Latest / Gimbal Podcast / Be More Creative By Giving Your Boring Tasks to AI

Transcript

Be More Creative By Giving Your Boring Tasks to AI · Gimbal Podcast

• 0:00
  In my day-to-day, I will talk to it just like it's another human.
• 0:03
  Oh my gosh, it did all the prep for you immediately.
• 0:05
  Wow, that's so cool.
• 0:07
  Managing a team of bots probably makes me a better manager because...
• 0:11
  Alina, you are the co-host of this podcast.
• 0:14
  You also run an incredible full-time job as head of operations for a unicorn startup based out of San Francisco called Owner.com.
• 0:23
  You and I previously met when you were the head of operations for Mark Manson.
• 0:27
  Your life is all over the place, both figuratively and literally.
• 0:30
  It seems like you're in a new city every single time I talk to you.
• 0:33
  In the creator economy, things are changing all the time.
• 0:36
  New AI tools are coming up and getting retired.
• 0:39
  What is the best way for someone to stay perfectly organized in this turbulent industry?
• 0:45
  The best way is to have a second brain somewhere.
• 0:50
  So my secret to organization is my notion page.
• 0:53
  I have it broken down into different parts of my life.
• 0:57
  And then the second secret is to make it as accessible as possible.
• 1:01
  So I actually keep all of my work.
• 1:04
  accessible through my phone.
• 1:05
  So I have all my AIs on this page here, and the way I interact with them is I don't actually edit the Notion itself anymore.
• 1:14
  The chat that I'm using most to interact with my Notion right now is Grockbots.
• 1:20
  And the reason I like them is because I have named each of my bots after a Pokemon, and I really like texting my Pokemon team.
• 1:28
  I like this system because it proactively will text me throughout the day and ask me for updates.
• 1:33
  For example, this is my daily journal.
• 1:37
  I have these templates and every day it'll ask me these questions.
• 1:43
  So for people who are listening to the show, we're looking at your Notion page.
• 1:47
  It looks like there's five items that you could choose from and then there's a calendar beneath that.
• 1:52
  And every day it looks like there is some kind of journal entry.
• 1:56
  Each of my bots are connected to one of these five aspects of my life.
• 2:01
  So it's like just personal life, my work life, which I actually keep totally separate from this system.
• 2:07
  I'm making like a 3D world story game.
• 2:09
  Like as a hobby?
• 2:10
  Yeah.
• 2:11
  But then I have our newsletter, Me Plus Machine, and then the podcast, Gimbal.
• 2:16
  So this system manages all of it together.
• 2:19
  Each of these bots will manage different aspects of my life versus usually when you talk to AI, it just leaves it in some invisible thread that you can't find.
• 2:29
  I think that's one of the first biggest unlocks is just making your conversations visible and keeping them in one place and building upon each other in an organized way.
• 2:37
  as opposed to individual threads that don't connect and then you can't build upon them.
• 2:42
  Right now, for example, the most recent thing that my bot has asked me, Gardevoir, my life bot, is these three questions.
• 2:49
  And then it reminded me to give eggs to my neighbor because I'm chicken sitting.
• 2:56
  So I will just like write back and say like, what am I grateful for?
• 2:59
  What happened?
• 3:00
  And yes, I did give my neighbor the eggs.
• 3:03
  The reason I'm liking GrokBots right now is because I can separate out different parts of my life into the different bots, as opposed to like dots or open claw or any of the older ones.
• 3:16
  It's kind of like one single thread that you just keep building upon.
• 3:20
  It's kind of hard when you're like working on five different things at once at every given moment.
• 3:26
  It kind of looks like iMessage or something like that.
• 3:29
  And you're selecting the bot or the friend that you want to text.
• 3:32
  And you would only talk to each bot about one different aspect of your life.
• 3:37
  Yeah, it's kind of like when you're working with a team, you work with one person to do like certain roles and another person to do another set of roles.
• 3:46
  So Arcanine manages my newsletter and the way I have it set up is there's a main bot that coordinates all of the other bots.
• 3:55
  So kind of like if you think of like a project manager or like a content manager...
• 4:00
  He manages a team that each has their own jobs.
• 4:04
  The way that I have my system set up is I have all of my instructions actually written out into the pages of Notion instead of in memory folders within the bots themselves.
• 4:16
  That way it's easy for me to like swap between bots.
• 4:19
  So even though my GrokBot is the one like coordinating a lot of the management between steps...
• 4:26
  I actually have Codex or another AI come in and actually do a lot of the thinking work.
• 4:32
  I don't actually give Codex a big prompt.
• 4:35
  I just say like every day at 8 a.m. sweep through this page and see what tasks need to be done.
• 4:42
  And it will just come here, follow all of these rules and then look at the next article, for example, that needs help.
• 4:51
  Interesting.
• 4:52
  So it sounds like all of the substance is housed in Notion.
• 4:56
  That's kind of like your source of truth.
• 4:58
  Yeah, basically.
• 5:00
  Because all the AIs are coming out with like new models every literally like 10 days.
• 5:06
  And like, I always want to use the best one because when you're not using the best one, it's just like, why not use the best one?
• 5:12
  Right?
• 5:12
  I...
• 5:13
  have had to switch so many times in the last few months that I've found that the best way to manage instructions is just in the files themselves as opposed to with the AI providers.
• 5:25
  Yeah, that was a big problem for me because I did a lot of my work in Claude when Claude was like obviously the best.
• 5:30
  And then ChatGPT came out with Sol and it got better than Claude for a moment and I jumped ship, but it was such a headache to actually migrate over.
• 5:39
  because so much of my life was embedded in clod already.
• 5:42
  So I like this line of thinking where you actually store it externally and then you just point your tools to that source.
• 5:49
  I want to ask you about using one single GrokBot versus using many, because I'm trying a GrokBot as well.
• 5:56
  It's still fairly new, as are dots.
• 5:58
  And I haven't gotten to the point where I have made many bots.
• 6:02
  I just have one, and I just keep talking to it about everything.
• 6:05
  What is the pros and cons of having multiple?
• 6:08
  I think for me, because I'm doing so many different things and like I'm in a position where I'm not actually making content a lot of the time.
• 6:17
  I'm overseeing like a broader pipeline.
• 6:19
  So for me, it's like knowing where everything is and keeping track.
• 6:24
  the whole pipeline in my head at the same time, it's a lot easier for me to think about each task separately as opposed to just like I'm making one thing or doing one thing.
• 6:35
  And the reason I like GrokBots or having, it doesn't even have to be GrokBots, it's just like multiple threads for it.
• 6:40
  I don't muddy up my conversation of working on a newsletter article specifically
• 6:47
  with a podcast that's like different topics while also trying to like work out my like travel plans or sending eggs to my neighbor.
• 6:55
  It's so many different aspects of my life in one chat gets really messy.
• 6:59
  But the problem with like a traditional threads is that like
• 7:03
  Each thread kind of is, it's like a non-recurring task in many cases.
• 7:07
  You start something new and you have to give it all the context all over again.
• 7:11
  And then that's it.
• 7:12
  And then it's done.
• 7:13
  And it like maybe saves its memory somewhere.
• 7:15
  Maybe it doesn't.
• 7:15
  But with the GrokBots, it's more of, it's not really a task.
• 7:19
  It's like a job.
• 7:20
  It's something that keeps happening.
• 7:21
  So it's doing this constantly and proactively.
• 7:24
  I think that's the main difference.
• 7:26
  When you're communicating with your Grokbots or with any AI, do you speak to it as you speak to people or do you have some kind of different vernacular in order to communicate more effectively with AI?
• 7:39
  both.
• 7:40
  In my day to day, I will talk to it just like it's another human.
• 7:43
  They're good enough where they can just take what you said and like understand you.
• 7:48
  But there is also a level of like, the better the prompt, the better the output.
• 7:52
  So if you're trying to do something like very complicated or thoughtful, then I would spend more time writing out a thoughtful prompt because there are things that like if you're not descriptive enough, it might not understand or
• 8:04
  You know, when you're making content, when you're trying to, we were just working on carousels, for example.
• 8:08
  And it's like, if you say, make it prettier, like it doesn't know what that means.
• 8:13
  But if you say like, increase the saturation or it won't know like, oh, this looks messy.
• 8:17
  But you can say like, organize it in a grid.
• 8:20
  It depends on what you're trying to do.
• 8:23
  But for the most part, you can speak to it like a human or a teammate.
• 8:27
  And then just know that, like, if it's not understanding you, you might need to be more descriptive.
• 8:32
  But that's the same way that you would need to talk to a human, too.
• 8:35
  Like, if I just told you, like, Devin, your carousels are ugly.
• 8:38
  Like, you don't know either, right?
• 8:40
  So...
• 8:41
  Right.
• 8:42
  Yeah.
• 8:42
  You know, it's funny that you mentioned it because as I've started communicating with AI a lot more, I've gotten more granular in how I communicate with like video editors or graphic designers as well.
• 8:52
  I remember when reviewing people's work, I would say things like, oh, you know, add a beat here or slow down the pacing a bit or speed up the pacing a touch.
• 9:02
  And all of this terminology is very vague, and you kind of rely on the editor's intuition to know what the difference between a touch or a beat is.
• 9:10
  But now that I'm communicating with AI and I'm using AI to generate carousels and doing thumbnail mockups and things like that, I find that even when I'm speaking to editors, I say things like, oh, you should add three to five frames of a gap before moving on to the next scene.
• 9:24
  And that communicates the same thing, but it's much more granular.
• 9:26
  And I'm finding it really helpful for humans, and I'm finding it really helpful for AI as well.
• 9:30
  especially with thumbnail mockups, because I spent a lot of time working on a bunch of different mockups, right?
• 9:36
  We're now testing at least three thumbnails per every YouTube video that we're putting out.
• 9:40
  And AI can do a really good job getting like 80% good.
• 9:43
  But in the CTR game, in the thumbnail game, you have to really polish it so that it is a perfect execution of your vision.
• 9:50
  And if you're saying things like, oh, brighten it up a bit, AI doesn't know how to handle that.
• 9:54
  And so you have to say, brighten it by 20%, increase the saturation of the reds by 25%.
• 10:00
  I think managing a team of bots probably makes me a better manager because it's like they do as you say and it forces you to be more thoughtful in what you say.
• 10:10
  There was like a real I saw on Instagram a few months ago where this father was teaching his daughter how to like give instructions and it was just like she would have to give instructions to him on how to make a peanut butter sandwich and like he would carry out the instructions exactly as she said and it was always wrong.
• 10:28
  And it was so funny.
• 10:29
  And I feel like that's a really good example of like how you present information.
• 10:33
  I've seen that reel and I think it's actually hilarious.
• 10:36
  I want to ask you about using different AIs for different things.
• 10:39
  So we've talked about GrokBots and how they're kind of coordinating and doing their own individual task.
• 10:44
  Is that the best AI for every kind of tool?
• 10:47
  Like, are there AI tools that are better at deep thinking, at executing and modifying your Notion page?
• 10:53
  Okay, first off, I think that the best tools are always going to be changing.
• 10:57
  So what I say right now is going to be different from what's true in like three days, because there will be a new model that does something better.
• 11:04
  But as of like right now, what I found is I really like Rockbots for extracting information from me and adding that, like logging it.
• 11:14
  So that's basic coordination work and like project management work.
• 11:18
  And I have not liked it for actually executing on any of the work.
• 11:23
  So I use ChatGPT or Codex in all of my personal work.
• 11:28
  But I found that Claude right now is better for a lot of like designer work or just like thinking and writing.
• 11:35
  The reason I'm using Claude right now is it has really good computer use, which means it like takes over my computer and just starts like clicking things.
• 11:42
  Yeah.
• 11:42
  which is really good for me plus machine and Gimbal because I am a little bit neglectful of it.
• 11:48
  So it kind of just like takes over and does it on its own.
• 11:51
  That's your newsletter and our podcast.
• 11:54
  Yeah, I think maybe we can walk through like maybe a Gimbal guest or Gimbal workflow.
• 12:00
  So now we are going behind the scenes into Gimbal podcast and the infrastructure that holds it all together.
• 12:05
  The official team one is a little bit, it's very simple.
• 12:08
  It's like, this is the simplest way to set up a Notion system.
• 12:12
  This is our gimbal behind the scenes database.
• 12:15
  This is our Notion.
• 12:17
  I don't actually look at it that often because I interface with it through AI through GrokBots.
• 12:23
  But just to like get you a little bit oriented, the way that I have Notion set up is I have
• 12:28
  I call it like a database of databases and all information, like everything about Gimbal is in one of these pages.
• 12:38
  And then each one of these database pages is managed by one of the bots.
• 12:43
  Oh, wow.
• 12:44
  You have a different bot for each of these.
• 12:46
  Yeah.
• 12:47
  And you can see them here.
• 12:49
  So each of these databases corresponds to certain bots.
• 12:53
  And Chutney is our like gimbal manager.
• 12:56
  Chutney is the only one that I really talk to.
• 12:58
  And then Chutney will then go in and talk to all of these bots to like move things along the pipeline.
• 13:04
  So these are the three bots that manage all of our Gimbal content.
• 13:08
  So we have production tracker, guest researcher, and content drafter.
• 13:12
  And this is a list of like all the podcasts that we've ever recorded.
• 13:18
  And then if you go into each one, like each of these bots is doing a different job to prepare us for these episodes.
• 13:24
  So we booked Ernesto Perez as a future guest.
• 13:29
  I just got the calendar invite added into my calendar and it saw that
• 13:34
  Oh my gosh, it did all the prep for you immediately.
• 13:36
  Wow, that's so cool.
• 13:37
  Yeah, so what happened is it saw that I am interviewing Ernesto soon, and it went in and updated the pipeline and said, okay, the status of Ernesto is interview booked, so I'm going to go in and tell Codex to do the background research for Ernesto.
• 13:59
  Codex then went in.
• 14:00
  I have a prompt for Codex that's just like, look through the pipeline and see if any statuses have changed.
• 14:07
  And if the status has changed, complete the, like fill out the template.
• 14:11
  That's basically all it says.
• 14:12
  It doesn't have any details.
• 14:14
  All the details are on the page.
• 14:17
  I want to ask about setting up these automations, because it sounds like the work is impressive once it's there.
• 14:23
  But how do you think about setting up an actual automation that does all of these things independently?
• 14:29
  For me, I know the steps of what I want to happen.
• 14:33
  I just have this like broad pipeline view of like this goes to this, goes to this, goes to this.
• 14:38
  If you wanted to start and you didn't have that in your head already, I would just start with one bot and tell it you are the manager of this system and...
• 14:48
  Spin up whatever bots you need and then I will talk to you.
• 14:52
  You will not do any work.
• 14:53
  You will tell other bots to do work and then just kind of like just trial and error from there.
• 14:58
  So every time it it'll do something and you'll be like good, but next time do this instead and just like kind of keep iterating and it'll it'll like naturally trend towards this kind of system.
• 15:09
  I'm setting up automations within my team right now, and I have an idea of what I want.
• 15:16
  And what I've noticed is I'm like dictating my process to Claude or GrokBot or whatever AI tool that I'm using.
• 15:22
  And one unlock that I've had recently is actually before doing that process, asking it how it would do the process by just defining the goal and the problems and then seeing what it comes up with.
• 15:33
  And I find more times than not, there's at least some of its answer that gets incorporated into the actual process that I just overlooked.
• 15:40
  So I'm finding it really helpful from like an asking it perspective, but also then a telling it perspective.
• 15:45
  Yes, I think there's a lot of like little tricks.
• 15:48
  So one, for example, is you said you dictate.
• 15:51
  I love Whisperflow.
• 15:52
  I just like blab into Whisperflow all day.
• 15:55
  That's like my job.
• 15:57
  I'm such a big proponent of it.
• 15:59
  Like typing is not a thing anymore.
• 16:00
  It's all.
• 16:01
  I don't know how to type anymore.
• 16:03
  Yeah.
• 16:05
  I talked to Whisperflow all day.
• 16:07
  I just go between chats.
• 16:08
  It's like the Slack chats from work, the Slack chats from Gimbal, the AI chats, chats, chats, chats.
• 16:14
  So that's the first thing.
• 16:15
  The second thing is like exactly what you were saying.
• 16:17
  There's a there's a lot of like little tricks.
• 16:19
  Like I used to say, do you understand?
• 16:22
  And it always say like, yes, I understand.
• 16:24
  But if you say, can you repeat this back to me, then it'll like repeat.
• 16:28
  what it thinks you said back to you and then you can correct it before it actually goes on.
• 16:33
  Oh, that's so interesting.
• 16:34
  Yeah, it's like the more you work with it, the more you realize it has certain tendencies and then you can kind of like control for those tendencies.
• 16:42
  Yeah.
• 16:42
  So you have all this research that leads you to the interview.
• 16:46
  What do you do with Gimbal once the interview has been recorded?
• 16:49
  So after the interview has been recorded, that has different humans involved.
• 16:54
  So we have an editor that goes off and like edits.
• 16:57
  We have a thumbnail designer that makes our thumbnails.
• 17:00
  We have a social media coordinator manager who helps with like making and scheduling and coordinating everything.
• 17:05
  And that's where Notion is really helpful because everyone can come and see the status of each episode.
• 17:12
  This is where the information lives.
• 17:14
  Like this is the source of truth for everyone, both humans and bots.
• 17:18
  What my bots do after the episode is, for example.
• 17:22
  We have Content Drafter and after you shot with Anthony last week, it came in and just created all of these potential posts.
• 17:33
  And I would never just like straight up send this out because I think like AI does a good job but not...
• 17:40
  A perfect job.
• 17:41
  Like AI gets you to 70%, maybe 70% good, but it's really that like 70 to a hundred of human effort on, on this like content type stuff that makes something good.
• 17:53
  And also humans are good at catching AI.
• 17:57
  I think it's just good hygiene to have human eyeballs and fingers go over top of all of the work.
• 18:04
  Like, I think it's okay to use AI for like the foundation of something.
• 18:07
  But I really truly believe that good work has humans doing the last step or at least modifying it or like rearranging it in, you know, using our own human brain.
• 18:19
  I agree.
• 18:19
  After an episode, we have a bot that comes in and extracts the transcript.
• 18:25
  And out of the transcript, it will come in and find frameworks for basically everything that it said in the episode that was like something that we should take away.
• 18:49
  I'm so impressed with how good of a foundation we have with Gimbal.
• 18:53
  The transcripts instantly turn into a really strong newsletter.
• 18:57
  So in case we ever want to roll out a newsletter, we've now got one single newsletter for every single episode that we've recorded.
• 19:03
  The frameworks thing that you just pointed out is excellent, and I'm using it right now to build out carousels.
• 19:08
  And it's such a seamless process.
• 19:10
  You know, I think there's a world where those get turned into scripts that we film for reels.
• 19:15
  That's the part that I'm like really excited about with Gimbal right now.
• 19:17
  Like we've got this...
• 19:18
  the interviews, which is feeding new content, and then the way that we're planning to disperse it over the next couple of months is really energizing and cool.
• 19:24
  I found myself in my daily work life, like my real job, referencing things that I learned from guests on Gimbal recently.
• 19:33
  That's made me really excited about what we're doing with Gimbal.
• 19:36
  And it's kind of made me extra bullish on creating this frameworks database because we have this like body of knowledge that we can share in different formats.
• 19:46
  I agree.
• 19:46
  I find gimbal really helpful for my regular job.
• 19:48
  One of the things that I think is extremely underrated in a Notion system is using all of these properties.
• 19:57
  For each episode, I have the AI filling in these properties.
• 20:01
  And the reason a property is better than a page is because this is very easy for the AI to identify and fill in and just know that it's doing the right thing in the right place in the right way.
• 20:15
  And it's helpful for humans because
• 20:17
  Riverside link is here.
• 20:19
  I see a lot of teams using a page, which is, it's okay.
• 20:23
  Like the AI is getting better and better and it can read it, but just having structured data is so much easier for everyone to like identify information.
• 20:32
  AI has a tendency to just keep appending information.
• 20:36
  Like it'll just keep adding more and more and more to the bottom of a page.
• 20:40
  That just makes the information messy, which is why I like having the structured place that to know where everything is.
• 20:47
  On my team at Mark Manson, we use Notion as our second brain as well.
• 20:51
  Actually, it was built by you back when you worked there and it's been modified since and catered to like the strategy that we're running on right now.
• 20:57
  But we have a lot of properties and we also have the body of every Notion card.
• 21:02
  But you've kind of changed my line of thinking, like maybe the script of a short could become a property and it could be more manageable by AI rather than keeping it in the body.
• 21:10
  I will say, though, like the body of our Notion records do also have templates that are followed very explicitly.
• 21:17
  And everybody on the short form content team or whatever team tells their AI tools to follow that exact template so that it all looks uniform.
• 21:24
  I remember when I first got onboarded to Notion, like it was really daunting because it's full of so much information, full of so many opportunities.
• 21:31
  You can literally make it into anything you want it to.
• 21:33
  And that's so overwhelming.
• 21:35
  But with Notion AI as a tool that's been like integrated into the system since then.
• 21:39
  You actually don't have to know everything about Notion to use it effectively.
• 21:42
  You can just vocalize what you want or what you think you want or the problem that you have, and it will generate the solution for you.
• 21:50
  So I think Notion is like more accessible than ever before.
• 21:52
  And I love it.
• 21:52
  Where do people begin with this new information that you've just shared with them?
• 21:57
  Using Notion as a brain, using GrokBots or any other kind of AI to interface with Notion?
• 22:03
  How do people do this as a starting point?
• 22:05
  If I were to just start out, I would start with one AI, any AI, and tell it what you want it to do.
• 22:14
  You have a blank page, you copy the link from that page, give it to AI, any AI and say, this is our source of truth.
• 22:23
  I want you to document everything that happens in this page.
• 22:29
  Every decision that we make, add into a decision log, every like
• 22:34
  process add it into like a process doc and then from there I would just have it keep documenting what you want you can say like interview me until you understand what I'm trying to do and you can go through like multiple rounds of talking I will just tell it to store all information in a database I think that's really important and like add properties for anything that's like reoccurring
• 23:00
  It's funny, Ali Abdaal has this saying, get going, then get good.
• 23:04
  And this is something that I apply to like every degree of my life.
• 23:08
  And so if I were in the situation where I am seeing all these people do incredible things with AI, like what you just demonstrated with AI,
• 23:15
  you know, as soon as a guest gets confirmed, having all of the research and questions automatically populate for you.
• 23:21
  As soon as an interview gets completed, having all of these things turn into a bunch of frameworks, a newsletter, you know, X, Y, Z.
• 23:28
  That's really overwhelming and daunting.
• 23:30
  But if you zoom out and you just want to like get going, I really like what you just described, like take a very blank notion brain with nothing inside of it.
• 23:39
  And then just say to Claude or CrockBot or whatever AI tool you choose, look, I'm new to this.
• 23:45
  I don't know anything about anything.
• 23:46
  This is the problem that I have and that I'm trying to solve.
• 23:49
  Please help me solve it.
• 23:50
  And I love the little tip that you added there about getting it to interview you.
• 23:54
  When I was like really ramping up with my AI usage, I didn't know what I didn't know.
• 23:59
  And so when Claude was interviewing me, I found it did a really good job pulling stuff out of me that it needed in order to achieve the goals that I'd set out for it.
• 24:07
  Because I didn't do a good job like verbalizing that to it.
• 24:10
  It thinks about so many things that I don't think about.
• 24:13
  And it's like, oh, that's actually really important.
• 24:16
  Another, like something that I tell all of my AIs is learn from our conversations and then fix our systems so that we don't have to go through this again.
• 24:25
  It's like that same thing where you can make a mistake once, but you can't, like, don't make it twice.
• 24:30
  In a lot of my work with, like, writing the newsletter articles, for example, it'll often make some silly mistake that's, like, you know, like the over-structuring of words.
• 24:40
  Or, like, even in the carousel designs that we were making over the weekend, it was, like, kept adding the eyebrow text or kept adding, like, random squiggles that really didn't need to be there.
• 24:49
  So I would say, like, I don't like this.
• 24:51
  Make sure this doesn't happen again.
• 24:53
  And it will add that information into the decision log.
• 24:56
  I think as long as you think of your work as a work in progress, you will have a higher degree of success.
• 25:00
  Because if you are trying to create something big and impressive, you'll either never get started because it's too daunting or you'll stumble along the way and then give up or start over.
• 25:12
  And the realization is like it's always a work in progress.
• 25:15
  Even for someone who is an expert like yourself, you just mentioned like that you are constantly reiterating on your own work.
• 25:21
  So just creating something by getting going, as Ali Abdel says, and then refining it along the way seems to be the right way of leveraging AI as, you know, a second brain, as a way of doing all of these tasks that are recurring and boring for you, but easy for AI to pull off.
• 25:37
  And if you do that, I think your productivity will 5x, 10x.
• 25:41
  I know mine has in the short time that I've been using AI.
• 25:44
  My last question is, who is a creator, maybe even in the AI space, who you're learning a lot from, or who is on your radar right now and is worthy of a shout out?
• 25:53
  Okay, so for AI things, I think it's better to not look at AI creators, but like the creators of actual AI.
• 26:02
  The person that I'm really obsessed with at the moment, her name is Lauren Tan.
• 26:06
  She is a developer, I guess, of GrokBots.
• 26:10
  And she's just so smart.
• 26:12
  She thinks of like all of these very smart ways to use AI to be so much more productive in work.
• 26:18
  One of my favorite GrokBots is actually Dr. EggBot.
• 26:20
  It's one that she made.
• 26:21
  You can get it from like the GrokBot marketplace.
• 26:24
  And it just, it has a lot of her systems built into it.
• 26:26
  Like her philosophies for how to set things up.
• 26:29
  It's like very efficient.
• 26:31
  I use Dr. EggBot to create all of my bots in a way that's most efficient.
• 26:35
  I'll just tell it what I want it to accomplish.
• 26:38
  And it will create like the recurring schedules and tasks and like set it up for success.
• 26:46
  How do you follow somebody like Lauren Tan?
• 26:48
  I follow her on X.
• 26:49
  She's at Potato.
• 26:51
  Is that where the majority of conversation around AI is happening right now?
• 26:55
  I would say not just her, but there's a lot of people that are actually working on AI that are so smart and like they're not really influencers, so they don't have huge followings, but they'll just share things that are huge unlocks.
• 27:08
  Great.
• 27:08
  Well, thank you for sharing your knowledge around operations and AI.
• 27:12
  I know that I've learned a lot from you over the last like two or three years that we've worked together.
• 27:16
  And I think all my foundational knowledge around Notion or AI as a whole has come from you.
• 27:22
  And as a result, I feel like I'm five to 10x more productive than I was this time last year.
• 27:26
  So thank you.
• 27:27
  You're welcome.
• 27:28
  Thank you so much for watching or listening to this episode.
• 27:30
  If you're new here, please consider subscribing or leaving us a review on the audio platforms.
• 27:34
  It helps us as a very new show.
• 27:36
  If you want to check out another episode, just look through our library.
• 27:39
  It's starting to build up.
• 27:39
  We're at like 40 episodes now.
• 27:41
  And we have a new one coming out every single Thursday.
• 27:43
  So thank you for being here.
• 27:44
  And I will see you then.