Hey, folks, and welcome to another episode of the Creator Toolbox. This is the show all about the nuts and bolts behind your creator business. I'm Colin Gray from thepodcasthost.com Join me, Jacob. How you getting on, Jacob? I'm good, I'm good. I'm in another country once again enjoying the world. True digital nomading this last year, Jacob. I know it's good. Never one place I know I'll get. Too used to it and I'll never go back home for more than a month.
Yeah, yeah. Well, you're avoiding. You were just telling me avoiding a nice tropical cyclone, which is not exactly the weather you usually hear about in the uk. It's coming up very soon. Is that right? Yeah, yeah, for sure. We're going to avoid that. But good luck to you and everyone in the uk. Thanks. Thanks. That's nice. Good. Well, it's nice and blue skies today, so I'm not sure I even believe you. Never mind. Fair enough. Fake news,
right? We've got a few things coming up today, so we've got. I've got a few topics I wanted to dig into a little bit more of my work around, try to figure out. Shorts, short video. We talked briefly about this a few episodes back when I went into Filmora and a few of the other video tools I'm looking at. But I've got a couple of things I'm thinking about now. Productivity update as well. I talked about that in the annual planning one
back in. Was that the start of the year or the end of last year? I think it was the start of this one, wasn't it? Changing up the way I manage my daily and weekly tasks as well. And then we've got organic search as well to look into, don't we? You pointed out a graph just earlier on today that's absolutely mental. So we shall have a look at that too. What have you got in your menu? I've got a few things I wanted to just touch on.
The threat to my New Year prediction was that there'll be no significant progress in AI this year. So there's the $500 billion Stargate project I want you to just touch on and pull apart a little bit. And also another AI thing, just because that's kind of where my head is at the moment. Deepseek R1, which is a model you might not have heard of, but it might actually be very impactful. Oh, nice. Yeah. Okay, cool. That's good because. Yeah, I've been a bit skeptical about that
too. I thought that was a good prediction that there won't actually be. It'll be a bit of a flatline through this year, maybe into a wee bit further, because I don't know, I've not seen half as much development in this area, so. Yeah, that'd be cool to hear about. Right, I've got three, so let me jump into one of my little ones first. So we productivity update. Then I talked about the fact in a previous episode that I have been struggling a bit with asana, so trying to manage my daily weekly
tasks. And it's because we use asana to manage our team tasks and it's great for that, actually. It's really good for collaboration, it's really good for working all together and managing all the stuff we do, but it's just actually not that well suited for just individual tasks, especially for things like being able to just log ideas and have spaces to store it. And, you know, to be honest, there probably is a way to do it better,
but I don't think it's a perfect place for it. And I really remember using Todoist in the past and really enjoying that. So I thought I'd have a look around and see if there's any modern recommendations, anything better. And I came across something called Ticktick, which I've been recommended very heavily. Have you heard of this one? I've not heard of it. I'm just looking at the page just now. It looks very like Todoist. It's very similar, to be honest, it's very similar.
Very similar. The one benefit I think with Ticktick is that it's very much more tied into your calendar as well. So you can bring in your calendar events quite nicely. But. Yeah, so I won't go to. I suppose the main thing here is just to kind of tell you how I've set it up. So over the last week, what I've done is I've transferred all of the
kind of mess. It did take a bit of time investment. I had to sit down and I spent maybe a whole morning, I think maybe three, four hours going through all of the mess of my asana, the backlog I've got in there, all of the tasks I've got in there that are currently kind of active and bringing them over into TikTok. And that was almost a review of what should be still active at the moment anyway. So I kind of cleared out maybe half of the stuff that was in
there. That was all. One of my struggles is whenever an idea comes up, I tend to put it into My calendar, I tend to put it into my task manager because I don't really have a better place or I didn't have a better place to put it. So that was to review it essentially, but it just ended up with too many things in there. Overwhelm carrying stuff over from week to week. So now I've moved it all over.
I've cleared out a whole bunch of stuff that's kind of redundant, kind of out of date and in ticktack now what I've got is I've got four main areas. I've got my inbox, I've got my backlog and I've got my scheduled and then I've got my calendar as well. So calendar events I do think are important. That's one of the things I think is missing in a lot of to dos, like the fact that you've. I've got three calls today and I need to actually plan around that. I can't do as many
things today because of that. So having a calendar view in there and the fact that Ticktick splits it out as well, you actually look at the calendar view and it shows your task list at the top, but it also shows all the calendar events below that separately. So it kind of breaks them out in a really nice way, which I like. So now what I've got is. And one of the things I mean I like most about Ticktick todoist had this as well, was the fact that you can get a really good filtered view of both
today, this week, this month. So you can plan out. We plan in cycles. So you can do two months, I can filter for that as well, but I can look at the week and I can plan out my week and make it much more realistic. Don't see all of the backlog stuff, don't see all of the ideas that I've still to process. This is just the stuff that's really, truly active that I've planned to do and it separates it out and I can look at.
And then I can break it down just to today and I can only see the four things that I've got active for today. Nice. And the key thing, I don't know, the thing that I have found with myself certainly over the last few years is that I need to have. It's the difference between having this kind of bank of ideas, this bank of stuff that you plan to do, you certainly do plan to do. Maybe some of them are just ideas
that are pie in the sky, maybe you won't do them. But then there are Things that I do plan to do this cycle, but I just haven't scheduled them in yet. But I do definitely have to do them. I can tag those ones as like, definitely do, but certainly they're all in a different box and you don't have a date assigned to them. And if they don't have a date assigned to them, it means that they're in this other box. They're either
the inbox or your backlog. The backlog being the things that I do plan to do, Inbox being just those kind of pie in the sky ideas that I need to process and think, is this going to go in the back burner or is it actually going to be scheduled in and have them separated out from the stuff in the calendar? Because as soon as you have dates on things, as soon as it's in the calendar, that's when I get overwhelmed with things.
So that's the way I'm thinking about it just now. And it's actually relatively simple, but it's kind of the. It's a wee bit of the getting things done GTD methodology. Like whenever ideas come in, they have to go into the inbox. And then every week I have to have this routine where I look at the inbox and I process everything that's in there. One of the things as well I think is really important is that it
almost stops you. The rule should be if you have an idea, if you have something new that comes up, it should go into the inbox and it never should be scheduled straight away because you've already scheduled your week, you should have planned your week on Monday and the things
in your week should fill up your week pretty much. So nothing should really make its way straight onto your calendar from coming in from external, whether it's an email, whether it's a request from you, whether it's something like that, it never really should jump straight onto your calendar, always should go into the inbox and only be processed when you get through to the next Monday. Obviously in real life that doesn't quite work. There's always
these fires that do need done straight away. But the key is that if you do put it onto the calendar, you really need to think, right, okay, I actually need to take something else off in that case. So if I'm adding a date to something and I'm midweek, I need to take something else off the calendar. So it makes you think about priorities, it makes you think about what's this going to replace actually, rather than add to my list. So that's Where I'm thinking about it just now. Yeah, that's really
interesting. I'd never. I've never been one to put my ideas into my to do list, but the way that you're talking about it does kind of make sense, to be honest. Yeah, I tend to just put it. Whatever. Whatever the plan is, I guess, rather than the to do. To do is always my like, it's definitely getting done thing. Yeah, that's the. I think this is a real. This is a real struggle with people that do the kind of work we do, whereby we have. Especially when you work with content,
you're a creator of any sort. You're. There's always so many things you can do. There's always like, oh, I can add in this little LinkedIn tactic. I can go and do this thing over on YouTube. I want to try this method of making a video, this way of promoting my podcast. There's so many things that we consume and get as ideas and they go into notion and they go into a plan somewhere, but it's the act of going back and looking at them again is the struggle, isn't it?
Yeah, they just kind of get lost. Many times if it's in some other random place. Yeah, you just don't go back. Yeah, I think that's probably. That's why I put it in. It has to be somewhere where you're always looking. So like, for me, if something comes up mid cycle, it goes into the bottom of the cycle plan, you know, because I know that's somewhere I'm going to go back to. And to be honest, like, I've tried to use
different to do lists for different things. Like, I know you used to use todoist and you would almost like copy the things out of asana into the todoist and that'd be your. Your immediate to do. Right. I tried to do that as well because it kind of made sense to me and it just never really worked. Yeah, the more, the more places, the more complexity I try to add to that, the less, the less it works. Yeah, that's totally true. Yeah,
I totally get that. I actually, I like that a lot as well. I remember doing that. I have done that and I still do with our cycle plans. So our cycle plans are two months long. So if there's something big that doesn't fit into this cycle, then we put it in the future. And yeah, when we go to review that cycle, that's almost the inbox for that particular planning type. Like when we're. Cycle planning isn't it that little future box at the bottom of our cycle plan is
like, oh, right, here's what we need to look at. Exactly. And sometimes it pushes forward even to the next cycle, but sometimes it doesn't. But it gets recorded. That's the main thing. Yeah, exactly. And key recorded in a place that we know we're going to review because we know we look at that every single two months. I think that is the base core thing. Yeah, yeah, exactly. It's scheduled in there to be looked at. There needs
to be a routine. There needs to be a routine around anywhere you log ideas, anywhere you log anything that has to be you want to action at some point in the future. It has to be in a place that you have a routine around reviewing. So, yeah, that's my process is going to be look at my inbox every single week, process that. The backlog I'm going to have a look at every month,
or perhaps even that might be my cycle planning. I'll just go back and have a look through the backlog that'll be part of my cycle review. And that's where I look at all those kind of ideas that are waiting for a good time, but also taking the initiative to get rid of stuff like things don't have to remain in there, like being a little bit brittle. And I did this with the asana stuff like saying, do you know, that is a nice idea, but it's just like,
I don't know if it's really suited to us. I can't see a place where we're going to do that in the next six months. And do you know what? If it's a good idea, it'll come back again. And I've heard, you know, the 99, not 99 designs, 37 signals. Am I getting that right? Yes. Yes. Basecamp, have you heard Jason Fried's kind of thoughts on this? No. But his principle is that they don't do backlogs, they don't do long term plans in that way. They don't kind of log things in that sense.
Because his principle is that good ideas keep resurfacing and by the time they're worth doing, they'll be resurfacing so much that you almost can't ignore it, which I'm not sure 100% agree with, but I like the principle. Yeah, I mean, I guess it's a good way to stay focused and actually really sieve out kind of naturally what's worth looking at and what's not. I guess you are going to lose things down the cracks, aren't you? But I mean, I do kind of agree with the principle. If it's
not worth losing, you would have remembered. Yeah, yeah, it'll just keep coming up. Yeah, totally. Great. Okay. Well, that's my productivity update though. That's how I'm using Tech Tech, my new system. So I'll bore people by updating on that over the next few months. Very nice. Just quickly though, before we move on, what is the difference between that and todoist? What. What's stopping you from to do it? To be honest, very little. It was just. It was.
It was really just my. My novelty addiction. I just wanted to try something new rather and go back to someone old. And it felt like there was a couple of little things in there that intrigued me enough to make me have a look at it. The calendar view, for example. Todoist, you know, it was two years ago or so that I used it. It might do exactly the same thing now. And I do believe todoist is better if
you're working with other people. So if you are wanting to do some task sharing, collaboration, that kind of stuff, if you have two people all using it, todoist is supposedly better for that. But it was literally just. I wanted to try a new tool to see if it was better, just to give it a shot. Really. That's pretty cool. Yeah. Yeah, Nice. I want to try. There's one I want to try. I might try it and come back
and talk about it. Called Motion. Oh yeah. I keep looking at it and keep thinking, well, trying to figure out whether it's a good idea or not. Basically you give it all your tasks and you connect it to your calendar and it schedules in the work for you. Sounds possibly a bit overbearing, but it gets advertised so much that I'm interested. AI powered,
type of. Is that what you mean, like scheduling? Yeah, totally. I mean, before the AI hype started, I don't think they called themselves an AI project management app, but now they do. So I do question how much AI is actually going on in the background. But regardless, it does look interesting. It's worth having a wee look at. Yeah, perfect. Okay, cool. Let's jump to one of yours. So what is Stargate? Are we about to start moving to different star systems? Well, yeah, first of all,
cool name. Fair enough. I'll give points for the name and the reason I wanted to talk about this, because if they did manage to somehow pull this off, it will impact us all. It will impact us all as creators and it will most Importantly, it will completely ruin my prediction for this year that there's going to be no significant development. And that'll be very sad for me, very embarrassing. But I mean, essentially it's the Manhattan Project for AI is
the way that they're pitching this. They're trying to get the best minds together, they're trying to get all of the money together and they're going to build, I think, five massive, massive data centers that are bigger than any existing data centers at the moment. And they're just going to throw. One interesting point I saw is with what data? Right, because you've got OpenAI,
scrape the Internet, fine, they've got that data. The data that you need to make super advanced models to really advance AI needs to come from somewhere. So that's one big question. But anyway, I think, yeah, a lot of people are talking about this, especially on social media, especially among more techy circles, and I just have questions and it's one, where is the money going to come from for this? $500 billion is more money than it's ever been spent
on anything ever really, probably apart from the Manhattan Project. You know what I mean? You look at the names that are involved with it. You know, SoftBank, you know, you watched the, well, you watched the crash. They were involved in WeWork, they were involved in Theranos, which was another. Well, WeWork wasn't really a scam, it was just. Yeah, but Theranos really was, and they were involved with that. Oracle is in massive, massive debt. I mean, I've written down
the numbers here, but I don't necessarily need to go through them. They are in debt. Same with SoftBank. They don't have a lot of operating cash. In the tens of billions. Yes, tens of billions. Yeah. OpenAI is living off investment. Nvidia is meant to be involved and they have all sorts of accounting issues at the moment and just very, very skeptical
about. I made a comment last year that I think you raised your eyebrows at and I keep thinking about it because it was so kind off color for me and I said, I think people are too optimistic at the moment, you know, and for, for someone that's generally really optimistic about like the way, the way, the kind of direction that technology is going and what that might look like and have to be optimistic to not see the, the, the, the dark side of
it and what could happen. I don't know, I just worry that, I wonder what this is going to do without anything happening. How is this promise of $500 billion being invested into super advanced AGI, we're talking about past human level intelligence is what they want to do. How is that going to actually affect the way that people make content? How's that going to affect where money flows? Like, how is that going to affect like our industry, podcasting and all the VC
money that sloshes around in here. And yeah, what's the real benefit at the end? Yeah, I know it's a really big, it's a really big statement to make. A really big sort of, I don't know, dangerous promise to make if they're not going to follow through on it. And I guess, yeah, I'm just, I'd like to, like to know what you think. I mean, so the, so 500 billion, it's quite a big collaboration between a whole bunch of different. Is it one company I don't like?
What's the organization about? Yeah, Project Stargate is to become a private company and I think it's to be essentially a roundtable of various tech companies funded by, you know, those companies that I mentioned in the main. I think there's some other investment bank that I'd never heard of. And yeah, I mean on paper let's get all of the best minds and all of the money, but I
don't know. I think there's already been like billions and billions of money churned into AI, probably with like minimal improvement. We got that. I mean we talked about it a few times. Like there was that magical moment the first time you used ChatGPT and realized that, oh, this is actually amazing. That was like a leap, that was a chasm, it was a mountain of difference. It was crazy. But now you have to do a lot of work to tell the difference between the models, don't you? I know you
do that work. You have told me you can tell the difference between Claude, ChatGPT or even in ChatGPT, the difference between the different models in there. I don't really notice the difference, honestly. I just ask the questions. I use the latest model and if they put me back on a previous model, which they do sometimes because they rate limit you, don't they? If it's too busy, like I only notice later on
potentially. So it's like all the money that's been invested in this spent on this for real improvement and really it's just still half the use of it is make a bedtime story for my kid or like my brother, actually one of his son's favorite activities now is just going on and generating pictures and it's always just crazy pictures and the prompt is make it cooler, make it shinier, make it like that is.
So he's generating 10, 15, 20 pictures, spending all that computing resource just to make a robot with like teddy guns more. Cool. Yeah, yeah. Like, I don't know. Yeah, it's the last mile that's going to cost all of the resources that we have on Earth, essentially. I mean, we're not just talking about money here though. Colin, Someone very helpfully worked out the power consumption requirement of this and I've just lifted this from them. So five Stargates, they'll need at least 6 gigawatt
each at all times. This works out to about $4 billion per year per Stargate. That's already a massive amount of money. And not to, not to even speak about the environmental impact. Yeah, yeah. For reference, Microsoft has some of the biggest data centers in the world. They pay $100 per megawatt hour. But that is, you know how they power that? They power that by. They restarted the Three Mile island nuclear power plant project, an entire nuclear power
plant going towards this data center. So actually Stargate's going to need five of these. They're going to need to build more. They're actually going to need to build nuclear power plants just to fund this or just to, just to power this. Rather if you were to take out the current fleet, you would be taking 5% of America's current nuclear power generation just offline and just dedicating it to that. You're talking like, like entire
cities worth of, worth of energy and it's for that last mile. And like the last mile is good. So I think you're right. And the models all feel quite samey at the moment. And it's just people using it for basic things where
they want to get to. And the reason I thought it was worth talking about is what does it look like if they actually do it, if they do get to that AGI point where we're not talking about something that can answer questions very well, very convincingly and can be helpful with code and all of these things where they want to get to a place where we have an artificial
intelligence that can invent things and solve problems. You can give it the problem of, of curing cancer, you know, of, you know, like actually applying it to hard science rather than just the sort of text based, almost just content generation applications that we have today. What does a world like that look like where computers can think that deeply if we ever get there and God, yeah. What does that mean for us as creators? I don't know we'll have to make better content and evolve,
but I do. Yeah, yeah, it does give me that sort of itchy feeling feet of we do need to evolve as creators. And it's just not clear exactly what that looks like yet because we can't really compete by writing more articles which. We'Ll get to as soon as you have. Yeah, exactly. As soon as you have a system which can basically, I mean you can already as a human, you can generate a fake influencer,
you can generate fake content for them. You can kind of run Instagram channel or a YouTube channel completely not with yourself like just auto generated the entire thing. But as soon as somebody else, as soon as a system can do that, what does YouTube look like? Suddenly it's going to be all right. You need some id. You need
some like genuine proof of humanness to register a new channel. But even then, like if you're an artificial general intelligence that can solve problems, invent shit, then why can't you just figure out how to fake an ID? Yeah, yeah, 100%. Well, I, Colin, I, I got Claude 3.5 not haiku sonnet with, you know the computer use thing that we talked about pre film another show. I got it to complete a captcha which blew my mind. You know the whole point of that is to, is to distinguish between
computers and humans. And this wasn't me like spending days trying to cleverly do it. No, I just did it. Yeah, just say complete this please. Yeah, I gave it a task and that was in the way and it just solved. Do you know what I mean? Yeah, it's definitely not, it's. There's definitely a lot of negative. I think the energy stuff is just mental. I don't know how. It's just kind of a race that people. It's like it's an arms race exactly the same as the Manhattan Project.
And a lot of governments think that they have to spend this money and burn this energy because other governments will get further ahead and just all that. It's just horrible. Yeah, I don't know, I feel like right now my kind of end point desire for AI is where it is more the agentic type approach where it can actually go and complete a few tasks. For me. Like I would be satisfied if it stopped at the point, well okay,
I'll finish this. Satisfied if it stopped at a point where I could use it to book a trip, a holiday and it could actually go and book the flights. It could find good hotels, it could, you know, that kind of thing where you can say, right, here's what I'm doing, please go and do this for me. And it's the kind of thing you would give to a, a kind of an assistant, potentially a human assistant. It's kind of grunt admin type work.
If they can go and then actually complete stuff like that, that would be like genuinely game changing for so many businesses because suddenly they could get so much more done. They could so many barriers getting out of the way and I don't think that makes people lose jobs either. I think assistants do still exist. They just help the. They just ask for more tasks, they think of more cool things to do because then it's the idea generation type of thing.
But beyond that. But the thing that made me pause halfway through that was when you say things like curing cancer and solving the climate problem and things like that. I suppose if you do get an AGI that can genuinely do that in five to 10 years, that is a huge benefit to humanity. Totally is. But it's going to be so expensive to run. It's only going to be the Silicon Valley tech bros that can afford to run it. So what are they going to use it for? They fully control the output.
They fully control the input. Yeah, exactly. It's all right anyway, but kind of along those lines, it's maybe worth us chatting about the chart that I shared earlier today, because part of that narrative is that that's been an AI involvement there, that people are just getting answers elsewhere. Yeah. So we looked at a chart earlier on today, which is HubSpot's traffic over the last year. Was it? So tell me what this looks like, Jacob. It looks like.
It looks like a poorly engineered Boeing aircraft crashing into the ground. Very, very ugly. It's horrible. And HubSpot are, I mean, say what you want about their software, but let's be honest, of all of the software companies in the world, they have probably the strongest SEO team. They have invested so much money into it, they wrote the playbook on inbound content and if the reaper can come for them, none of us are safe.
Yeah, it's a weird graph as well, because it doesn't even necessarily follow the trend that we've seen over the last couple of years around the helpful content update and stuff. Because it's got, if you're watching on video, I've got it up in front of me, but I'll describe it for audio. Obviously their rise started 2016 into 2018 and then really took off around 2018, going from something in the region of 4 million per year,
actually that's per month. 4 million views per month in 2018 up to a kind of start and peak in 2020 of 16 million. So four times over two years potentially. But then they had a big peak in 22 as well, where they were up to 23, 24 million. And then helpful content update kicks in 23. It was more September 23rd actually. So they. Maybe it's related anyway, but they had a bit of a drop 24 to 16. So I mean, that's a significant drop, about a third.
But it's more the last six months for them, isn't it? Yeah, yeah. Or even six months. It's actually more even the last two months. They've had a huge drop off, so who knows what that's about. But you're right. I wonder, I mean, you know their content way better than me. Were they good quality? Were they adding value? Were they unique? Like, were they helpful? Well, they have such a vast amount of content that it's a bit of both, like part of the sort of meme of people talking about
this graph at the moment. Yes, they were putting out content that was like 20 famous quotes from this person. You know, some of it was just programmatic. Like when you get to that scale of an operation, you have to start looking at some lesser value keywords to be writing for. Do you know what I mean? But they still, you know, I still found myself reading their content pretty often. They have good writers. They have really
like, they're. They kind of set the bar for the ultimate guides kind of pillar page for me, you know, and when they do go in on a topic, it was always pretty good. So I mean, it's hard to say exactly what has caused this, but I think certainly part of it is that a lot of the probably kind of quite technical marketers and content creators that were using this, they are fairly early adopters of AI. And I think that that's definitely had an impact. You know, it's probably been
part of the, part of the slip and slide. That's a good point. Yeah. But I don't think this can be explained entirely by like a voluntary switch. I think there's got to be some Google loss here as well, some SEO loss for that amount of change. Change. The argument that I've seen is that so many of the of the questions that they try to answer are getting answered by Google itself
and they're AI generated. So not necessarily actually into chat GDP or anything like that, but just the search engine itself is taking the traffic. That makes sense to me. Yeah, yeah. Quite straightforward, single kind of answer questions. Yeah, yeah. There's so many examples of this though as well. Like over the last year or so we've followed this pretty closely because we were pretty
lucky. Like we did have a slide. We lost I think something in the region of a quarter of our traffic around that area, 25% ish. But it stayed relatively steady. And then we had a bit of a decline over a year or so we had a bit of an increase again. So it was nothing in the realms of this kind of loss, but it made us very aware of it. Certainly one of the guys more comparable to us that talked about this a fair bit was Spencer.
Spencer Hawes. I've forgotten the name now, actually. Nichepursuits. Anyway, he runs a site called nichepursuits.com Forgive me if I've got that name wrong, but he was hit really hard. Was up to. They had traffic in the region of 300, 400,000. More like Ark in a region back in. Yeah, exactly. Per month. And then has just gradually, gradually dropped like pretty much almost a straight line when you kind of average it out.
And right now like semrush hrefs and stuff are showing him a traffic of about 30,000, so 10%. And as far as I was concerned, his content was pretty much exactly what Google always asks for, which is the eaat, like authoritative with a person. You've got this person who actually is known to be authoritative space this guy. I'm sure Spencer was. I should check that. But anyway, presuming it is Spencer, he knew what he was doing, he had social presence, he had a good audience
elsewhere as well. And it was like really helpful stuff. It was, I don't know, it was like good content written with personality and humanness, which was my impression of what the helpful content update was designed to tackle. So I didn't understand this at all. Totally. I don't actually see what's in it for Google as well. Just going back to the AI generated thing at the top. I don't see what's in that for them because they're not getting ad revenue from that. They're not
getting. They're getting engagement. But how are they making money from that? I don't know. Know. Yep. Don't know. Don't know. Yeah, it's crazy. I think it's worked really heavily against them. What do you think? I mean, what's the. Let's actually make this kind of practical. Like what are our plans? Like, I've certainly got a few things in my head around how
to try and future proof for this. So presuming that SEO is going to be more and more hit, that AI results are going to take over more and more search results, what does that mean for content creators and people selling products that they use content to sell like us, or people that just create content to educate like us as well? What are your thoughts in terms of what do you think we'll
see in the next year or two? I think we're going to see more of a conversation emerge around how do we optimize for AI answers. To be honest with you, I think that search isn't going anywhere, but maybe the way that we do search is going to change. Right. And I already see quite a bit of traffic increasing coming from places like Perplexity Store articles and they've got their own
sort of search API for that. They don't use Google, they use some proprietary ranking presumably to try and find out what the best answers are. So I think we're probably in the way that we've had to kind of figure out over the last 20 years or so how to optimize for a traditional search engine. We're probably going to need to figure out how do we optimize for AI. And I think my gut tells me that it's going to be a lot less sort of of editorial. It's going to
be a lot more fact based. It's going to be, we think about how these things are trained. They're trained on quite structured content like you might find on Wikipedia and then unstructured content like you might find on social media, on YouTube, on podcasts and that kind of thing. So I mean, I think a safe haven until we figure out what we do is still going to be podcasts, video, you know, these channels where we're mostly talking more about pleasing some other algorithm.
Sure. But it's not so much based on. I don't know, I don't know. That's the question, like, you know, how do we not the same risks exist, you know what I mean? Yeah, yeah, I think you're right though. I think there's a few things coming out around how to be optimized for that. I've been really
pleased to see you've pointed out a few as well. I've spotted a few recently around and ALITU showing up fairly well in AI search and I think that's because we've done the work over the last few years not only to write a lot of content ourselves about alitu, that's pretty Educational, pretty fact based. We have some editorial in there, we put some personality into our content, but it does answer questions. Our principle for content has always been, let's say Marcus Sherdo Jordan is a guy that
says this. His thing is they ask, you answer. So if there's questions out there that your audience are asking, you need to have answers for that in your content. And so our content has always been heavily built around questions, whether it's the title of the article or whether it's the sections within the articles. And I think that is exactly, my gut tells me that's exactly how you work towards being included in AI content. So our own content is certainly a base for it.
And I think everyone should still be writing their own content based on answering those questions. I don't think that is going away. Even if you're not showing up in Google Search, it needs to be there to help the AI to find those answers, to include in its own answers. But the other part of it is doing the work to actually have partnerships with other people in our space. So we've managed to get AI, Alatu and the podcast host content and all that kind of stuff mentioned in other websites
and as many other websites as we can. So other people that are answering these questions, you then want to be mentioned in there. And that's about making good product, it's about making good content, it's about making relationships, going to events and really getting to know these people and offering value and making it worth their while to include you in their stuff. So I think that's a big part of it. And the other interesting one I spotted come up was Reddit.
So Reddit obviously a big source of content for AI in general, a big source of content for SEO these days, isn't it? Like shows up two or three in Google pretty much every single time. So running things like running a Reddit channel for your product, for your content, for your audience, certainly a thing. Participating in other channels, certainly important as well. But one tactic being to run AMAs
as well. Well, quite often. So if you run an AMA for your audience or a different audience, if you've got the expertise, you've got a bit of reputation potentially to allow you to run something like this, then that AMA can turn into a really valuable bit of content on Reddit that then the AI picks up and starts taking answers from because it's got
so much interaction with loads of other people in there too. So that's something that I've seen has supposedly worked well for other people in terms of Getting their products mentioned and the right info mentioned in AI as well. So I've found that interesting. Yeah, that is really interesting. I mean, that's. I guess that's kind of part of a trend that's already been ongoing for a while now, right. The idea of an influencer over a sort of publisher.
Yeah, yeah, yeah. And I think that's probably, maybe that's the way that that's how humans find us and actually can persevere and continuing to write quality content that answers questions and helpfully, that's, you know, that's how we get recommended. I think there's going to be a teething, really, you know, and I think that probably the. The creators that persevere and keep creating that quality content, even if the.
The channels that they're using to get used to getting eyeballs from are in a weird way, you know, not to say that Google is going to disappear in the next few years or anything like that, but it's definitely going to change and the way that search works is definitely going to change. Yeah. And I think, yeah, there will have to be an element of that. We're going to have to persevere. I think probably a lot of people will give up when they see graphs like that HubSpot graph. It's very,
very scary. But, yeah, ultimately people are still going to need information from the Internet. Yeah. I think my impression of it over the next year as well is it's going to be that we've talked about the barbell as an analogy for a few things over the last 20 episodes, but I think it applies here in Tert as well, in terms of humanity, humanness in your content and that kind of
middle ground. We've often fallen into the middle ground where all of our articles for the podcast host or for anything else we write have been very factual and educational, but we also put personality into it as well. So it's human and educational at the same time. And I wonder if we're going to end up over the next year going much more towards the extreme ends and ignoring that middle. So it's the barbell, whereby one end is purely human content, where it's nothing but story
and speech. So media into that YouTube, podcasting. So it's just conversational. It's like pure human. You're not even so worried about the educational content because it's more just about having this human contact and hearing the stories and the human experience and all that kind of stuff, and then the other end is just the pure facts. So articles that are Much more concise and much more just. Here's exactly how to do this. Step by step. Let's dispense
with a fluff. And it kind of really pains me to think of that because I kind of hate to think of just writing stuff like that. And that is the stuff that AI can write as well. So that kind of worries me too. But yeah, I'm not sure. I think there's something in the fact that the substacks and the beehives have even more so taken off in the last year, I think because suddenly newsletters feel much more as if they're
from a human because they have a name and a picture on them. Them, they feel more trustworthy somehow and they feel more curated because you're selecting a particular person to then hear from every single week. And they are educational but you almost always get a bit of story, a bit of the person about the background, the context in there. But then they're a blog as well. You can go and just read
the substacks on the website. So there's something around that either end, which I think think we still need the factual stuff in one end, but we're probably gonna have to lean more into the human, the really extra human stuff on the other. Yeah, yeah, totally agree. Yeah. Hi. All right, well we'll see where we go with that. We'll try and keep up, keep updating. Okay, cool. Right, what have we got? So I'm gonna say we've got. So, yeah. What's Deep
Seek? Deepseek army one. Yeah. Do this real quick. And I think it's just an interesting thing to note. So I think one of the most exciting new models, AI models came out of OpenAI, which was 01, particularly their Pro model, which I've not tried because I'm not going to spend my money or your money on a $200 a month subscription for it. Deepseek have come out with a similar model. The reason that 01 is really interesting is that it thinks before it talks. Essentially it will reason first and
then it will give you a response. So it's almost like a two step thing. It's an extra layer in there. Deepseek is a Chinese company and actually you mentioned earlier on that sort of arms race of AI and very, very much this is a step in that direction, but very good for us as creators because it means that actually more advanced AI that can do more helpful things is going to become cheaper.
The big thing about this is that, well, two things. One, there's a version of it that's small enough that it can run on your phone locally without having to connect the cloud. And that is almost as good, pretty much. Well, actually, sorry, not almost as good in. In lots of benchmarks, it outperforms GPT4O and Claude 3.5. So that's one thing that's very interesting, but for much, much less money than it costs to run 01 for. For. For any sort of application. The sort of larger version of it
kind of. It's not quite 01, but it's 01 Mini. It still has that ability to think before it talks and it's just very, very impressive. And it costs a fraction of the amount. It's actually cheaper to run overall than cloud 3.5, which is crazy. Yeah. And it's sort of half open source as well, which means that there's probably going to be some more innovation coming out of this. There is absolutely no explanation as to why it's so cheap or so
small other than. And there's been an innovation in the way that they've built it. But, yeah, they're definitely making it cheap as a sort of. Yeah, a bit of. A. Bit of a war with OpenAI, which I think is a good thing, because OpenAI is getting a bit of control with cost. I think $200 a month is mental. Yeah. But they're still losing money on that. So actually, maybe it's just that they, you know, there wasn't enough innovation there in
creating that model and keeping the cost down. It's a massive, massive model. It does cost a lot to run. Maybe this pushes them to figure that out. You know, it kind of felt like they were trying to justify it rather than try and make it cheaper. Well, that would be a nice trend if we went towards efficiency rather than power for a little while. Exactly, exactly.
Cut down that crazy energy usage they're talking about. Like, what if we spent just two or three years, like, pausing in terms of capability and just making it far more cheap, far more easy to run, far more efficient. Yeah, totally. Yeah. So can you use this right now? This is available. You totally can, yeah. Lots of different places. I think Deepseek probably have a demo on their website that you can use. You can download the model yourself and you can run it locally on
your computer, if you're technically into that. Or. Yeah, you can just look for a demo online. Yeah, it's pretty cool. Perfect. All right. All right. I'm just going to do a quick update on my thinking around shorts. So I've been trying to figure out Shorts for the last little while, like how to do them more effectively. We talked a little bit about this. Like the fact
that I was. Did I mention the fact that I was taking these episodes and rather than AI clipping, I would actually take a section and just record a one minute, two minute clip just by myself. And I find that just so much more effective than the AI clipping method. But I do wonder, like, AI clipping I just don't think is effective yet. But you do see AI clips from other podcasts, from YouTube channels
doing really well. So I don't know, do you have a sense of what makes a good clip of that sort when it's taken from a larger space, a larger piece of content, what makes a good clip like that? Yeah, from what I've seen of good clips that are maybe AI generated, maybe not, but most good clips in general, they seem to come from interview shows where there is a setup of ask a question, get an answer, you know, and if you can get a short question with a short answer, it makes a good clip.
Whereas we're, the way that we do this show, it's a bit almost like I take a turn, you take a turn and we do, yeah, we do have a bit of a discussion. But yeah, maybe you're right. Maybe you're right. There's like more, especially if it's a much more conversational interview, if there's somebody's a good interviewer and they're making it much more back and forth, forth. That's a good point actually. Maybe that's what leads to better potential for AI clipping.
Because I find most interview shows you'll ask a question and it'll be quite a long question and then there'll be quite a long answer. And that's like, that's kind of a. When you're less good at interviewing, you build less rapport, all that kind of stuff. It's definitely more monology from both sides. So yeah, maybe that's a good point. So. Well, okay, so think about that a bit more. How to build that in.
The other thing I'm thinking is there's a lot of channels just now that I like like, which are just doing much more kind of point of view type videos, particularly around things like gear and software. And so I'm thinking about a setup where I can have my phone on a little stand with a nice light and be able to move it around really simply around my office and have a look at bits of gear and tour what's in my bag.
Things like that work really nicely, because people in our audience are really interested in the microphones, the mixers, the lighting setups, all that kind of stuff. So I'm thinking about a setup and I'm going to get some advice from you actually on this because you've got this a little bit for your filming for this show. Like what stand do you have for your phone just now? So one I'm using at the moment, because it's really light and folds up nicely, is actually the Joby gorillapod.
Yeah, I just. I can attach my microphone to the top of it. I've got my phone on it. But actually I bought something interesting just before I left. I've not had a chance to properly use it yet. I bought a DJI OSMO mobile style gimbal which I was gonna use because you can set it up on a tripod and it will actively track you around. So it'll actually move
the stand as you go. Yeah, yeah. Obviously as a gimbal, it's designed to sort of be a sort of steady cam type thing that you can walk around with. Yeah. Which you can. Yeah. Use for lots of different things. But yeah, I mean, just keeping it simple with this. Yeah, just something like a microphone on top with my phone. Yeah. I'm not fussed out the mic, actually, because this would just be. I'd probably do a voiceover over or I would do a lav mic on me if I'm going to hold it and speak
at the same time. But I think the idea with this would be to actually just record the visuals, go through my bag, show, demonstrate a bit of gear, show a bit of software even on the screen and then do the voiceover afterwards. And I tried one of these yesterday, just really simply, just recording a couple of things to show a way that I've mounted my lights on top of my monitor just with some action cam mounts. Just a really simple thing, but it's just a little kind of
hack. And I just filmed, filmed four or five little clips, maybe 30 seconds long each. I brought them into Filmora, I cut them down into just like specific little bits and then it just recorded a voiceover as I played it. And it actually only took me 20 minutes to do that. It didn't take long at all. Kept it really simple. Again, like. Like I said last time, used Filmora to do some really simple little filters to make the. So the lighting, because I wasn't that fussed about
the lighting when it was recorded. So I just made it, shined it up, essentially Instagrammed it up, I guess, with the filters and some trans. And I thought it came out really polished, actually. And with shorts, they're meant to be pretty raw in any case. I mean, you're watching it on a phone screen generally, so you don't exactly need a big, fancy, mirrorless camera for that. Yeah, cool. So I'm going to play with that. I'll come back with some more updates on that in the near
future. Yeah, cool. All right, Jacob, I think that'll do it for this week. Yeah. All right. Thank you very much. Yeah, you too. So if you're out there listening, one thing we'd love you to do right now is we'd love to get more people into the community, so we have a little space in our larger Indie Pod community. So it's a. A community of creators, podcasters primarily, but a lot of wider creators,
too. But we've got an Indie pod. We've got a creator toolbox space in there where you can give us some feedback, you can give us some questions to answer on the show, tell us what you want to cover on future shows, and we'd love to get some conversations going on in there. So we'll put that link in the show Notes. So if you click in your podcast listening app right now, you should see a link in there for the community. Just go and click that and you'll pop straight over to that.
If you want to use the URL, you can go to IndiePod, thepodcast host.com and you'll find it there. But we'd love to see you there. You'll get some control of the show. You can tell us what to do in the near future. We'll take over. Indeed. Indeed. All right, thanks very much, everyone. Thanks for listening and we'll talk to you next time.