Episode Transcript
[00:00:08] Speaker A: Welcome to Examining, a technology focused podcast that dives deep.
I'm Eric Christensen.
[00:00:16] Speaker B: And I'm Chris Hans.
[00:00:23] Speaker A: All right, so welcome to another episode of the Examining podcast, the technology focused podcast that dives deep.
I'm Eric Christiansen and I am joined with my co host, Chris Fons. How's it going, Chris?
[00:00:40] Speaker B: It's going well. Yeah, we're, we're experiencing some decent, I don't know, unusually warm weather in Alberta.
[00:00:48] Speaker A: Yeah, this is great. Well, maybe not great for like the world, but it's great currently. It's comfortable. So that's, that's fine. Yeah, yeah, yeah, I agree. But yeah, it's nice and warm. Today we have kind of a show that's kind of all about search and automation. So I think we were going to just kick it off by talking a little bit about what's happening between and there's a bunch of sources associated with this story, but kind of what's happening between Apple and Google, specifically the kind of the two way street of money pouring in between the, you know, Google being the default search on Apple's devices. That's an old story, but there's some updates to that as well as the more recent trend where Apple is now sending money to Google the other way, where they're going to use Gemini to help power at least part of the new personalized series. So we have kind of a weird, strange rivalry, frenemy relationship here. So we're going to talk about this before we go into some of the other details. So just a bit of a primer to people.
It's been known for several years now, but basically Google pays Apple roughly $20 billion. Is it a year?
Yeah, it must be a year.
So that's as of 2022. So the number may have changed to be kind of the default search engine on iPhones, iPads, Macs and Safari, essentially. And so this was revealed through kind of unsealed testimony from the Apple svp, Eddy Q. During a Department of justice antitrust trial. I think that was the antitrust trial of Google. And so then also there's internal kind of Google analyst estimates that the company would lose between 28 to 33 billion in revenue if it lost its position as the default search status on Apple devices. So that would see kind of a 60 to 80% drop in search query volumes according to Apple Insider.
And so this is just a history.
According to the Search Engine Journal, it's estimated that that drives roughly half of Google's total search traffic because of course mobile makes up so many computing devices. So that's two and a half billion devices worldwide. So it's an enormous scale.
And in 2024, August when there was the, when Google was going through the antitrust case that the U.S. department of justice, you know, they basically the DOJ found that Google was guilty of operating an illegal search monopoly. Because of course that deal kind of freezes out other potential competitors to the search market. I mean search is absolutely dominated by Google and as a result Google Ads and their ad platform, which is where they make most of their money, dominates the ad network.
But the September 2025 ruling, despite that, allowed Google to keep making these payments.
Just can't enforce exclusivity anymore. And the deals are now limited to one year terms.
So they didn't really do anything about it. It was kind of an unveiling, but nothing happened. But now we have money going in the opposite direction because of course Google has partnered with Apple officially to use Google Gemini to power the new personalized Siri. So what's happening with that, Chris?
[00:04:08] Speaker B: Yeah, so they're paying apparently about a billion dollars to go and access the, the cloud infrastructure from Gemini model to go and power this personalized in house Siri platform and. Oh damn my, my phone's just by.
[00:04:29] Speaker A: Saying you set it off. Oh my God, I set it off.
[00:04:33] Speaker B: So yeah, it was interesting because Google beat out, you know, OpenAI and Anthropic were also in the running for this and prior to that OpenAI was the one that they partnered up with. But you know, probably because of Jonathan Ivy joining OpenAI and you know, their ambitions to go and develop some sort of hardware device, they kind of exited that relationship. And so then do you think they.
[00:05:03] Speaker A: Would have been the front runner had it not been for that?
[00:05:07] Speaker B: Probably. I mean if you look at it like, you know, OpenAI they're powering copilot as well for Microsoft. Right. And I think in all estimations right now, usually they, they have the more well used and I guess effective models.
[00:05:29] Speaker A: Yeah, I think from monthly user perspective, like I haven't looked it up recently, I can check now, but OpenAI historically has dominated average monthly users over all AI platforms.
[00:05:40] Speaker B: Yeah, totally. So but yeah, I mean at the end of this, I mean it kind of, I think strategically it probably makes sense. I mean you're, it's like, you know, you're getting $20 billion now, we knock off 1 billion, so you take off like you know, 5% of the revenue. You're essentially getting, you know, this 1 trillion parameter model, you know, for free from them.
[00:06:07] Speaker A: Right, so, so I mean, estimated currently as of so stats counter. I mean, take this with a grain of salt. This is January to January 25th to January 26th. ChatGPT still commands 80% market share.
[00:06:22] Speaker B: Yeah, there you go.
[00:06:24] Speaker A: And Gemini's third at 7%. Perplexity, amazingly, is second at 7%, just under 8. Microsoft Copilot is three and a half though, I mean, depending on how you measure that, because that's also chatgpt, so I'm not sure how you would dice that up. And Claude is 0.92 apparently. So I've really gone into the niche market by moving to Claude as my daily driver.
[00:06:49] Speaker B: Yeah.
And then again, like you mentioned, so Gemini is not going to power all of Siri. It's going to handle the summarizer, the planner functions. So synthesizing information and executing some complex features.
And while some of the other features will still run on Apple's own models, and again, this is a bit of a stopgap so that they can keep looking at building their own models to go eventually replace Gemini.
[00:07:20] Speaker A: Yeah. Are you able to turn up your mic volume a bit?
[00:07:23] Speaker B: Has it gone down?
[00:07:25] Speaker A: It just seems a bit quieter. I don't mean to interrupt you, I just.
[00:07:28] Speaker B: Yeah, no worries.
[00:07:28] Speaker A: My headphones are maxed and I'm like, is there something wrong with my headphones or.
Unless it could be my AirPods, it could be my fault.
[00:07:46] Speaker B: What's going on here?
How is that? Is that better?
[00:07:50] Speaker A: Seems like the same. I wonder if it's just quiet AirPods.
I can always switch headphones.
[00:07:59] Speaker B: I could just turn up the gain. We're gonna have to cut this hole.
Is that better?
Is it better?
[00:08:11] Speaker A: Hang on a second. I'm just gonna switch here. Can you talk again?
[00:08:15] Speaker B: Yeah. How's this?
[00:08:17] Speaker A: That's better. I just got rid. My AirPods are super quiet lately. I don't know what's wrong with them. Maybe they're busted, but yeah. Okay, so basically now with Money going in two different directions, Apple gets $19 billion a year instead of 20, roughly from Google, if you subtract the billion dollars that they're going to pay. This is a strange arrangement.
[00:08:40] Speaker B: It totally is. I mean, again, I think really, you look at it, it's like basically you're getting it for free. And so what's a 5% of our revenue that is being thrown off and can cover this off? But yeah, the interesting thing that I, I, I don't know. Especially when you're going, you know. Alphabet, which is the parent company for Google, they are publicly traded company. And during the earnings call, this, this whole topic was kind of avoided in terms of the discussion of what the particulars are in terms of the revenue and so on.
[00:09:24] Speaker A: Yeah, because they're going to be kind of not just a Gemini chatbot, it's a cloud infrastructure deal to talk about. I mean, so they must be making some money off this or. Anyways, it's just a strange kind of money management question.
[00:09:43] Speaker B: Yeah, totally. And so, yeah, I mean, at the end of it, yeah, Google is now going to be its preferred, Apple's preferred cloud provider for this. And while they're still going and developing their own generation of models.
[00:10:02] Speaker A: Yeah, I mean, to me it's interesting how it impacts kind of the whole ad model. So Google is going to start experimenting with ads and it's kind of AI mode search interface, at least that's the rumor from TechCrunch.
So there's gonna be. And we already heard of this from ChatGPT, right. So like ChatGPT is gonna start putting ads, probably the free version, but then there's also a lower tier. And so I wouldn't surprise me if ads even make their way into ChatGPT+ at some point, which is kind of infuriating because that's what you pay for, to not have that stuff. Right.
So far, the only AI platform to say, no, we don't need any ads is Anthropic. But that's not surprising because they're not really bad company.
They have consumers that use it, but they seem to have a lot of enterprise deals, especially from a coding standpoint, because Anthropic is such a good coding co pilot. It's great for writing. It has a huge context window. That's why I've been kind of using it full time. So that doesn't surprise me because it seems like my limited understanding they're probably in a better position from a revenue profit standpoint. Even if they're only breaking even at this point, it doesn't seem like they're losing as much money on their platform. So there's like this huge ad push. And so you know what happens.
So it almost seems like Google wants this deal to go through to some extent, because if they're going to push Ads into Gemini, that further kind of solidifies their dominance in the ad market, which is kind of where all their money comes from.
[00:11:34] Speaker B: Well, yeah, I mean, they basically have built a $300 billion, you know, revenue empire based on search. And if everybody now is going and using AI to search, it's going to definitely impact Their overall, you know, revenue coming in right, just from the web surge.
[00:11:53] Speaker A: It's also interesting to me that like Apple has kind of treating Gemini. I mean, I guess they've alluded to this, that it's a temporary measure. Like ultimately, historically, Apple doesn't like to not own the core technologies. I mean, so it partners with companies all the time for things where it doesn't need to own the core tech. But ultimately they like to own the core tech. You can see this in hardware, for instance. So Apple moved away from using cellular modems from Qualcomm to designing their own.
And I think originally maybe designed them with intel, but ultimately they design their own chips, they have their own software. I mean, of course they don't manufacture the chips, the TSMC manufactures the chips, but Apple designs them. So they own the whole stack. And that's one of the reasons that they've been so successful, because things integrate really well. Everything works well together.
So Siri to me is kind of a core technology. I mean, it's funny, people make fun of this stuff, but they were the first to come out with an assistant. And as limited as it is, it does work for the limited things that it does. And so to me that's still.
Even though it's been almost 15 years since they though it has been 15 years since they launched it, it's never really gotten to where they've advertised. It is a core technology. It is a core integrated feature throughout all their stuff. So it's interesting that they have to rely on a third party. I can only imagine that their goal is to do this, that they're not going to go into competition with ChatGPT or Anthropic. Their goal is to have as much of this their own models in the secure cloud or working offline with their. Because all their devices have NPUs for doing kind of local AI workloads and stuff like that. I can only imagine that their goal is to get away with this. And I would say the same thing for Microsoft. They probably also don't want to be dependent on OpenAI's models, even though they tend to adopt them right away when there's an update.
And so it's weird how this is going to work like all this money changing hands, these relationships, eventually they're going to go their own way.
One thing that interests me though is that why didn't Apple go their own way when it came to search?
So my understanding is that, and maybe I'm wrong and we should double check this, but my understanding is that Siri which will come back with search results, has a web crawler.
And so that's how it does it. It's not, you know, results from Bing anymore. That was like early days. Right.
So why, why use.
Why?
I guess, I guess it's just because that makes money. That's why they're using Google's product.
[00:14:44] Speaker B: Yeah. I think it comes down to that $20 billion that they're getting. Right. They just, you got the revenue coming in.
So. But yeah, I mean there was even talks of them. Yeah. Creating their own search engine at now Apple Bot.
[00:14:59] Speaker A: Apple Bot is their web crawler. So I mean they could, they could have done this.
[00:15:03] Speaker B: Yeah, yeah, yeah. But again, I mean now you have an influx of revenue and again, I think in terms of this now it might be a pivotal kind of point where everybody is transitioning from the traditional search engine to like an AI search. And so it might be a good timing for them. Right.
[00:15:27] Speaker A: Well and that's a really good segue into what we were going to talk about as well. Right. So one of the things that came out of the Nielsen Norman Group, so there's a great article that we'll put in the show Notes called how AI is changing search behaviors. I think we've talked about this before but we wanted to touch on it because we're going to talk about our own experience in a second. And they basically said that AI, they make the argument that AI is shifting shape, shifting search. So all these kind of decades old habits of going to search and everything going through search is going to be upended through the introduction over time of generative AI. So users still default to Google, but now you're seeing a little bit more, a larger share of people defaulting to ChatGPT or even Google's Gemini as a starting point to find information. So traditional search isn't going away, but it has competition now in terms of an interface. So they argued from their kind of nine person, nine participant study that AI overviews are kind of the gateway drugs. So by posting AI overviews and you know, brave search, Google search, even DuckDuckGo people are getting exposure to these AI summaries, which is kind of enticing because they give a much more concrete kind of to the point answer versus kind of digging through search results. So that gives the aha moment. Oh, this is what AI can do in terms of retrieval.
And then they also argue that AI solves kind of tedious research problems. So you know, defining queries. So we talked about this before in our or I talked about this in the solo episode. Like why prompting is so hard. Right. Like, because you're used to creating search queries out of, you know, complex prose that are running through your mind. Well, now you can just spit the prose out and you can have AI interpret that as a search query and then sift through the information after the fact that tell it to tweak it.
They also said there's familiarity wins, like things like ChatGPT and Gemini dominate because they got there, essentially because they got there first, and then discoverability. Right. So many consumers don't realize what AI can do in terms of information seeking, but when people see it, it's kind of like this aha moment. So does that kind of overhaul people's habits in the long run?
[00:17:48] Speaker B: Yeah, no, for sure. And again, I mean, we've seen all sorts of creative ways that the. These large language models have been used, like, you know, for travel planning. All of a sudden now I can give you a full itinerary and, you know, ideas of what type of things to go and visit. Right. And to go and compile that through search, it would have taken, who knows?
[00:18:13] Speaker A: So, interesting. You mentioned travel planning, so, I mean, I did it in reverse to see how well this is before I moved kind of over to Anthropic, my daily driver. But I asked ChatGPT to do some travel planning for me to a place I had already been knowing that it knows my preferences and background in international relations. And I asked it to plan a trip to Florence, Italy and Rome and, you know, what would the sites be? And blah, blah, blah. And it basically told me to go see exactly what I did.
I mean, it was remarkably close. It even told me to go see, like, some of the niche galleries where some of the. Oh, I'm trying to remember the names of the galleries.
Where, like, some of, like, the Benigni sculptures are, like, really, you know, in addition to going to see this coliseum and the David and stuff like that. And it totally guessed it and it gave, like, you know, here's where you buy tickets, here's where you can get advanced tickets here. Are the hotels close? Here's a bed and breakfast. You know, gave me all these options. And it was just like, at that point you could just kind of go pick and book and choose. Right. It could play in your whole day if you said, I want to do one big thing per day and then just go explore. It would build a whole trip around however many days there were. It's incredible.
[00:19:30] Speaker B: Yeah, yeah, no, for sure.
But yeah, again, right now I think we're at a point where like these habits they're somewhat sticky and so you know for many people that Google search is like muscle memory for people.
[00:19:46] Speaker A: I still use it, I shouldn't say it has upended it for me but do you go to Google's default still or not?
[00:19:52] Speaker B: I, you know I still use it here and there but especially when, when I look at the AI is a, a big shortcut when there's some sort of, what would I call it like research friction, you know, like if I want to go like one of the things and you know these are some of the, the ways that we were going to discuss like how we're using AI But I mentioned to you I'm I'm looking at buying tires right now and so in the past what would I have done? I would have gone through, you know, done a search on Google maybe I look at various vehicle type of websites and where they have like Consumer Reports or what have you and then I have to go through a whole bunch of different sources and you know, then figure out which is the definitive source. Now instead I basically, you know, the it's not that AI is like replacing search, it's just kind of reordering it. And so now I can go and get the, the AI kind of compiles everything together and then I can go and double check all the links and look at it for myself. So it's just a, it's a different workflow.
[00:21:05] Speaker A: It's interesting you mentioned tires because I did that a while back. I mean we use all weather because those are basically winter tire certified tires you don't have to take off for people they don't know it's different than all season.
And so I don't have the patience to deal with changing tires especially in the Calgary climate because I can never guess when the snow is going to come back. It's like all over the place. So I just leave these tires on but they don't last as long so you have to replace them more frequently. But I, you know I did a search for all weather tires just to see because I had already done the research several years ago and it came up with some new results. But again because I know what the best tires are, what I was interested in is that can you come back with what I already know is the right answers in terms of the top three? And it did and it gave a table and it gave a comparison laybout it was like, you know, your best choices are either Michelin or Nokia.
And I already knew that but. And it, and it did a really good job of summarizing features, brought up points that I had never thought of.
And so it's scraping something and putting this together.
[00:22:07] Speaker B: Yeah, yeah, exactly.
[00:22:08] Speaker A: And I guess that's a good. Just segue into what we were going to talk about because you and I were going to chat about how, personally how AI has changed some of our habits in terms of search, but as well as just like automating and productivity in general. Because we mentioned a few episodes ago we've talked about how it's impacted, but we said we were going to come back to this. I don't remember which episode it was. So this was always kind of at the back burner. And I thought we would touch base again to see and share with our listeners, you know, where our places, new places are continually useful things that we've continued to do where AI has helped. So maybe did you want to get started?
[00:22:45] Speaker B: Yeah, yeah, no, for sure. And maybe prior to that, like, when I look at this, I was thinking too, like, you know, search itself, it isn't really a feature, it is a habit. And so, you know, to break these habits, it's because something new exists. And so, and that whatever that new thing is, it has to be easier. So I mean, the way I'm kind of looking at it, it's like Google is muscle memory and AI is like a new reflex that you have to do. And most people, they don't, they don't choose Google, they just end up in, in Google because it is the, the dominant search engine. But again, I think that's where like the AI, it really has become a shortcut when you're looking at that research, friction, it's, it's not changing what I'm trying to do, it's just changing the, the cost of getting started. And so the, the time sync, it wasn't really like typing up the keywords and it was sort of the, the middle part where you're trying to sort through, compare, synthesize. And so the AI eats a lot of that part out. And so it's, it's really good at the boring parts of thinking then that frees me up to go and focus in on the hard, hard parts.
[00:24:00] Speaker A: I've noticed as a search habit, one of the ways that AI has changed things is that, I mean, you and I are both big readers. So I read a lot. I read articles every day. So like, I know, I know that there's things out there in the news cycle that have happened.
For instance, excuse me, I'm still Coming over my cold. So I don't know if I'll be able to edit that out, but people will understand. I have kind of the long tail of this cold.
But one of the things I've been writing about, I've been working up a post about the Mac menu bar and how it's kind of a necessity now because of the introduction of the hardware notch to have a proper menu bar manager. Because of course things fall under the menu bar. Even prior to the notch, the menu bar can become this unwieldy kind of interface. And it's interesting because it's different than how Windows operates. Because of course Windows maybe sacrilege. I think Windows does this actually better. Where Windows has the taskbar at the bottom, which is kind of both open applications and functions is an equivalent to the Dock. But all the running applications are kind of in a little box that has a built in pop up where you can see like, you know, if you have a, like logitech for your mouse and all that stuff, it's there where Apple's interfaces are kind of sandwiched. You have the dock apps on the bottom and the menu bar items on top. And that creates a strange kind of dynamic from kind of a usability perspective. So I've been working on this and one of the things that happened a few years ago is that there was a very popular app called Bartender. And I promise this example is relevant. But it was controversial because the app was quietly sold to a third party. It was an independent developer and it was never disclosed that it was sold.
The privacy guidelines were updated and one of the things that Bartender requires is like constant screen recording access to be able to do management of your hidden menu bar items. And so, you know, you sell it, you don't tell anybody, you sell it to a third party who then updates like the privacy requirements. It's super opaque, but you're giving this tool potentially full access to your computer. Right? So it's a bit of a concern. So it kind of blew up as a story and I'm talking about how that happened. And I tried to find an alternative, but none of the alternatives work and they suffered from similar glitches. But I couldn't remember exactly how the story unfolded. I knew it had to do with Bartender and I knew that there was an independent developer and they were sold. And I kind of knew pieces of the story, but I didn't know the names and the players and the details. That's where I'm going with this.
So I could have googled this and found it and gone through different links. But instead what I did is that I asked Anthropics Claude and I described this and said, here's what I'm trying to write about and why it's important to me and I need you to fill in the gaps. And can you find kind of some of the articles that broke the story and give me some of the names, like who did they sell to? Who was the original developer?
And it came back and gave me a list and all the links worked, all the links were true. And so by describing this search query in prose, having known part of the information but missing big chunks, it was able to fill in the gaps for me in a way that traditional search would have been very painstaking. And so I do my own writing on my blog. I don't like to have AI write my blog posts for me. I may get it to help me with transitions or edit something if I'm struggling. But I like to write, so I'm not going to have it do that for me because that would take away from the whole purpose. But as a research tool, especially to fill in gaps from my memory, I found it to be enormously useful.
[00:27:51] Speaker B: Yeah, yeah, no, for sure. So I mean, basically you're, you're not searching for answers anymore. It's, you're, you're basically searching to provide that shape.
[00:28:00] Speaker A: Yeah, like I search for answers too, if I'm, you know, if I'm curious about a trend. And I don't know, but it's almost kind of like.
Yeah, it's like having a research assistant follow up or it's kind of, in some ways. I used to painstakingly save articles and I still save a lot of stuff in Pinboard. I love Pinboard. That's my favorite web bookmarking tool. I've subscribed for many years, but I'm, I now only have to save things that I'm really interested in, like reference resources, things that I know I'm going to return to because quite frankly, saving all these bookmarks that I might want is a lot of kind of management. Right. It's a lot of work to manage and save and archive things that I might never use. So the ability to have a digital research assistant bring those things back on the fly based on my partial memory is a much more efficient way to kind of fill in the details when you're doing background research.
[00:28:58] Speaker B: Yeah, yeah, no, for sure. And again, like, this is where I looked at it. Like Google would give us all the links, but now with the AI, what it's doing is it's giving you your first draft of understanding.
And so again, the workflow is changing.
[00:29:14] Speaker A: I do find, I mean, and the summaries, the AI summaries and Google search results are helpful, but I find that extremely limited.
Actually brave search and DuckDuckGo have some of the better AI summaries in my experience.
They give a lot more detail and you can actually ask it to generate even more.
So it kind of does it in the browser in a private tab. So that's kind of cool. But yeah, I often go to an AI search first to give me kind of a comprehensive and you know, I may also do some manual searching to fill in gaps. I'm not saying I never use Google search, I just, it's not depend. Given the kinds of problems that I have to solve, I just find it some more effective.
[00:29:56] Speaker B: Yeah, yeah, no, for sure.
I mean I, I think the AI, I mean this is the thing like it, it does help you get oriented fast, but then you do gotta go and verify what matters. And I mean you've seen like it, it makes mistakes, right? And I mean it's, it's kind of like that. You, you know, you described it as a research assistant or as like an intern that doesn't sleep, but it also swears it's never wrong. So you gotta go.
[00:30:23] Speaker A: Yes, it's, it's very compelling. So one of the ways that I, is that I, you know, my instructions for Anthropic and Anthropic has some pretty good instructions. It's, it's to chatgpt but like if you go into Settings for instance, I have, I have like custom instructions of what I've asked it to do.
I've also told it, you know, to give me very blunt writing that's not congratulatory or anything like that. So it tends to follow my instructions of just being kind of cut into the chase a lot better.
One of the things I also found that's helpful is that I made a, I mean, I guess in ChatGPT you would make a custom GPT with instructions and you could upload documents to it. In Claude it's called Projects. And so one of the projects that I have is Tech News Helper. And so what I've done is that there's no documents uploaded to it. But what it is is that there's custom instructions with a big list of, of the tech blogs that I consider to be the most authoritative.
And essentially when I create a new chat using that project, the instructions are to, for whatever I'M looking for.
Stick to those sources only and bring me back articles that those particular sources have written about. Because those tech blogs have real journalists who do real background research.
I do, you know, and I say, you know, if there are other equally good news stories from, you know, New York Times or other places, also bring those in. But the core instruction set is to focus on, you know, Engadget, the Verge, Ars Technica, and the places that I know are going to give the deets.
So I've almost created my own search engine, like a tech blog search engine, and I find that to be very helpful because in the old days you'd have to go to Google and then, like, you could use, I guess, Google search to search the particular domain using the site colon function. But that's a very tedious exercise.
[00:32:35] Speaker B: Yeah, yeah, exactly.
[00:32:37] Speaker A: I mean, do you do it in that way or do you take a different approach?
[00:32:42] Speaker B: Well, I mean, I, again, like, for myself, like, I think what I would describe it, it's like with the AI, you're outsourcing the, the starting point, but I'm not really outsourcing the thinking.
So at the end of it, you know, some of the things that I'm asking is like, what's true first? And then I look at the, the competing stories.
Yeah. So again, it's just a matter of figuring out what is like the, it takes what I'm curious about and then it's, it's putting together like maybe the, the five things I need to go and confirm.
[00:33:24] Speaker A: Yeah, yeah.
We were also going to talk about just automation and productivity in general. So just like even expanding away from search a little bit. Are there particular things that you continue to do or new things that you've been using AI for?
[00:33:41] Speaker B: Well, you know, one thing that I've been doing is if, let's say if I'm copying and pasting something, especially like more than twice I've been, I, I take that and then I turn it into a prompt. And so that's one thing. Like, I have like a list of prompts that I, I have. And so I think for me, I don't know if it's the goal is automation, but it's.
I'm trying to go and reduce the amount of touch points and then, you know, creating fewer workflows. So.
And this. Yeah, this technology kind of helps me with that.
[00:34:19] Speaker A: Okay.
I find that I take a lot of notes, I write a lot of notes, pen and paper, I type a lot of notes.
Basically, it's eliminated all time. I spend reformatting Things that I know I'm going to return to. So if I'm documenting processes for something I have to do at work, I just take notes in whatever order that comes to mind because then I'm more focused on getting things down.
Similar to health practitioners who've been like, look, I use AI, health based AI for transcription when we're having a discussion in a medical setting. So I can focus on you rather than trying to take notes while you're talking.
Kind of similar to that, right? Like, it's like people are giving me information. I'm just typing it out, I'm asking questions, I'm providing details. It's a mess.
It's not an inordinate amount of work to go and you know, reformat notes, put in bold keywords, headings, especially for. Again, I take that time to craft notes that I'm going to return to. They serve as reference for me, especially if I'm developing like an email template to something where I'm going to respond to people over and over. There's all sorts of examples, examples of this.
I don't have to do that anymore.
I'll just dump that into Claude and be like, remove all personal identifiers. Obviously, I'm very careful about that.
And it will just reformat those things. And especially if I tell it like, this is a template, this is going to be something that I'm going to return to. These are instructions that I'm going to need to reference later. It does a terrific job of just formatting it in a way, even suggesting keywords for tagging if I wanted to search them in Apple Notes or something like that. Like, it just, it's incredible. So I just don't have to do any of that formatting anymore. And it sticks to my prose. Like it'll fill in words and details, but it looks at my writing style because it's in my custom instructions like stick to what I would write. And it does a very good job of writing, expanding on what I would have written had I done it manually.
[00:36:21] Speaker B: No, that's a good point. I mean, I've done that where I was.
So I mean, I'm taking some PhD courses right now. So I, I, I just write up my notes and then one of the things that I've been able to do was take my notes and put it into and format it as a table so that I can go and compare the various, you know, concepts or, you know, theories or what have you at a quick glance.
And so it's just, you know, that would have for me to go do that, it would have taken a lot longer for sure.
[00:36:53] Speaker A: And like, one of the things I've done. I don't know if this is a good idea. I'm a little skeptical of Claude Cowork, where you give it total computer access.
But I have allowed it to connect to Apple Notes. And one of the things I've been, because I have hundreds and hundreds of Apple Notes and a lot of them are like professional development, you know, personal development stuff. And, you know, I've asked it to kind of comb through and look at some of these things. Are there redundant notes? Are there things that I can compile together?
Can you bring me a list of notes that seem like maybe they're unfinished ideas so I can go through an archive, delete, whatever? And it does, and it gives me a list and it says, hey, these are a bunch of notes on a similar topic. You could combine this into one thing and I could reformat it for you.
And so it's helpful in that regard. You can do the same thing with Notion or connect it to OneNote if you want. I think ChatGPT will do the same.
And I find that helpful because we gather these kind of snippets of advice as we go through life.
And what good is this stuff if you can't find it?
So to be able to query your own database. Remember years ago we talked about the Zettelkasten and notes and that sociologist who developed that kind of interconnected web. I love that idea.
I love that idea of being able to take notes and put it together and being able to.
But the work of managing an interconnected web by hand is very difficult and it's a huge opportunity cost because you have to go in and do this kind of personal knowledge management with all the stuff that you may or may not ever return to. So this idea that you would have a better search that you could query stuff, I mean, you still want it to be organized because I don't query everything, but just that idea that I can go in and, and pull in things. Hey, I want to write about this. And I'm looking for notes. I've taken on books that cover this topic. Can you bring back all of that? And I will do it.
To me, that made all of this useful again, all this untapped knowledge. I heard years ago there was a great quote from an entrepreneur that I had met, and he said, don't take this literally, but everybody has $100,000 worth of ideas in their Google Drive. And. And that's kind of A general. He said, and that's kind of a general argument for all the things that we save that we've forgotten about. And I wonder if that's kind of one of the future uses of AI, especially as we get into the Voice era, this idea that it can uncover all of these maybe potentially very useful things that I have filed away that I would never think to look for again.
[00:39:29] Speaker B: That's a really good application, for sure.
[00:39:32] Speaker A: I'm trying to think of other things that I've done in terms of productivity.
You might like this one. I mean, I realize I'm going on and on, but for teaching tools, and I showed you some examples today.
Working with students next week, or I guess when people hear this, it will have been the week prior that you're listening to, because I think we're depending on when we release these Wednesdays, ideally, but sometimes Friday, showing them how to take a research question and do a mind map. And a lot of this stuff I like to do manually. I like to develop my own questions and brainstorm keywords because it's more analogous to what students would do.
But I was with Claude and I said, I'm going to make a mind map out of this research question. I need to break it into themes. But I was tight on time, so I was like, can you help me generate keywords from this research question that would be useful for demonstrating to students how you would translate this into a search for an academic database? And that's something that an AI doesn't do very well. Academic searching in databases is something that there's a lot of advantages to doing it manually, at least at this point.
But what it took was from that query because it was a bit sloppy. My prompt was like, build a mind map from scratch. I mean, I was just going to go into Miro, open a template and start building the keywords. But what it did is that it Vibe coded a whole HTML mind map and it looked really great. And it kind of did a very different take on a mind map that I never thought of. And I just thought, wow, I don't know that that's what I would have built, but that's a great example. So, I mean, I even just used that for my class. I took a screenshot of it. I'm going to make the HTML local HTML document, which would open in a browser with this coded mind map. And hey, I Vibe coded this with Claude and all transparency, but ephemeral teaching materials. We need to generate examples.
I took all these notes. I have all this documents, diagrams, things I've compiled for these capstone courses that I work with students in.
This is my content. There's no personal information.
I gave it to Claude and it built out an entire local webpage with my own wording, my own content, everything. And one of the things I'm just thinking is generating kind of high quality HTML based graphics for teaching and learning. Examples like these are things that I would have to go use a tool to build to go into PowerPoint or get Vector images and put things together. But now I can have it actually vibe code things in HTML that look really professional and these are things that I'll use once for an hour and a half that I would have had to spend two hours to develop and.
[00:42:13] Speaker B: Maybe even longer, who knows. Right. But yeah, that was pretty impressive. I mean I saw the mind map and also the HTML page with the.
[00:42:21] Speaker A: Guide like a color coded mind map with a legend and everything. Right. Like remarkable and agree.
And it came up with examples like I was just going to do a mind map of keywords and ideas from the question but it even made sections of the mind map of like here are organizations that might have answers. Here are the types of documents you might want to look at outside academic literature. It did a version of a mind map that I wasn't even thinking about.
[00:42:49] Speaker B: Yeah but I mean you make a good point especially when you need to go and make like, you know, I mean this is like in the education sector but I mean I think even in some sort of workplace scenario if you need to go and show an example it can generate pretty quickly, right?
[00:43:06] Speaker A: Yeah. If it generates multiple. You can kind of edit it and you can take from with it. It's a great brainstorming tool. Do you do stuff like that with it? Have you found used it for things like that?
[00:43:15] Speaker B: Well, I mean and especially like for from a teaching standpoint, I mean to go in. I try to mix it up with and when you're teaching various sections and just to see what happens.
But yeah, I mean I'm. I'm kind of less impressed with the answers and it's like more what I'm trying to get them to look at is how did you get here? And then where you have to go and actually look at the evidence you have to take a look at.
It's more about the sense making sure. Right. And so questioning it and, and so on. I find it interesting too like sometimes especially when you're teaching multiple sections, the AI starts getting better.
And so that's, that's another reason why I'm Trying to go and, you know, use different models or even just different examples just to see what would come out.
[00:44:14] Speaker A: Yeah, I mean, it's incredible.
I'm wondering if the future of some of these tools for me is just how they connect to other tools. And you had mentioned a little while ago about AI is kind of like a wrapper or like an integrated features, another way to kind of flip that around. I wonder if that's really the future. It kind of unleashes potential in tools that we use for a long time, but kind of got stale.
[00:44:37] Speaker B: Yeah. Well, and. And this is where now there are some developments where people are going. And I mean, again, I think you should be careful with some of the ones that I've seen. People have been talking about maybe taking a. A small computer, like a, you know, a Mac Mini or something, and installing some of these agents to go in, and then maybe take like a dummy account that you don't really care about to make some of those automated decisions. But, yeah, I mean, it's interesting how the things are developing for sure.
[00:45:10] Speaker A: Yeah. Yeah. It's just incredible.
So, I mean, is there anything else that you'd like to add in terms of automation or search preferences from your perspective?
[00:45:20] Speaker B: I don't think so. Yeah, I mean, it's just, again, it's looking at a different workflow and I think there's opportunities there where, you know, you're not. It's not the automatic default of just going and, you know, Googling or even like with these.
There's all these apps. I mean, what I'm trying to do is use fewer apps. You know, it's not about finding better apps or even.
It's. It's. Yeah, it's in terms of the, the mastery, having fewer touch points.
I. I don't know if it's even about more productivity. I just want less friction.
[00:46:00] Speaker A: The capture. That's why we need that ring that has the audio from pebble that records everything. What do you think?
[00:46:08] Speaker B: Yeah. Yeah.
[00:46:09] Speaker A: Are we gonna be ring buddies?
[00:46:12] Speaker B: Maybe.
[00:46:15] Speaker A: I just think it's funny.
Yeah. I'm still thinking about it, but I feel like. Like it's not necessary. It's just like another thing I'd have to manage.
[00:46:25] Speaker B: Yeah, exactly.
[00:46:26] Speaker A: But I. I do want it.
I think you should buy it for.
[00:46:30] Speaker B: Me if it makes you happy.
So. Yeah. Yeah.
Again, I. Yeah. I mean, even if I look at it like it's like, you know, it isn't that even AI is killing search right now. It's just reordering, you know, the whole process.
So how you're getting oriented, the workflow, sorting, scanning, synthesizing, verifying. I mean, it's just. Yeah. Different approach now.
[00:46:59] Speaker A: Yeah, it's very interesting.
Well, that is probably a good place to end it.
This was a successful kind of recap on what's happening with AI and search.
So until next time. Well. Well, next time we'll get to. I think we maybe have another news and then another interview.
We have a few things lined up.
[00:47:20] Speaker B: Yeah. Yeah, for sure.
[00:47:21] Speaker A: Okay. Take care, Chris.
[00:47:23] Speaker B: Yeah, you too.
[00:47:24] Speaker A: By.
