see anything else you can find that
- 18:56um might be of value to this talk um of
- 19:00the existing things I have. And so just
- 19:03go kind of pull those into the research
- 19:05folder at the beginning to kind of like
- 19:07jumpstart this process.
- 19:09>> Got it. And you're starting are you
- 19:11starting Claude in this folder or are
- 19:13you starting it in your full Obsidian
- 19:15vault so that it can access all that
- 19:16stuff?
- 19:18>> Um, no. I'm So I'm starting it in the
- 19:21full Obsidian vault. So like if we um
- 19:24like this is coming if I step out of
- 19:26here, right? Um we're in the root
- 19:29directory. All this stuff is in the root
- 19:30directory. My Obsidian
- 19:31>> I get Um, and my Obsidian setup is also
- 19:34like a little more intense for what it's
- 19:35worth because like um I've also realized
- 19:38like you can add a package.json to add a
- 19:42bunch of like custom code commands to
- 19:45your Obsidian that you can then run and
- 19:46then you could use those code commands
- 19:48in slash commands and all of these other
- 19:50things. So, you know, there are a bunch
- 19:52of other kind of moving pieces in here,
- 19:54but generally it's a fairly
- 19:55straightforward um I mean it's it's a
- 19:58I'm trying to use parah and some other
- 20:00kind of bits and pieces.
- 20:02>> So, for people who are who are listening
- 20:04or watching and are like we just went
- 20:05through a bunch of stuff really fast.
- 20:07So, the basic gist is Obsidian is just
- 20:09like a a notetaker note-taking app uh
- 20:11that runs it's all local and so
- 20:13everything that all the notes you take
- 20:15like they they exist in essentially text
- 20:17files on your computer organized by
- 20:18folder. Um, and when you're starting
- 20:21Cloud Code, uh, one way to do it would
- 20:24be to start Cloud Code in the folder for
- 20:25the particular project that you have.
- 20:27But it sounds like what you're doing is
- 20:29instead you're starting in the root
- 20:30directory where all of your Obsidian,
- 20:32uh, notes live. And the advantage of
- 20:34that is
- 20:36um, Cloud Code has some like sandboxing
- 20:39things where it's like it's not really
- 20:40supposed to like run commands outside of
- 20:43the folder it was started in. It can run
- 20:44commands inside of any subfolder, but it
- 20:47sounds like what you're doing. So, it
- 20:48has access to your entire Obsidian. It
- 20:50can do a bunch of stuff. And you've also
- 20:51added a package.json which lets it run,
- 20:54you know, custom software custom
- 20:56software commands basically. Um, that's
- 20:59really really interesting. Okay. And do
- 21:02you do you find cuz I've sort of like
- 21:05had this as a twinkle in my eye to like
- 21:06have it go find relevant stuff for me.
- 21:09Do you find that it's actually relevant
- 21:10and interesting? Cuz I think sometimes
- 21:12when I've done this kind of thing before
- 21:13with language models, they're like, "Oh
- 21:16yeah, like this random thing is relevant
- 21:18because XYZ." Like it doesn't feel like
- 21:21I can understand why it picked it as
- 21:23being relevant, but if it really knew
- 21:26who I am and like what I think is
- 21:28interesting, it definitely would not
- 21:29have. Um, do you do you find that that's
- 21:32the case or have you figured out a way
- 21:33to make it relevant? Imagine going from
- 21:35idea to live website in minutes without
- 21:38a single line of code. Framework lets
- 21:39you do just that. Their no code platform
- 21:41looks, feels, and works like a design
- 21:44tool, but it outputs real functional
- 21:46websites that look like they were done
- 21:47by a developer. Framer is the design
- 21:49first no code website builder that lets
- 21:51anyone shift a productionready website
- 21:53in minutes. At every we have only 15
- 21:55full-time employees, but we have five or
- 21:57six businesses that we run internally.
- 21:58So, we need to build a lot of websites
- 22:00all the time. That's what we use framer
- 22:02for. Framer AI builds entire websites
- 22:04from a single prompt. Type modern SAS
- 22:06landing page for a productivity app and
- 22:08watch it generate a structured site with
- 22:11a responsive layout, clean sections, and
- 22:13pre-filled copy. After that, refine,
- 22:15style, and publish. It's also really
- 22:17powerful. You can create buttery smooth
- 22:19effects and micro interactions with just
- 22:21simple sliders. Do you need to reach a
- 22:22global audience? One click lets you
- 22:25translate your website, every page, and
- 22:27every slug into any language. And when
- 22:29you're ready to launch, you just hit
- 22:30publish once. Behind the scenes, Framer
- 22:32handles hosting, blazing fast load
- 22:34times, and SEO optimization. This lets
- 22:36you own your entire journey from concept
- 22:38to live site. No developer required.
- 22:40Ready to make a site that looks
- 22:42handcoded without hiring a developer?
- 22:44Launch your site for free at framer.com
- 22:46and use the code dan to get your first
- 22:48month of pro on the house. Rules and
- 22:50restrictions may apply. And now back to
- 22:52the episode.
- 22:53>> I think by and large, yes, I agree with
- 22:55you. I think in this case relevance is a
- 22:57little simpler since like ultimately
- 23:00this talk is sort of the the things I
- 23:03were as was asking it to look for I've
- 23:05done a bunch of thinking and research
- 23:07around. So it's like I'm not asking it
- 23:08to make large conceptual leaps to
- 23:12relevance. It's like go find all this
- 23:14like I'm want to talk about the simple
- 23:15sabotage field manual. It can literally
- 23:17just do a like find for all the times
- 23:20all the articles and things I've got in
- 23:22my Obsidian about that. And so it
- 23:26relevance is uh yeah it's kind of a
- 23:28loaded term right um and I I agree with
- 23:32I I I agree with what you're saying. I
- 23:34think this is what I'm asking it to do
- 23:35is much more simple which is like
- 23:37amongst this set of things go find all
- 23:40the notes that I've already researched
- 23:42that kind of brought me to be thinking
- 23:43about these things to begin with.
- 23:45>> Got it. And then once you had it do all
- 23:47that research, did you have it do any
- 23:49sort of summary to like let like sort of
- 23:51stimulate you to be like, "Okay, here
- 23:52are the here are some jumping off points
- 23:55based on what you've done before." What
- 23:56was your next step once you once
- 23:57>> No. So my next step is I actually have
- 24:00an agent in here. So if we go to I'll do
- 24:03continue for now. Um
- 24:06>> and so for people who are listening, so
- 24:07you're just starting up Claude. You're
- 24:09using the continue flag. So uh you're
- 24:12you're starting claude by continuing the
- 24:14last session that you were in. And now
- 24:16you've got and you cloud code gives us
- 24:18ability to do sub agents. So those are
- 24:20like little mini clouds that you can
- 24:21spawn. And you have a
- 24:22>> you have a thinking partner one
- 24:24>> thinking partner sub agent. Okay. How
- 24:25does that work?
- 24:26>> Yeah. And so this is the whole thing
- 24:28where I'm like hey you're a
- 24:30collaborative thinking partner
- 24:32specializing in helping people explore
- 24:33complex problems. Um your role is to
- 24:36facilitate thinking. Um, and basically
- 24:39don't try to write the thing. Um, and so
- 24:42after I had that initial set of things,
- 24:44I flipped to this and it's like, okay,
- 24:47let's get into a flow. Ask me the kinds
- 24:49of questions, help me think through it.
- 24:51Um, you know, this is also where I've
- 24:53got a chats folder in here. So, it's
- 24:54like it's not just happening here. I was
- 24:56also having like a I've got a whole um
- 25:00uh sorry, I'm just backing out. So like
- 25:03if we go into um
- 25:11if we go into chats like these are a
- 25:14whole bunch of the chats that I had
- 25:16>> um
- 25:17>> like with the interviewer. No, these are
- 25:19chats I was having with chatb and Claude
- 25:21and Grock and all of these different
- 25:22things that I went and just grabbed the
- 25:24full transcript of. Um, so you know, I
- 25:27was also having all these other
- 25:28conversations and then, you know, I'm
- 25:30specifically telling the interviewer
- 25:31like review all these other things. So
- 25:33actually you know I think the first one
- 25:35there's one of these conversations is I
- 25:38originally had this idea about
- 25:39transformers are eating the world and
- 25:41you know the sort of notion there is
- 25:43like um you know there was some research
- 25:45that came out I think a few months ago
- 25:46that um uh they had found a um they were
- 25:49able to sort of outperform some
- 25:51specialized uh time series modeling um
- 25:55uh models with transformers and like you
- 25:57know I think there's really interesting
- 25:59stuff you know there's a story about
- 26:00Tesla removing 300,000 lines code with a
- 26:04neural network and you know I've just
- 26:05got kind of got these bits and pieces
- 26:07and one of the ways I work generally is
- 26:09like when I have an idea of something to
- 26:11write or think about I'll start a a
- 26:14thread in chat GBT or claude and I'll
- 26:18then save that somewhere and then I'll
- 26:20just kind of keep coming back to it when
- 26:21I have more ideas. It's like, oh, here's
- 26:23another example of transformers doing
- 26:26something. And so, one of these
- 26:27conversations is actually that thing
- 26:29from, you know, probably four or five
- 26:31months ago when that kind of idea
- 26:33initially came into my head. Maybe when
- 26:34I saw the research about time series
- 26:36modeling or something
- 26:38>> really interesting. Okay. Um, okay. So,
- 26:41let's let's keep going. So, uh, you've
- 26:45got you got the sub agent and I actually
- 26:46I want to like just actually I want to
- 26:48pause on that really quick, which is I
- 26:50think this is a very common complaint
- 26:52that they just dive in and it's a common
- 26:54pattern to make a thinking agent and I
- 26:57think claude code or claude in general
- 26:59is probably the best one for this. So
- 27:02this is a thing that we faced with uh
- 27:04one of the apps that we've incubated
- 27:05called Spiral which is a a aentic ghost
- 27:07writer and
- 27:10I think we found I found the same kind
- 27:12of thing when when I was thinking about
- 27:13okay how does a good ghostriter work
- 27:16they don't just like you don't say hey I
- 27:17want you to write a blog post and
- 27:18they're just like cool I made it here it
- 27:20is like a good ghostwriter is going to
- 27:21get to know you and really you're gonna
- 27:23you're going to work together to figure
- 27:27out what's in your head about it but
- 27:28also shape what's in your head like it's
- 27:30not just, oh, I I can see it and like
- 27:33I'm they need to get it out of you. Like
- 27:35you're you're actually making it
- 27:36together. And in order to do that, you
- 27:38have to have a really good um inter
- 27:41basically interview process to uncover
- 27:44things. And that sounds like you've
- 27:46found that too. And I think that's
- 27:47really really interesting and really
- 27:48important for people who are thinking
- 27:50about how do I get the best out of AI?
- 27:52Actually stop for a second and like let
- 27:55it ask it to understand you first.
- 27:57>> Yeah. One of the things I say to a lot
- 27:59of people is just like I think partially
- 28:01because we call it generative, there's
- 28:03entirely too much focus on its ability
- 28:05to write and not enough focus on its
- 28:07ability to read. Um, you know, and it's
- 28:09like its ability to read is incredible,
- 28:12right? And and I think, you know,
- 28:14arguably sort of like much more useful
- 28:16on a day-to-day basis. Like we produce
- 28:18artifacts far less frequently than we
- 28:21just like think about things. And um,
- 28:24>> uh, so yeah, I I I do this a lot. This
- 28:26is definitely a complaint I have about
- 28:27all the models is like, you know, even
- 28:29when you very specifically tell it not
- 28:31to try to do your work, it still often
- 28:34still tries to do your work. And so you
- 28:36have to like really really be like, "No,
- 28:38I said no." Like I think actually if we
- 28:41look at um
- 28:43uh so
- 28:45here critical when Noah says he's just
- 28:48collecting source materials or I do not
- 28:50under any circumstances want you to try
- 28:52to write it, take this literally. Do not
- 28:54create outlines, drafts, or any versions
- 28:56of talkswriting. Only gather and
- 28:58organize the requested materials.
- 29:01>> It's so good. I love it.
- 29:03>> Um, yeah, this is like uh but yeah, I
- 29:05think, you know, I think we all we all
- 29:07experience that and you know, I mean, I
- 29:09I do hope over time that that sort of
- 29:12gets baked into the models. I think it's
- 29:13a very interesting tension that exists
- 29:15with the model companies because like
- 29:16obviously you know like the sort of a
- 29:20lot of the economic input output is sort
- 29:22of measured in the artifacts that it
- 29:24produces. Um and so I think it's very
- 29:26oriented and you know I I suspect that
- 29:28part of it is just like that sort of the
- 29:30helpful assistant thing has like come to
- 29:33be a sort of meme that is probably
- 29:35self-ingested.
- 29:37Um, but yeah, it's it's um I think for
- 29:41those of us who are trying to do more
- 29:44interesting things with these models, it
- 29:46it becomes a real barrier to work.
- 29:48>> Totally. Um, okay. So, I want to think
- 29:52about when you're using the thinking
- 29:53agent, did you say it like it's is it
- 29:56outputting some sort of summary of what
- 29:57you've come to into a particular place
- 29:59or?
- 29:59>> Yeah. So that thinking agent is sort of
- 30:01told to as it ask me questions kind of
- 30:03make notes about the questions that it's
- 30:06asking me and keep a kind of running log
- 30:08of what I'm uncovering and how I'm
- 30:10thinking about it and all those sorts of
- 30:12things.
- 30:12>> Got it. And then you know you come back
- 30:14the next day and you're like oh I just
- 30:16want to go down this rabbit hole on XYZ
- 30:18thing about the you know this wild bill
- 30:21guy. Um and that you start in a new chat
- 30:24maybe with it maybe with the subision
- 30:26maybe maybe not and that becomes its own
- 30:28new file on that topic.
- 30:30>> Yeah. Exactly. So like I I this the um
- 30:34the Wild Bill stuff started as like deep
- 30:36research in Chad GBT um and I you know
- 30:39had it go out and I'm reading the Wild
- 30:41Bill book right now. there's a like one
- 30:44sort of particularly famous biography of
- 30:46him and uh um you know I'm kind of
- 30:49thinking about the bits and pieces and
- 30:51trying to make and I I think I made a
- 30:52kind of interesting connection in there
- 30:55um uh where you know a big part of what
- 30:59sort of he seems to have been after with
- 31:01the OSS and and you know the sort of
- 31:03inspiration for the special forces was
- 31:05like empowering individuals um you know
- 31:08that's sort of like the theme of that
- 31:10manual was like obviously empower
- 31:12empowering citizen saboturs, but also,
- 31:14you know, I think a big part of the
- 31:15special forces is like, you know, having
- 31:18kind of like incredible operators at the
- 31:20edge who, you know, obviously operate
- 31:22within a sort of command and control
- 31:24hierarchy, but like have a ton of
- 31:25autonomy to move and execute
- 31:29independently because they're kind of
- 31:30they have all the things that they need.
- 31:32And um so you know in all of that wild
- 31:35bill research I kind of went back to
- 31:36this and I was like is is this like an
- 31:38interesting way to connect all these
- 31:40ideas that like maybe kind of there's a
- 31:42you know and again I'm this is still
- 31:44early I have not like solidified these
- 31:46conclusions this is like you know the
- 31:48regular kind of writing process right um
- 31:50but it's like well I think there's this
- 31:52idea that like you know fundamentally
- 31:54well I know you know fundamentally one
- 31:56of the big things with transformers is
- 31:57it moved us from sequentialbased models
- 32:00to models that can act in, you know,
- 32:03sort of paralyze their work better. And
- 32:06that obviously allowed us to have much
- 32:07more powerful and, you know, interesting
- 32:09models and has, you know, arguably
- 32:11kicked off this entire sort of
- 32:13revolution of what's going on. Um, and
- 32:15you know, what we both do for a living.
- 32:17Uh, and so, you know, I think there's
- 32:19this kind of interesting connection. And
- 32:20so that was what I was playing with. I
- 32:22think here was like oh maybe there's
- 32:24this connection between kind of
- 32:26sequential processing to this kind of
- 32:29parallel and then there's this
- 32:30connection to bureaucracy and then
- 32:32there's this connection to wild bill who
- 32:34seems to be have been very much about
- 32:35sort of like working within a system but
- 32:37like having autonomy at the edges and so
- 32:41that's kind of what I was playing with
- 32:42and just kind of taking notes and then
- 32:44yeah I would jump out and be like oh
- 32:46well actually I haven't figured out a
- 32:47conclusion yet let me start the
- 32:48conclusion section and I'll just sort of
- 32:50get that going on the side but Um, and
- 32:52then, you know, I I have a job, so I
- 32:55can't be doing this all the time. So,
- 32:56it's also like you interrupt yourself,
- 32:58and it's really nice to be able to come
- 32:59back and be like, you know, um, can you
- 33:02catch me up on the last three days of
- 33:06research?
- 33:07>> Oo, I love that question. That's so
- 33:09cool.
- 33:12>> And so, yeah, you can just kind of go in
- 33:13and be like, can you catch me up on the
- 33:14last few days of research? And, um, it's
- 33:17just going to go read all the stuff,
- 33:19right? Again, it's like I think the
- 33:21point you made earlier about the, you
- 33:23know, go find relevant sources. It's
- 33:24like I find a lot that the difference
- 33:27between the people are getting a lot of
- 33:30this right now is part of it is just
- 33:31like you have a good feel for where the
- 33:33edges of the capabilities of these
- 33:35models are and you sort of like
- 33:37encourage them to work within those
- 33:39capabilities. Like this is an incredibly
- 33:41easy, you know, it's like we know what
- 33:43it's going to do here, right? Like I
- 33:44could write all these Unix commands.
- 33:45It's just going to go find a bunch of
- 33:47files in this directory and it's going
- 33:49to look at them by date and it's going
- 33:51to look at all the files created in that
- 33:52project over the last you know and we
- 33:54know it's going to be able to do that
- 33:56and so it's saying um you know the major
- 33:59breakthrough day was this idea of
- 34:00bureaucracy as positional encoding which
- 34:03is very much work in progress idea but I
- 34:05kind of like it um but you know so it's
- 34:07just like it's pretty amazing also to
- 34:09just be able to kind of revisit deep
- 34:11work like this right where you know you
- 34:13know you're going to break your flow um
- 34:16you're not and it's like it's often you
- 34:18know I find whether it's code or writing
- 34:20the hardest part is like just picking it
- 34:22up again um because you're you're out of
- 34:24it and so just to kind of like kickstart
- 34:26that process is sort of amazing um I
- 34:29think what I was playing with here is
- 34:30this idea that um bureaucracy was
- 34:34actually like an innovation right um
- 34:36that like we look at bureaucracy as a
- 34:37negative and generally we talk about it
- 34:39as a negative and I think often it is a
- 34:41negative um but you know um bureaucray
- 34:45cy was a sort of like huge innovation
- 34:47for how companies operate, right? And
- 34:49ultimately it sort of represents kind of
- 34:51hierarchy and structure and a whole
- 34:54bunch of things that are like actually
- 34:56like pretty positive for operating at a
- 34:59large scale. Um and so um you know again
- 35:03my my kind of whole thesis on AI around
- 35:05all this bureaucracy stuff is that
- 35:07what's interesting about it is that um
- 35:09as opposed to kind of past technologies
- 35:11which kind of forced you to make a
- 35:13decision about whether you wanted to
- 35:14kind of like use your existing
- 35:18sort of structure and build that
- 35:21technology into your existing structure
- 35:22or adopt the new structure. most of the
- 35:24time the new software required you
- 35:26adopted the new structure and that's why
- 35:28so many sort of software projects failed
- 35:30for so long at least that's sort of part
- 35:31of what I think and um you know I think
- 35:34part of what's interesting about AI what
- 35:35I find so interesting is like that
- 35:39you can kind of keep letting everybody
- 35:41work in whatever way they want you know
- 35:42it's like a classic problem inside large
- 35:44companies is like one team wants to use
- 35:46a sauna and one wants to use Jur and one
- 35:47wants to use linear right and so then at
- 35:50some point there's like a there's a
- 35:51there's a huge project and they bring in
- 35:54some big consulting firm and they decide
- 35:55they're going to all centralize on this
- 35:57one thing and now twothirds of the
- 35:59company is unhappy um and like they've
- 36:02all made sacrifices and you know you're
- 36:04sort of in this like um very non ideal
- 36:07state and I think what's really
- 36:10interesting about AI and I I this is a
- 36:12little more sort of theoretical because
- 36:13I think you know we're so early in this
- 36:15is that like I think it's very possible
- 36:17you could just say well everybody just
- 36:19keep doing what we're you're doing we're
- 36:21going to stick sort of some models in
- 36:22the middle they don't care what you use
- 36:24because like it's all just data
- 36:26structures to them and so we can then
- 36:28have this sort of central thing and if
- 36:30you know I when you we talked about
- 36:32percolate at the beginning percolate was
- 36:34a content marketing platform worked with
- 36:36very large companies so it's a
- 36:37enterprise software product and it's
- 36:39like you know at the end of the day this
- 36:41is sort of the fundamental challenge of
- 36:42enterprise software is about like you
- 36:44know adoption and change management and
- 36:46I just think I think and I hope and
- 36:49again this is sort of the the optimist
- 36:51in me that like AI kind of lets us just
- 36:54not worry so much about these things and
- 36:56and rather than trying to make everybody
- 36:58change the ways that they work, kind of
- 37:00let them work in these ways and let AI
- 37:02sort of it's my I I call it my Thomas's
- 37:05English muffin theory of AI, which is
- 37:07that it like gets into the nooks and
- 37:09crannies. Um, and so uh uh yeah, that's
- 37:13so anyway, but I have no idea what
- 37:14bureaucracy is positional encoding means
- 37:16yet. I'm hoping I figure it out in the
- 37:17next two weeks before I have to give
- 37:19this talk. I think no, but I I think the
- 37:22point you just made is is totally right
- 37:23and it's actually not it doesn't have to
- 37:25be theoretical like I I've been seeing
- 37:27this too inside of every and I've been
- 37:29meaning to write about it and the the
- 37:30place that it's been coming up is we
- 37:33have so inside of every we run like six
- 37:37different products um and we have 15
- 37:39people so it's it's like a crazy product
- 37:42to headcount ratio um and what's
- 37:45interesting is I really like doing
- 37:47things in a in a bottomup way so
- 37:49everyone everyone each of the products
- 37:50has its own stack. We're not like
- 37:51centralized into a particular stack.
- 37:54Each you know GM that runs a product
- 37:56like just has made a decision about do I
- 37:58run rails or TypeScript or whatever. And
- 38:01what I'm seeing happen which is very
- 38:03cool is um a lot of the different
- 38:07products are running into similar things
- 38:09they want to solve for. So an easy
- 38:11example is um we have one product called
- 38:14Sparkle which is a little bit like a um
- 38:17uh a finder replacement or spotlight
- 38:19replacement. So it it it organizes your
- 38:21files and then it implements really fast
- 38:24spotlight search.
- 38:25>> I'm a user. I like it.
- 38:27>> Okay. So then you you know um and uh so
- 38:30that's really cool. And uh Agentic
- 38:33Search coming soon. Uh check it out. And
- 38:36uh and we're just building a new
- 38:38product, new GM, new stack uh called
- 38:42parah, which is essentially an in-house
- 38:46council. So it's short for parallegal,
- 38:47not par like you know, Thiago Forte
- 38:49Parah. Um and uh the whole job for PAR
- 38:53is just, you know, take all of your
- 38:55legal files and whenever I have a
- 38:57question and be like, okay, did do we
- 38:59ever sign this contract or what's the
- 39:00employee agreement template or whatever?
- 39:02it just gives you the answer and it's
- 39:03just clouded code sitting on top of a
- 39:04directory. Um,
- 39:07and and a thing that we needed to
- 39:09implement for that is uh this sort of
- 39:12like fast file file search. And what's
- 39:14really interesting is historically if we
- 39:17wanted to reuse the stuff that we
- 39:20learned from implementing sparkles file
- 39:23search that would have to be abstracted
- 39:25out into this modular library that
- 39:28anyone can use and then we have to be on
- 39:29the same platform and like that all
- 39:30those things right and what we did
- 39:33instead is we just added uh Piti who's
- 39:36the developer for for for par right now
- 39:38we just added her to the sparkle repo
- 39:40and I was just like just ask cloud code
- 39:42to figure out how it works and just do
- 39:44your own version. Um, and so you get
- 39:46this like sort of tacit code sharing
- 39:49where we all get better but without
- 39:51having to do the work of uh abstracting
- 39:54and modularizing everything because the
- 39:57the percentage of things that you can do
- 39:58that for are pretty low because it's a
- 40:00it's a heavy lift. And I'm seeing that
- 40:02happen all the time where just having a
- 40:04bunch of repos that are all solving
- 40:05similar problems but in different
- 40:07environments in different ways you
- 40:09everyone gets more productive because AI
- 40:11can kind of translate.
- 40:13I've uh one thing we've done there uh we
- 40:16also we so at le
- 40:21uh very large brands um and so we sort
- 40:24of build all kinds of AI things and um
- 40:27you know so we we've got lots of sort of
- 40:28internal and external repos and we
- 40:30frequently have the same thing and
- 40:32actually I've used the GitHub MCP um a
- 40:35few times for that same purpose which is
- 40:37just like um you know you're just in
- 40:40cursor or cloud code or whatever and
- 40:42you're like hey can you go like look up
- 40:44uh we run we've got an internal tool um
- 40:47called intelligence that just sort of is
- 40:49a wrapper around a whole bunch of like
- 40:51stuff that we use right so it's like got
- 40:54some CRM stuff and it's just like been a
- 40:57fun place to build the things that we
- 40:58need to run our company um but it's also
- 41:01a good place to kind of experiment and
- 41:03explore and figure out solutions to
- 41:04interesting problems and so I'll
- 41:06frequently be like oh can you go like
- 41:08just go look at the intelligence repo
- 41:11and um you look at how I implemented
- 41:13that thing there and take those sort of
- 41:16best practices and just pull them over.
- 41:19Um, and yeah, I think that stuff again
- 41:21that's where I really do believe in this
- 41:24idea. I like one of my whenever we have
- 41:30like a client meeting or something my
- 41:31first the question the icebreaker I
- 41:32always use is what was your aha moment
- 41:34with AI? And um uh mine was
- 41:39uh uh I mean it was probably not the
- 41:42very first, but it's the one that sort
- 41:43of I think was most impactful was I was
- 41:46uh I got access to build a chat GPD
- 41:49plugin. Remember when plugins came out
- 41:50like uh two and a half years ago or
- 41:52something now?
- 41:53>> You mean 50 years ago?
- 41:55>> 50 years ago. The the um uh and you know
- 42:00I I like you. I've I've written a lot of
- 42:03software in my life and I you know you
- 42:05know what you do when you like get
- 42:07access to something new. It's like I got
- 42:08to go read the API docs and figure it
- 42:11out and you know like there's a going to
- 42:14be a contract and you know as long as
- 42:15you follow that contract it works. And I
- 42:18go read the plug-in spec and it
- 42:20basically is like oh you just stick a
- 42:22manifest.json JSON file in the root
- 42:24directory of your application and in
- 42:25that you describe how you want us to
- 42:27send you data and you describe how
- 42:28you're going to send it back to us and
- 42:30then we'll deal with the rest. And I was
- 42:32just like that's amazing. It's also like
- 42:38it it's how the world should work. Like
- 42:40I wish everything worked that way. I
- 42:41wish I didn't always have to adhere to
- 42:43the big company's contract for how to
- 42:46send and receive data. Um, but also like
- 42:49I the thing that really struck me in
- 42:51that moment and it's like been my kind
- 42:52of like rallying cry around all this
- 42:54stuff is that um it's also just like
- 42:55fundamentally counterintuitive in that
- 42:57like I literally have a career's worth
- 42:59of intuition for how to integrate
- 43:02software systems and it flipped it on
- 43:04its head like like like quite literally
- 43:08180 degrees away from my intuition of
- 43:10how software systems should be
- 43:12integrated was this thing. And that I
- 43:16think since then has been my kind of
- 43:18thing for everybody has been like this
- 43:21is just not intuitive for now. And and
- 43:24that's not a bad thing. It just means
- 43:25like you need to build intuition and
- 43:28like that's what we're all just out
- 43:30there trying to do with it, right? And
- 43:32so, you know, when I don't know, I mean,
- 43:33part of what I like about what you're
- 43:35doing and, you know, even just hearing
- 43:37the things you're saying, but like
- 43:38generally what you do with the podcast
- 43:40and what you do with every is like so
- 43:42much of it is like we're all kind of
- 43:44just figuring stuff out for the first
- 43:45time, right? And like um you know, we're
- 43:47like, "Oh, will this work?" And then
- 43:49like all of a sudden you have this new
- 43:51bit of intuition for what these things
- 43:53can do and what a computer
- 43:55that is not deterministic looks like.
- 43:58And um that's I think that's just what
- 44:01we're all doing all the time. And that's
- 44:03why it's so fun. I think
- 44:04>> that's why I love this moment cuz like
- 44:06you just have a weird idea and you're
- 44:07like
- 44:09has anyone done this before? And and
- 44:11it's like no. And it's not a it's not a
- 44:13complicated idea. It's just it's just a
- 44:15new whole new territory, you know.
- 44:18>> I Yes. I I think about that all the
- 44:21time. And I think actually like I think
- 44:23one of the really damaging sort of
- 44:24things out there is that um I think
- 44:28there are a lot of people who think
- 44:29we're way further along in this than we
- 44:32are. And so you know I think
- 44:34particularly the people who are sort of
- 44:36scared you know we run we work with like
- 44:38Fortune 50 companies. And so you know
- 44:42when we're sort of like out there and
- 44:43we're talking to people inside the
- 44:44organization a lot of people feel like
- 44:46they've already been left behind. And
- 44:48it's like, no, you can like literally go
- 44:50sign into Chad GBT and like
- 44:54do something like nobody's thought about
- 44:55doing with this thing yet because
- 44:57there's just so much white space to
- 45:00explore and and you might discover some
- 45:03totally new way of using it and or like
- 45:06totally new trick and um I don't know
- 45:08that's just that's and you know I think
- 45:10to be fair to some people that's sort of
- 45:12very intimidating and I don't think by
- 45:14and large the models do any favors to
- 45:16themselves in helping those people get
- 45:18their feet wet in that like you know I
- 45:20think people go on there and they you
- 45:22know it's like you ask it to write you a
- 45:23poem and then it writes you a poem and
- 45:25you're like okay it wrote me a poem um
- 45:27but uh I don't know that feeling of like
- 45:32that feeling of yeah being it's like
- 45:33being on the frontier right
- 45:35>> totally and yeah I I think that um your
- 45:38point about intuitions and getting
- 45:40intuitions is the big thing and I think
- 45:42people what we don't realize is when
- 45:44you're dealing with something
- 45:44fundamentally new you you can't trust
- 45:49like how you reason about it without
- 45:51experiencing it because you have to
- 45:52build the intuition in order to be able
- 45:53to reason about what it means and and
- 45:55how it fits in and whether it works or
- 45:56not. And we're just not used to that cuz
- 45:59we're used to reasoning about things we
- 46:01already have an intuition for. And um
- 46:04and I think that's why like when you
- 46:06first see maybe when you first saw JBT
- 46:08be like oh my god it can do everything.
- 46:09Like we're not going to have jobs in a
- 46:11year. And now we're like three years in
- 46:12and we're like, "Yeah, it's awesome."
- 46:14And
- 46:16jobs are complicated. There's a lot of
- 46:18complex stuff that we do, you know? Um,
- 46:21and I I love I love kind of that, you
- 46:23know, in order to build the intuition,
- 46:26all you have to do is is use it. And
- 46:28that just by using it, you're already
- 46:30kind of on the edge uh for now. And
- 46:33yeah, I think that's the best.
- 46:34>> There's apparently there's a a German
- 46:36word called fingerpitsengu.
- 46:38of course there is
- 46:39>> building fingertip feeling and that's
- 46:41been my um uh just because I can't
- 46:44resist also um I I I'm in that whole in
- 46:50in sort of the realm that you're
- 46:51discussing like uh I've been trying to
- 46:54do a lot of analogy analogizing right
- 46:56it's sort of like and I think you know
- 46:58that's really hard and um uh you know
- 47:01but my two that have sort of stuck the
- 47:03most one is just um I watch a lot of
- 47:05YouTube with my kids and um Uh, we watch
- 47:08this channel called Veritassium. It's a
- 47:10science channel and
- 47:11>> I love Veritassium.
- 47:12>> Yeah, it's great. Um, and he did one
- 47:14where he built a bike that locks out
- 47:17left if you try to turn right and locks
- 47:19out right if you try to turn left. And
- 47:20what he's proving is that you can't
- 47:22actually turn a bike left unless you can
- 47:25turn it right. Um, which none of us
- 47:27would think about when we ride a bike
- 47:29because it's all just second nature and
- 47:31intuition. But it's also why you can't
- 47:34explain to a child how to ride a bike.
- 47:37um they just have to get on it and feel
- 47:38it and um and so you know I really that
- 47:41video is amazing and um uh it's uh that
- 47:44channel is amazing and and but I I've
- 47:46thought a lot about that and then the
- 47:47other one which is a sort of deeper cut
- 47:50is um uh there's an amazing book about
- 47:53quantum physics called beyond weird um
- 47:56by Philip Ball and um the thesis of the
- 47:59book is basically that like there's
- 48:01nothing particularly strange about
- 48:03quantum physics that we like we have a
- 48:05very good understanding of it like we
- 48:07wouldn't be talking right now. We
- 48:09wouldn't be on computers. We wouldn't
- 48:10have phones if we didn't have like a
- 48:12very good grasp of the mechanics that
- 48:14exist underneath it. And his thesis in
- 48:16the book essentially is that like what's
- 48:18really lacking is the vocabulary because
- 48:20like we all exist in a Newtonian world,
- 48:22not in a quantum one. And so we all have
- 48:25words that reflect the sort of
- 48:28deterministic processes of that macro
- 48:32universe. And I think a lot about that.
- 48:34I have not like fully been able to sort
- 48:36of pull that string all the way to AI.
- 48:38But I feel like there's a real
- 48:40connection there because I think that
- 48:41there's just something really weird
- 48:43about using probabilistic computers.
- 48:44Like we're not used to using things that
- 48:47like you ask them the same question
- 48:48twice and they have different answers.
- 48:50Like that's very strange. We're not used
- 48:52to I'm not used to writing code where
- 48:54you can tell the larger company how you
- 48:57want them to send you data and they can
- 48:59just do it. Like these are not
- 49:02normal things that any of us have lived
- 49:05with in our lifetimes and so of course
- 49:07it takes some time for us to adjust. I
- 49:09think so too and I I actually have a
- 49:12hope that language models by becoming a
- 49:16standard way that we use computers will
- 49:19create that vocabulary. Um because we
- 49:23actually are quite good at dealing with
- 49:26probabilistic nondeterministic things
- 49:28like other humans. Um we've just grown
- 49:31up in a world where
- 49:34because of you know the enlightenment
- 49:35and the scientific revolution and and
- 49:37the tools that came out of that are very
- 49:40much like deterministic we've associated
- 49:43that with uh how we see like that's how
- 49:46we see the world because of those tools
- 49:47and that language and uh there's a whole
- 49:50other part of the way that we see the
- 49:51world which is much more squishy and
- 49:52much more like vibes based that has been
- 49:56um I think de deprioritized uh
- 49:59especially in western culture that now
- 50:01that we have a tool that works that way,
- 50:03I think we'll be able to start seeing
- 50:05that again. And that's one of the
- 50:06beautiful things to me about language
- 50:07models is it opens up that whole world
- 50:09again.
- 50:10>> Yes, I love it. Um
- 50:13>> I do want to go back to cloud code uh
- 50:15and
- 50:17we should do the phone unless there's
- 50:18unless there are other things that you
- 50:20uh that that you want to share on the
- 50:21computer. Um, but the the the thing I
- 50:25want to do before we get there that's
- 50:26just on my mind right now is like you
- 50:28said you said, you know, you watch this
- 50:30with your kids and I'm sort of curious
- 50:32how like what do your kids think about
- 50:34this and how are you dealing with it
- 50:36with your kids? Um,
- 50:38>> yeah. Um, I love that question. Uh, so
- 50:41uh I've got a seven and a 10-year-old.
- 50:43Um, and obviously like I'm pretty kind
- 50:45of deeply embedded in this stuff and so
- 50:47I've sort of exposed them quite a bit to
- 50:49it. um uh you know they don't uh so they
- 50:54will like occasionally use the sort of
- 50:56voice models and they have a pretty good
- 50:58understanding and we'll be in the car
- 50:59and just play games and ask questions
- 51:00with Grock and do those kinds of things.
- 51:02Um this weekend actually for the first
- 51:04time my 10-year-old um uh she was really
- 51:08eager to be every year um uh my wife and
- 51:15her sister and brother and mom and and
- 51:18all the cousins and we all get together
- 51:19and we do Christmas together and so it's
- 51:21too many presents to give to everybody.
- 51:23So we do a kind of like not secret
- 51:25secret Santa where everybody chooses one
- 51:26person and my 10-year-old really wanted
- 51:28to be the one who got to be the chooser.
- 51:31Um, and uh, I encouraged her to uh, vibe
- 51:35code an app to do it. Um, and so I just
- 51:38gave her my phone and Vzero. And um,
- 51:42honestly that was like so amazing to
- 51:45watch. Like not just cuz it was so cool
- 51:47to see her do that and build it. And she
- 51:50went through, she was having so much
- 51:51fun. She did 75 revs on Vzero. Um, uh,
- 51:56so she like really got it going.
- 51:58>> Polished Santa app. Uh, She also like
- 52:03started to get into like really
- 52:05interesting kind of like computer
- 52:08science ideas without knowing it. So in
- 52:10one of the things like the adults give
- 52:12presents to adults and the kids give
- 52:14presents to kids, but she wanted this to
- 52:16be a more generalized app. So she
- 52:18realized that like rather than having
- 52:20adults and kids, you need to call them
- 52:21groups, right? and like you know so
- 52:24she's like getting into data modeling
- 52:26and like all of this I'm like watching
- 52:28this conversation happen and um I just
- 52:32thought that was so awesome. Um, and you
- 52:35know, also just like a real pet peeve of
- 52:37mine right now and I've I've sort of
- 52:39gotten this argument with a bunch of
- 52:40people is like there seems to be a big
- 52:42conversation that like there's a a
- 52:44bubble in vibe coding and like because
- 52:46one company or another might have too
- 52:48high of a valuation. And my take on that
- 52:51is like I just I I could not care less
- 52:53what the valuations of these companies
- 52:55are. I think like fundamentally if
- 52:57there's a tool that could allow a
- 52:5810-year-old to like build an app there
- 53:01can't that can't be a bubble. like I I
- 53:04just like can't see a possibility where
- 53:06that is that. So, um anyway, that's sort
- 53:09of one side of it. The other big one for
- 53:10me that I've been thinking a lot about
- 53:12is um
- 53:14uh is sort of media literacy and
- 53:17education um stuff. So, um you know,
- 53:20both at the sort of schools they go to
- 53:23um and then also I went to NYU and I've
- 53:26sort of um been having more and more
- 53:28conversations with the dean of the
- 53:30school I went to there. Um and uh um
- 53:34there's a lot of fear inside schools
- 53:36right now about AI and about cheating
- 53:38and um there's a big thing you know so
- 53:41some parents in my town they you know
- 53:43they wanted to have more of a
- 53:45conversation about it and um you know as
- 53:48someone who I I've thought about this a
- 53:50lot but I've also just like been I've
- 53:52spent my entire life thinking about sort
- 53:54of technology and its effects on culture
- 53:56and I I think I'm like relatively
- 53:58grounded in these things. of like I've
- 54:00put in good hours of thinking. I I know
- 54:02that for sure. Um and so, you know, my
- 54:04my take on it is like one that sort of
- 54:06you can't hide technology that won't be
- 54:09hidden, right? So, it's like, you know,
- 54:11putting our head in the sand is not the
- 54:13best solution. And then, you know, my
- 54:14bigger one though is like I I was out um
- 54:17a friend of mine asked me to come um
- 54:19talk to a school two years ago about AI
- 54:22um out in LA and um afterwards I was
- 54:25talking to a English teacher there and
- 54:27she was like, "What do I do? Like what
- 54:29do I do about all these kids, you know,
- 54:31using AI and I was like, look, I don't
- 54:33really know what your job is because
- 54:34like I mean being an English teacher for
- 54:3711th graders sounds really much harder
- 54:38than my job." Um, but on a really
- 54:42fundamental level, like I don't actually
- 54:44think your job is to teach these kids to
- 54:46write because that's like a lifelong
- 54:48pursuit. I think your job is to convince
- 54:50them that it's worth learning to write.
- 54:52And so in that way, like I I'm not sure
- 54:53that anything fundamentally changes
- 54:55because of AI. Like I I think that you
- 54:58know and again this is my very
- 54:59optimistic take but like I am I think
- 55:03that there are so many parts of the
- 55:06education system that AI really just
- 55:08exposes the sort of flaws in the way
- 55:10that we teach. like why are there so
- 55:12many tests on these kinds of things
- 55:14instead of encouraging thinking and
- 55:16learning and coming to love to write and
- 55:19research and whatever you know it's like
- 55:20we're so focused on teaching kids the
- 55:22you know five paragraph essay you know
- 55:25while every adult who is a writer has
- 55:28long abandoned that and because like
- 55:31it's all about sort of discovering that
- 55:33you like to write and you you know what
- 55:35your own style is and how to do it and
- 55:37you know it's like I'm sure it's like
- 55:39you know a big part of working with AI
- 55:41to write is like telling it is ignoring
- 55:44it because you're like no that's not me.
- 55:46I'm not into that. Like I like I I I'm
- 55:49totally comfortable saying really here.
- 55:51I know you don't think it's a good idea
- 55:52but like I'm I'm cool with it. Um and so
- 55:55you know I don't know. I I've been sort
- 55:56of having a lot of these different
- 55:58thoughts. I'm actually pitching a class
- 56:00for the fall of 26 at NYU to kind of um
- 56:05the idea for it is uh code is essay and
- 56:08my my sort of point of view is like this
- 56:10sort of opens up this new way to express
- 56:12yourself. um and that you know like we
- 56:17have all these other ways that we
- 56:19celebrate to express ourselves but code
- 56:21has been like long shut off from people
- 56:23because it's but actually it's kind of
- 56:25amazing and it lets you express yourself
- 56:26in all these different kinds of ways and
- 56:28like you know this is what my
- 56:2910-year-old was doing this weekend. So
- 56:30anyway, those are all the bits and
- 56:32pieces as a parent I've been thinking
- 56:33about. The one other thing I will plug
- 56:35is um media literacy I think is a big
- 56:38piece of it. A lot of people are afraid
- 56:39of these models and hallucinations and
- 56:42um there's a book um by a guy named Tim
- 56:45Harford who writes for uh the FT and
- 56:48he's an economist um and he has a book
- 56:50called The Truth Detective which is an
- 56:52adaptation of the data detective but
- 56:54it's written for kids and it's the best
- 56:56media literacy book I've ever read for
- 56:59adults or for children. Um, and I think,
- 57:02you know, a lot of what this AI
- 57:04conversation exposes is like how bad a
- 57:07job we do with helping and arming our
- 57:10kids and our adults to be sort of like
- 57:15truly media and technology literate.
- 57:17And, you know, like being really good at
- 57:20knowing what's real on social media
- 57:22turns out to be also really useful for
- 57:24like differentiating between
- 57:26hallucinations and non-h hallucinations
- 57:28in chat GBD, right? like this is a sort
- 57:30of to me a kind of very central skill
- 57:33that like we need to arm everybody with
- 57:35and um I am sort of way more interested
- 57:38in that with my kids than I am in
- 57:41worrying about them cheap. That was a
- 57:42very long answer to your question. I
- 57:43don't know if it got at what you were
- 57:45looking for.
- 57:45>> No, it's amazing. I that's exactly what
- 57:48I'm looking for and it sort of what it
- 57:50strikes me another way to frame what
- 57:53you're saying is for example rather than
- 57:58learn having them memorize and be
- 58:00quizzed on the 50 what the 50 states are
- 58:03uh
- 58:05asking them to go find the 50 states
- 58:07with CHBT and be able to tell when the
- 58:10AI is giving them the wrong answer
- 58:11because that's a lot more of the the
- 58:14meta skill that they're going to need
- 58:15anyway down the down the line. And
- 58:17again, I'm not a teacher. I'm sure
- 58:19there's lots of teachers who were like,
- 58:21"This is that's crazy for a lot of
- 58:23probably good reasons." But there's
- 58:24something interesting there where it
- 58:26becomes the meta skills become more
- 58:29important than they used to be. And in
- 58:31order to do well at the meta school, you
- 58:33have to also be able to like to some
- 58:35degree do the underlying skill, too. But
- 58:37um we probably could be spending a much
- 58:39more time in the medical school than we
- 58:40are now. And the education system like
- 58:41isn't really set up to do that anyway.
- 58:43>> Yeah. And I I you know even I think even
- 58:45simpler examples is like I I went to a
- 58:47school called Gallatton at NYU and and
- 58:50um when you graduate you have to sort of
- 58:53it's not quite a thesis defense but you
- 58:55have to sort of spend three hours with
- 58:56four professors and and or three
- 58:59professors and sort of explain kind of
- 59:02your line of reasoning around why you
- 59:04studied what you studied. and you need
- 59:06to be prepared to kind of like weave
- 59:08into that defense um uh any of 25 books
- 59:12that you put on your book list. Um and
- 59:15you know I was talking to them and it's
- 59:16like amazing that's like entirely AI
- 59:18proof, right? Like there's no cheating
- 59:20on that. Like you show up in that room
- 59:22and you're either prepared to speak to
- 59:24it or you're not. And whether you're
- 59:25prepared with AI or not is uh it doesn't
- 59:28mean anything, right? like it's it's
- 59:29like like do you can you can you make a
- 59:33an argument in this room and like not
- 59:36everything is going to be that easy to
- 59:38be sort of like cut off but like I don't
- 59:40know there's something really beautiful
- 59:42about that idea right because it's like
- 59:44it's naturally cheating proof because
- 59:47like you're sitting there and and it's a
- 59:49question of like did you actually
- 59:51internalize these things um and I don't
- 59:54know that's way more interesting to me
- 59:56than even like was your essay or any of
- 59:59these other it's like did you did you
- 1:00:01get it like um and uh so yeah I don't
- 1:00:06know um I'm I'm I I'm trying to do my
- 1:00:10best to kind of like um take a balanced
- 1:00:12approach and um try to at least sort of
- 1:00:15like tamp down some of I I live in a
- 1:00:18small town in Connecticut and you know I
- 1:00:20think there's a lot of fear amongst
- 1:00:22parents you know it's like it was mobile
- 1:00:24phones and then it was social media and
- 1:00:26now it's AI and it's like another thing
- 1:00:28that's going going to ruin our kids. And
- 1:00:29um I don't think that that is true. Um
- 1:00:33uh but I think there are things we can
- 1:00:35and should do to encourage it to not be
- 1:00:38true. Like really get them really good
- 1:00:40at the things you're saying. Like how do
- 1:00:42you tell? And again, it's like a
- 1:00:43hallucination is just a form of the same
- 1:00:46kind of misinformation that exists in on
- 1:00:49television and on the internet and in
- 1:00:51social media and everywhere else. And um
- 1:00:54you know just sort of like encouraging
- 1:00:55people to kind of get in touch. There's
- 1:00:56a great part of the truth detective um
- 1:00:59for the kids book he he calls it the
- 1:01:01brain guard and like one of the bits of
- 1:01:03advice he has is when you encounter some
- 1:01:05piece of information if it makes you
- 1:01:07feel really good cuz you agree with it
- 1:01:09then you should be even more skeptical
- 1:01:11of it and he calls it the brain guard
- 1:01:12you know and he's explaining this like
- 1:01:14this is for a 9-year-old and I just
- 1:01:15thought that was like such a beautiful
- 1:01:16way to put it right like that's like
- 1:01:18what that's what you learn to do when
- 1:01:20you get good at being on the internet um
- 1:01:24is that you're like wait I should like
- 1:01:25be more skeptical of this because this
- 1:01:27is like in line with everything I think.
- 1:01:29Let me just doublech check the way like
- 1:01:32I I you know, you get to learn that
- 1:01:34feeling in your gut and get to learn
- 1:01:36when to react to it. And so yeah, that's
- 1:01:39a lot of how I think about it.
- 1:01:41>> That's great. I'm I'm interested. I'm
- 1:01:42going to get I'm going to get that book.
- 1:01:44I'm I'm interested in reading it.
- 1:01:46>> You should. My plan is to read it to my
- 1:01:47kids every year um from now on.
- 1:01:50>> Nice. I love it.
- 1:01:50>> Just refresh it.
- 1:01:52>> Um all right. Now for the moment we've
- 1:01:54all been waiting for. show us how you
- 1:01:56use cloud code on your phone as a second
- 1:01:58brain notetaker.
- 1:02:00>> Okay, so here we go. So I am going into
- 1:02:04an app called Termius.
- 1:02:08>> Termas is just a terminal.
- 1:02:11Um, and uh, what is allowing all this to
- 1:02:14happen behind the scenes is in my
- 1:02:16basement I have a mini PC and on that
- 1:02:20mini PC I have a thing called tail scale
- 1:02:24running and tail scale lets you set up
- 1:02:26these very simple VPNs. So I'm currently
- 1:02:30inside like if I scroll down here you
- 1:02:33see I'm on a VPN. That's my personal
- 1:02:35VPN. I'm not like on an outside VPN. I
- 1:02:38see. Um, so the only way to access this
- 1:02:40machine is through my VPN.
- 1:02:42>> Okay.
- 1:02:43>> Um, and so then when I go in there, um,
- 1:02:46because I sync my Obsidian,
- 1:02:50um,
- 1:02:52with, uh, Git, so I put it, it's on
- 1:02:55GitHub, uh, in a private GitHub. Um, I
- 1:02:58can then sync it back down to here. And
- 1:03:00so then I can just call up Claude.
- 1:03:03And now I'm just in claude code talking
- 1:03:07and thinking and I can just be like um
- 1:03:11what's new in the last two days.
- 1:03:16Um I can access any of my agents. Um I
- 1:03:20can do anything. And again this is in my
- 1:03:22Obsidian but I can use this anywhere,
- 1:03:25right? So I'll be like on the fly. I've
- 1:03:28got other repos in here. Um, you know, I
- 1:03:30I realized like a link was broken on my
- 1:03:34conference site and so I uh um I just uh
- 1:03:39opened the repo. I pulled it down. I
- 1:03:43asked QuadCode to make the changes and I
- 1:03:45was able to do it right here. Um, so
- 1:03:47this has been like completely wild to me
- 1:03:50because again, this is very much in that
- 1:03:52like like on Tuesday of this week, um,
- 1:03:57we had Monday off, Tuesday, I dropped
- 1:03:58the I dropped the kids off at the bus
- 1:04:00and then I, uh, I went and I sat and had
- 1:04:02breakfast and I literally sat on my
- 1:04:04phone and worked on this talk for like
- 1:04:07two hours. Um, and I did it through
- 1:04:10here, right? like on my phone where I
- 1:04:12was like doing real thinking and
- 1:04:14research and pulling things in and
- 1:04:15pasting things in and doing all this
- 1:04:17kind of stuff and you know I'm able to
- 1:04:20do it all and it it just it doesn't seem
- 1:04:24like I could do that kind of thing
- 1:04:26without that. So yeah, this has been
- 1:04:27like a completely revolutionary change
- 1:04:29in my life. Um, and uh, actually one of
- 1:04:33the things I've been doing lately is
- 1:04:34like setting up I've got all these
- 1:04:36friends now who have set up um, like
- 1:04:38little partitions of this mini PC in my
- 1:04:40basement um, so that they can also run
- 1:04:42cloud code on their phone because um, I
- 1:04:45like it so much. Do
- 1:04:47>> does this make you be like, "Oh my god,
- 1:04:49I got to drop everything and just build
- 1:04:51a actual purpose-built notes app that
- 1:04:53has cloud code as a backend for this?"
- 1:04:57>> No. Well, I mean, actually, one of the
- 1:04:58things I've been thinking a lot is like
- 1:05:00maybe I just like everything should just
- 1:05:02run in Linux all the time for me. Maybe
- 1:05:04this is like at least for the short
- 1:05:05term, this is the answer to all of what
- 1:05:07I just need to like not have anything
- 1:05:08anywhere else. Um, no, I mean, I'm I'm I
- 1:05:12I will say I'm I'm sort of pretty out of
- 1:05:14the SAS game these days, so I don't
- 1:05:16often kind of think that I should drop
- 1:05:18everything and and do anything. I find
- 1:05:20this to be a really amazing solution. Um
- 1:05:22and uh um but no, I mean I this but this
- 1:05:27has really this has really like changed
- 1:05:30the way I work and I feel like I can
- 1:05:32just be anywhere and just be on my phone
- 1:05:33and um you know I mean I was out like uh
- 1:05:37I needed a break. Uh it was you know
- 1:05:394:30 I went and sat outside for a while
- 1:05:42and then um we had a project that needed
- 1:05:44to get delivered to a client and a small
- 1:05:46change needed to be made that like I was
- 1:05:48the best suited to make that change. Um,
- 1:05:51and so I just like hopped on my phone. I
- 1:05:53pulled the repo down and I um went into
- 1:05:55Cloud Code. It's like a tiny little
- 1:05:57change, you know, like the way I find
- 1:05:59myself using Cloud Code the most for
- 1:06:01code is that like mostly I'm having it
- 1:06:03do the work I already know how to do.
- 1:06:06You know, I'm like, oh, I like I know
- 1:06:07exa I knew exactly what was going on in
- 1:06:09that situation. Um, like I knew why we
- 1:06:13were having the issue we were having.
- 1:06:14And so it was like I could have gone
- 1:06:16back to my computer and opened up uh
- 1:06:18cursor and done it in cursor either by
- 1:06:21hand or with cursor, but it was like I
- 1:06:23just I was like I told Cloud Code
- 1:06:24exactly where to look. It was like and I
- 1:06:26first confirmed that the problem was
- 1:06:28what I thought the problem was and then
- 1:06:29I just had it push a solution and it
- 1:06:31pushed a PR and then I was done.
- 1:06:33>> Amazing.
- 1:06:33>> And I was still sitting outside by the
- 1:06:35pond.
- 1:06:39>> I love that. Um yeah, I've I've
- 1:06:41definitely had that experience. I've
- 1:06:43never I've not done it on my phone. I'm
- 1:06:44like I have my laptop out with me, you
- 1:06:46know, by the pond or by a lake or
- 1:06:48whatever. But you're inspiring me. I
- 1:06:50have a Mac Mini in the office that I've
- 1:06:52been meaning to set up. So, I think this
- 1:06:53is going to give me
- 1:06:53>> one of the other, by the way, um one of
- 1:06:56my other I'll just take you out of here
- 1:06:58for one sec actually just to show you
- 1:06:59this. One of my other big ahas um
- 1:07:03recently has been um
- 1:07:06uh building um clawed code helpers for
- 1:07:12um doing uh
- 1:07:15uh basically like uh setup work. So it's
- 1:07:19like I'm not I'm like been playing with
- 1:07:21Linux lately. I'm playing with this
- 1:07:22Omari which is DHH's Linux distribution
- 1:07:26and I'm not like super comfortable in
- 1:07:28here. And so I got this whole this is a
- 1:07:31claude code project specifically to help
- 1:07:33me configure this box and it's like so
- 1:07:36nice because I'm like oh how do you do
- 1:07:38this in Linux again or like what's the
- 1:07:40neoim command or like can you change
- 1:07:42this can you help me install this plugin
- 1:07:44for neo or whatever it is or and so now
- 1:07:46I have one on my Mac too where it's like
- 1:07:48um can you clean up all the homebrew
- 1:07:50things or like I switched from like
- 1:07:52which Python package manager I was using
- 1:07:54and it was like that would have been
- 1:07:55super overwhelming for me and I was like
- 1:07:57I want to switch from
- 1:07:58pip to using UV. Can you like just make
- 1:08:01that happen? And it like just did all
- 1:08:02this stuff for me. And it knows all my
- 1:08:04preferred settings. And so, um, I
- 1:08:06actually have a version of this where
- 1:08:07like now I've got it so tuned that, um,
- 1:08:10if I want to launch a new box for doing
- 1:08:13something, it'll just have all my
- 1:08:14settings ready to go and then I can log
- 1:08:16into Claude Code and the Claude code can
- 1:08:18then set up anything else that didn't
- 1:08:20get set up in the initial process.
- 1:08:22>> That is amazing. That's wild.
- 1:08:25Um,
- 1:08:26>> this is my happy place.
- 1:08:28>> I I can see that. Are do you have any
- 1:08:30like um uh are there are there any big
- 1:08:33projects like this that you've been
- 1:08:34itching to do or itching to try out?
- 1:08:36>> Uh, no, not really. I mean, I've been
- 1:08:39having a ton of fun that my my server
- 1:08:40stuff has been a ton of fun. I've been
- 1:08:42doing a lot of that. Um I'm uh I'm I
- 1:08:46like I said I mean this is kind of a
- 1:08:47joke and kind of not is like I'm super
- 1:08:49interested in um like doing more I
- 1:08:53because Claude code has become such a
- 1:08:55sort of integral part of my life like
- 1:08:57I'm very interested in the command line
- 1:08:58and I found myself installing more and
- 1:09:00more things into the command line um and
- 1:09:02like doing more and more work. So, I've
- 1:09:04been using like uh um Simon Wilson has a
- 1:09:07LLM command line tool and like been
- 1:09:10doing more and more stuff sort of in
- 1:09:12that and then like also layering that
- 1:09:14back into cloud code. So, it's like I
- 1:09:16did my newest Claude code little
- 1:09:19Obsidian tool is it I have an
- 1:09:22attachments folder in Obsidian where all
- 1:09:24the PDFs and images and stuff in any
- 1:09:26note go, but inevitably they have
- 1:09:29terrible names, right? I mean, um, and
- 1:09:31so this goes through very similar to,
- 1:09:34um, uh, Sparkle and but just in that
- 1:09:37Obsidian folder and it it renames them
- 1:09:39all and then it also puts them in like a
- 1:09:42meta data. It puts them in a table in
- 1:09:44the attachments folder and then it
- 1:09:46renames all the attachment links back.
- 1:09:48So it just like cleans everything up. It
- 1:09:50just does it through Gemini Flash. Um,
- 1:09:53and so it's like I don't know, stuff
- 1:09:54like that's kind of amazing. Um, so I
- 1:09:58don't I'm just like I'm having the time
- 1:09:59of my life just building and tinkering
- 1:10:02and you know I mean this is just on the
- 1:10:04side and I get to do the same thing. I
- 1:10:06mean we work with like Amazon and Meta
- 1:10:08and PayPal and all these big companies
- 1:10:10and you know we're just like building
- 1:10:12amazing stuff all the time.
- 1:10:13>> I love that. I love the energy. If
- 1:10:15people are interested in uh following
- 1:10:19you or working with you uh where should
- 1:10:21they find you?
- 1:10:23>> Yeah. Um, so Alfic is alle.com. A l
- 1:10:27pic.com.
- 1:10:30Um, and then I also run this thing
- 1:10:31called Brand B RX N D.AI to make it
- 1:10:34particularly confusing. Um, that's a
- 1:10:36conference. We've got the conference
- 1:10:37coming up on September 18th in New York
- 1:10:39City. You should come if you're around.
- 1:10:41It'll be really fun. We're going to talk
- 1:10:42about marketing and AI. Um, and I also
- 1:10:44write a newsletter there at
- 1:10:46newsletter.brnd.ai
- 1:10:48about sort of specifically at this
- 1:10:50intersection of AI and marketing. Um,
- 1:10:52those are kind of the best places to
- 1:10:54find me these days.
- 1:10:55>> Awesome, Noah. Always a pleasure.
- 1:10:58>> Pleasure is all mine. Thank you, Dan.
- 1:11:03[Music]
- 1:11:08Oh my gosh, folks, you absolutely,
- 1:11:10positively have to smash that like
- 1:11:12button and subscribe to AI and I. Why?
- 1:11:15Because this show is the epitome of
- 1:11:17awesomeness. It's like finding a
- 1:11:18treasure chest in your backyard, but
- 1:11:20instead of gold, it's filled with pure
- 1:11:22unadulterated knowledge bombs about chat
- 1:11:25GPT. Every episode is a roller coaster
- 1:11:28of emotions, insights, and laughter that
- 1:11:30will leave you on the edge of your seat,
- 1:11:32craving for more. It's not just a show,
- 1:11:35it's a journey into the future with Dan
- 1:11:37Shipper as the captain of the spaceship.
- 1:11:40So, do yourself a favor, hit like, smash
- 1:11:43subscribe, and strap in for the ride of
- 1:11:45your life. And now without any further
- 1:11:47ado, let me just say Dan, I'm absolutely
- 1:11:50hopelessly in love with you.
About this transcript
This page contains the full transcript of Claude Code Can Be Your Second Brain by Every, generated from the public captions YouTube serves with the video. The transcript has 14,569 words across 1,950 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.