82: Peak Bubbly  ## Metadata - Author: [[Mostly Technical]] - Full Title: 82: Peak Bubbly - Category: #podcasts - URL: https://share.snipd.com/episode/9310fbe8-1abc-4f0a-8708-e4c94978e341 ## Document Notes ## Highlights - **Ian's Strange Fever Experience** - Ian Landsman experienced bizarre fluctuating fevers and dangerously low body temperature readings. - Urgent care offered limited answers, highlighting frustrations with healthcare responsiveness for unusual symptoms. Transcript: Ian Landsman Ready. Energized. I love that for you. You were sick this week, right? Oh, man. Aaron Francis We're on the street. Ian Landsman Not energized on this side, I'll tell you. I had one of those weird week-long sicknesses. So last Saturday at the kids' soccer game, I was like, man, I don't feel good. You know, whatever, normal. Monday, I had 102 fever. So I was like, oh, this is not good. The rest of the week, kind of like low energy, not feeling great. And then I think it was either Thursday or Friday. This has been going on. I got home and I was freezing cold. Just like I cannot get warm. I'm wearing pants and a shirt and a sweater and a jacket inside the house. And I'm like dying. So I thought, all right, well, let me let me take my temperature, see if I have a fever. Took my temperature and it read 93 degrees. And I was like, good, that's dead. Right. And so I took it a few more times and it said 93. Wow. Man, this seems bad, right? Aaron Francis You should go to the doctor. Ian Landsman And so I got in the shower instead of going to the doctor you know because i was so cold yeah so i got in the shower and i was in the shower for like 45 minutes two shower heads just blasting Me with hot water because i was i felt like i was gonna die okay so then you know then after i get out of the shower i've got a fever of 102 So I get in bed with the chili pad on. And, you know, then I'm waking up in the middle of the night just drenched in sweat. Not great. So then Saturday, we have a birthday party for the kids. And I'm feeling fine. Go to the birthday party. Come home. It's like 630 and I start to get really cold. And so I take my temperature and it says 95. And I'm like, this is not, this isn't supposed to happen. Yeah. I'd never heard of anything like this. No. And so at that point I was like, all right, this seems, this seems pretty bad. I'm going to go to the urgent care. So I went to the urgent care. By the time I got there, my fever was 101. So I like all over the place and they were just like oh yeah you know it's probably a virus just take some ibuprofen it's like are you okay are you concerned at all about my cold temperature And she's like oh no that's probably not real that can't happen and i'm like but what if it did my inclination i have to say what if it did happen like what if i am really cold and she's like Well that would be bad but i just i did that's not compatible that doesn't make any sense and i was like i don't think i've never heard of somebody at 93 that's okay it's like you're dead I don't know i know and i felt I was shivering. I was freezing in my bones. Do you have a second thermometer? We do have a second thermometer. Did you test that? I didn't do the forehead thermometer when I was cold. So next time, if it happens again, I'm going to get a second reading because I'm tired of these doctors being like, this doesn't exist. I'm like, okay, well, explain then. Explain then why I'm shivering because I'm quite cold um so they did covid fever too you know yeah but when you know the other times the thermometer read high so like you tell me i'm not A doctor but i'm basically a doctor i have a thermometer you have a podcast i have a podcast so yeah it was negative for flu and covid and strep and so she was just like yeah i don't know i mean If you want to do a spinal tap to see if you have meningitis you can go to the emergency room and i was like do you think i have meningitis she's like no i don't it's like okay well then i'm not Aaron Francis Gonna do that so kind of worthless you get the weird urgent care doctor it's like that's such a crap shoot like you get the good one you're like oh yeah i got the good urgent care doctor it's Like being at the er it's like we're on top of shit or you get the bad one and you're like boy this is like this is worse than chat yeah and me just she was she was super nice but was like i don't Ian Landsman Have any answers for you and i can't i can't do anything else here right like my phlebotomist is gone because it's you know seven o'clock on a saturday night they leave at four on saturdays I don't have any tools you can go you can go to the er if you want or you can just take ibuprofen and go home it's like all right i'm gonna go home then so so you seem okay now so you shook it off I don't know i think so i didn't wake up sweaty last night so that was the first night that was the first night and maybe two or three nights that i like didn't wake up with a fever so potentially Aaron Francis Potentially i have shaken it off i am always sick for at least a week when I get sick. Any kind of sickness, always a week at the minimum, like usually longer, but minimum a week. But I don't usually have this like, yeah, this is weird what you're describing where you had a fever and then you didn't or you felt somewhat better, it sounds like. And then you got refevered at the end of the week. And then that's a little weird. Ian Landsman I know. I think it's a kid disease my son has the fever today. And, you know, last week my youngest son had some, like, spots around his mouth, which is like, oh, do you have hand, foot, and mouth, my friend? But he never got a fever. So I don't know. Just freaking kids, man. Always sick. Aaron Francis They bring you all the disease. It's terrible. Ian Landsman They do. Man, all right're on you're on the mend you're recovering i'm on the mend i should be back last week was rough because it would be like oh i'm you know trying to work you know i didn't fully Register that i was sick because in the middle of the week i was ([Time 0:00:24](https://share.snipd.com/snip/24bc7a88-dc18-4595-addd-98bcf1fc95cc)) - **Misunderstanding Basecamp's Success** - Criticism of DHH and Basecamp often misrepresents their goals and misunderstands their market position. - Success can mean lifestyle and profitability, not necessarily beating large competitors like Jira. Transcript: Aaron Francis And you decided this was the week you were going to throw down with the internet con. Got a little feisty. So why don't you tell us what happened? Because we don't get feisty Aaron. Ian Landsman Here's the problem. Let me see if I can pull up the tweet. I've had to, obviously, I've had to mute the thread because I can't. I'm not. I got a life to live. I always got time for that no goodness gracious all right so uh somebody tweeted michael greenwich tweeted uh it was a quote tweet of some they were talking about uh dhh moving out of the Cloud right and quote tweet whatever his commentary was imagine losing to Jira, talking about Basecamp. Imagine losing to Jira, but boasting that you save a couple million dollars per year. DHH shares all this virtue signaling as if they are business masterminds, but seems they just got smoked by Linear, Asana, Monday, and even Atlassian. Does anyone actually use Basecamp? I literally can't think of a single company. That's a crazy ass tweet. So you got to admit I was set up, right? I was set up. This is not my fault. This is not my fault. Okay, so here's my problem. Here's my problem with the tweet. First of all, that tweet has 817,000 impressions. My reply has 30,000 impressions. And I got the ratio. I have more likes on my reply than he does on his original tweet. It feels good. That's right. The people are with me. Here's my problem. Here's my problem with this tweet. And here's why this scratched my like, this is an injustice bone, which will make me quite feisty. This is not like an intellectually honest argument. It's frustrating to me. Like, you can have an opinion that VC is better than bootstrapping. I'm totally fine with that. I'm actually not terribly interested in arguing that with somebody who's like VC brained. That isn't really, you know, that's just like an opinion. Like, I don't really care if you have that opinion. That's great. We can have a discussion. We'll both go away feeling more solidified in our own opinions. Right. Problem I have here is people ascribing their own points of view to someone else and then saying, well, your strategies don't match what I believe your point of view is. And it's like, well, that's because that's not their point of view at all. Listen, DHH, good, bad, indifferent, don't really care. But to to assume that DHH wants to take on Jira is just like you haven't done the reading. You have no idea what you're talking about. You have no idea what you're talking about. Yeah, it's crazy. So like, I don't even I don't even care. I don't even care if DHH or Michael is right. This whole like this whole like I'm going to put what I want out of my life onto you and then criticize you for not living up to it. It's like, come on. This is very, very stupid. And so my response to him and I was trying I was trying to just like, you know, keep it easy bree i was easy breezy and i said you could put it another way imagine making tens of millions of Free cash flow a year owning a dozen super cars and a mansion in the hills and caring at all about jira sounds good to me and that's the one that that's the one that got the ratio and his response Was having nice cars and a nice house is fine if those are your goals and why you were. And I was like, yeah, you made my point. So I responded again. Exactly. DHH has said many times it's a lifestyle business. And in that regard, he's unambiguously winning. And this is what I just described to you. I think a lot of people are projecting their own goals onto him and saying his strategies make no sense. Well, yeah, he's not working towards that goal. And that's the part that like really, that's the part that always bugs me is where people are making a bad faith argument. And I don't like more than like wanting to enter in the fray of the arguments. I want to point out to them, and this is a fatal flaw of mine. I want to point out to them how wrong they are in constructing their argument. Like, that's just a bad argument. There are great arguments you can make. And I will let you make those. But this one, this doesn't even make sense to me. So, yeah, I got a little feisty. Aaron Francis I think you're totally right. Mean i think it's unfortunate that like uh yeah i think that he doesn't seem to know anything about base camp or jira in the sense of like that they're not related yeah and that jira makes No money no one ever plays jira with base camp first of all like these are not the same tools in any way like it's just a silly comparison and so like just on that whole level it's like you're Trying to make this point but since you don't understand what base camp does and what jira does i guess maybe uh that this the point doesn't make any sense um also i think i've been on this Hill forever and we've discussed this on the show but i think everybody especially the vc guys really underestimate how much money base camp makes and i think so like if you're not winning And you want dollars to be the measure i mean i think base camp you're good you got a while with your vc startup to match base camp like you i think so tons and tons and tons of money more money Ian Landsman Than most people imagine that they make i think they really make a tremendous amount of money so like and that's that's to say nothing of like that if the if the revenue is the metric those Vc companies still have a long way to go if after tax money in your personal bank account is the metric i don't know if those vc guys are ever going to get there yeah like your vcs may get there But the vc founders i don't know if they're going to get there. Aaron Francis Really have to hit the super home run. Yes. To get there. Like you have to be in the ultra 0.01% or more of the, of the ones that even make it right. To make that kind of money. And even if your measure is, I think the other implication is like the measure is I want to change the world. Correct. And I feel like, boy, I don't know, like, uh, if I'm going to compare B2B dev tool product, which is a very nice product and I've even used it and like it. But comparing that to Basecamp in 2006 and like what it did for like software in general, SaaS in general. Small business in general. Small business in general. Like, there's pretty much almost nothing I would say you could do that's going to equal the sort of personal human impact that they had. Like, your B2B dev tool thing, it's just not ever going to be the impact of Basecamp on a human level either, which is fine. Like, hey, I built a product that doesn't have human impact on the level of Basecamp either. That's fine. Like, that doesn't have to be ([Time 0:06:52](https://share.snipd.com/snip/1081c8b7-bdde-4fe7-a6fe-9ffc32e02990)) - **Basecamp's Unique Market Niche** - Basecamp serves agencies and consultancies primarily, not software developers or direct Jira competitors. - This niche focus explains its continued relevance and user base, despite popular misconceptions about its scope. Transcript: Aaron Francis Know you think that i know that and i don't care down the like i mean the guy has accomplished some amazing things and to say that they're not as silly like also does anyone use base camp Ian Landsman Kind of got me i was like oh bud yes maybe time to drive outside the san francisco correct like three minutes just like get out and some of the replies were like i've worked at four different Companies and nobody's used base camp and i'm like guys he's not financing the house in the hills he's not financing his f1 race cars you know and it's just like again it's like that i don't Aaron Francis Really get why people's perception is if you work at a software company in silicon valley or anywhere that you would expect to see base camp. That is not who it's for. It is not a bug tracker. It is not even a software developer tool, even though they use it themselves that way and they do pitch that you can use it that way. It's like a planning tool, not as a bug tracking tool. Ian Landsman It's more for PTAs than it is for bug tracking. Yeah, it's for consultancies. Aaron Francis It's for communicating with your consulting clients and agency. That is who uses it. And practically every consultancy I know uses it. Maybe not 100%, but a lot of them. And yeah, there's plenty of people using Basecamp. Those guys aren't going on. They basically keep 60 people employed so that they can build side projects. That's literally what most of those people are doing. They're like, let's build this other thing. Let's keep Basecamp running. You know they're not they could fire half the people there easily and just pocket another 20 30 million dollars and there would be no difference in base camp itself the product like they Would have new versions it would be maintained and everything would be fine and they'd have extra 30 million dollars a year in their pocket but they don't because they want to just have People around do other stuff they they want to do, which is cool. And that's it. So like the whole thing is silly. ([Time 0:15:10](https://share.snipd.com/snip/ef742096-d842-49f4-a662-c45114aa7292)) - **AI Hype vs Reality** - Many current AI applications overpromise and underdeliver, confusing hype with practical utility. - AI will improve, but foundational knowledge remains crucial to avoid missteps. Transcript: Aaron Francis Don't even want to get into ai base camp i can't i can't do it but um yeah oh boy this just feels like you know this is like when you hit that, feels like we're like maybe approaching bubble, Peak bubbly. You know, there's just so many people out there with so much money. Yep. And they're like, this is the future. And like, I don't know if it'll be like dot com bubble level, but it just feels like maybe there'll be a little shakeout. Maybe tariffs and recessions and things like that will shake it out a little bit. But like, it feels like there's a shakeout out there where like, there's too many people with too much money, sort of trying to make up new things that don't even necessarily make sense. Ian Landsman Doing too little. It's all just like, what do you want to build today? That's what every app has turned into is what do you want to build today? And it's like, just a text box and you got to type a to-do app. And it's like, great, here's a React to-do app. It's like, now what? Like, is this good? Is this better? I don't know if this is better. There are certain things, of course, that are so much better. And that's like, you know, generating crud and, you know, doing transcription and some things that AI has done is just so much better but i don't know if like the majority of the ai implementations Aaron Francis Have made things better yeah and i think there's the case of like the ai is getting obviously this is the worst ai is ever going to be and it'll be better tomorrow presumably right it'll Be better next year and so like there's a case we made and like you should be building towards that towards that day totally years from now where it is better or 10 years from now whatever Like and you'll be there and ready when it as it evolves right i think that's fine but yeah this like the the just yeah you see like in my world you see a lot of like oh upload pdfs and point Us at your website and we suck up all your stuff and we make a chat bot that answers a bunch of stuff mostly incorrectly or partially incorrectly because just sucked up this stuff it doesn't Know anything else about it just like sucked up your bad website copy and is like answering questions and sure it answers some questions right but it also answers some questions wrong Pretty often so like it's like yeah again like part of it is it's going to be bad before it gets good and i think it's fine but there is absolutely this whole like, yeah, just what even AI Tools, I think often the text box should not be the interface to them. Like, I think this is like a huge area where people are going wrong, too. It's like I want to type a big description. Like we can have buttons that just do things. That's what software's always had. And behind the button could be an ai thing that happens or whatever but like i don't want to write everything i want something to do what if i don't even know what it can do right like there's Ian Landsman Just other interface affordances that are important yes that's the thing that's the thing that um is is slightly related but slightly tangential about ai is you said something like I don't even know what it can do that's how i feel about using ai as like your beginning point of education like i i'm not an ai doomer i think ai is is great i do think it is overused just like We've just talked about i think it's used in a lot of bad places or a lot of like poorly thought out places. But I think AI is incredible. What I don't think it's good for is serving as like your foundational underpinnings of knowledge. Because you go into the AI and it, as we all know, gets like rabbit hole. It's like it goes down one path. And if you don't have enough knowledge to say, bro, you got to back up. There's a whole different thing you could be using that is way easier. It's you're just going to take it as is and it's going to be terrible. And so that's that's the part from obviously from my side, I'm very, you know, keen to try to convince people of it's like, you still got to learn fundamentals like learn the fundamentals As you always should have and then use the ai to like uh help you do specific discrete implementations or solve specific problems but you still have to know the landscape of like the world Aaron Francis If you're going to guide this thing that is super smart but also kind of dumb yeah yeah i think that to me the key element like um i was blue skying some people that i'm not going to use whatever Stupid term they use let's say tweeted i was tweeting ([Time 0:18:47](https://share.snipd.com/snip/3561c281-847d-4809-91c0-19098dba98ae)) - **Verify AI with Knowledge** - Use AI tools to save time, but always verify with your foundational knowledge. - Recognize when AI outputs seem off and be prepared to correct or override them. Transcript: Aaron Francis I think that to me the key element like um i was blue skying some people that i'm not going to use whatever stupid term they use let's say tweeted i was tweeting about like uh that like whatever The ai people who are like and met on the ai but i'm like it saves me time every day like some days it saves me 10 minutes some days it saves me 10 hours like it saving me time every day. But I, and yes, you have to watch it. And the watching it takes one minute to 10 minutes sometimes or whatever. But net, it's still plus, right? But the key to that whole workflow, and maybe this won't be true forever, but it's definitely true right now, is that you have to have the foundational knowledge to be like, huh, that Seems, something seems not right here. Maybe I don't even know what's not right. You sure about that? Yeah. But like something seems wrong with either what it said or what it said to do or what the code it wrote or whatever. Like there's some kind of smell there that you have enough knowledge to detect. Yes. And you can bring it back on track or just do it yourself manually if it's dependent on what the issue is. Right. And like, whereas, yes, if you were just going in there totally blind you have no idea what's going on and you're just going to assume everything it says is accurate or even if you don't Assume that you don't have enough knowledge to be able to quickly narrow down where it's probably going wrong then yeah that's probably not ideal at least at this phase it's going to Be you're going to get sent down some wrong paths and some learn some things incorrectly and all that kind of stuff which is not great look ([Time 0:22:55](https://share.snipd.com/snip/fb0d6799-e68f-4dcd-b1b1-7e33cf7a1e11)) - **Enhance Search with LLM Summaries** - Combine LLM-generated summaries with vector searches for better content retrieval. - Use expanded queries from LLMs to improve search accuracy over raw vectors alone. Transcript: Ian Landsman So now I've got the full text, and then I've got the LLM generated summary, which is like, I don't know, one, two, three sentences or whatever. Both of those things together get vectorized. So I'm already taking, I'm already taking, to use your words, a little more juice. I'm taking a little more juice into the vector, which is the LLM generated summary, right? So then on my retrieval, I can take my question. So like, hey, I'm looking for this, give it to the LLM to kind of expand the question a little bit, right? Expand it and then vectorize that query that has been expanded by the LLM, vectorize that query, and then check that against my database. So I'm getting a little higher chance of finding the right topics because I'm sending both things to the LLM and saying like, you produce something and then we'll check both things against Both things. How's that feel? Aaron Francis Yeah, I mean, I think that seems fine. I just think that like at the end of the day here, like you'll be able to just like you kind of have an extra step with the vector that i think you probably don't need like just i like the potentially In the future i agree and then just take all the summaries and i even even a category even like not literally all them maybe right like if you know you have a question that's within this Scope of type of question right you can just get all the summaries of that type of question and send them all to the lm and be like which one's the right one and skip the whole like vector Which is going to give you kind of even if it's better than it used to be still not going to be as good as the lm and you can just give it all the lm and let it pick it now they still aren't great When you have tons yeah it's not great things but that's all the stuff they work on a lot like every time they release new ones i agree like the needle in the haystack stuff is like they always Talk about it every single new model release is like i'm doing the haystack improved xyz but i've got i've got a hundred hours of content there's a lot that's a lot and if i'm asking a question Ian Landsman About, you know, what have I said about focus versus like, you know, working on too many things, I don't need to send it a bunch of text about HVAC systems and how Disney World sucks. It's already so cheap. But it's not price, it's performance. Like I don't need to send it 10 encyclopedias when really it's just like these three pages are the ones that are relevant, you know? Aaron Francis You're only sending it the summaries. You're not sending it the full body. Ian Landsman Oh, you're saying I send it the summary of everything and ask it, which ones do you want for context? Right. Oh, potential. Aaron Francis I'm still doing the heavy lifting. You're not sending it 100 hours of transcripts you're just sending it whatever 50 a thousand summaries whatever if you did everything it's thousand summaries but maybe you can even Limit the scope if you're something like a vector search yeah well no i don't think well maybe the vector search i don't even think so i think it's like what type of question you're asking When did i talk about philosophy right well this is a certain type a certain category issue. And so you can only send the summaries of those categories of discussions or whatever. Like, whatever. Even if you send it all 1,000, though, I think it would still be fine. I think it would be more. You should try. I think it would be more accurate than the vector. That's my guess. Ian Landsman I find that hard to believe. Why? The LLM is smart and the vector is dumb. I don't think the vector is that dumb. I'm still on the vector is dumb bandwagon. I mean, I've done, let's say, related article vector searches, right? I've done videos on that. And they're all very good. You know, you vectorize the entire articles. And then for an individual article, you say, give me the vectors that are closest to this. And then like you're on, you know, year in review 2024. ([Time 0:29:54](https://share.snipd.com/snip/a8e61cb1-4688-4b43-a3dd-421e40dbc769)) - **Marketing Requires Multi-Channel Presence** - SEO's decline makes multi-channel brand presence critical for marketing success. - Being visible repeatedly through various mediums like YouTube and podcasts builds brand affinity. Transcript: Aaron Francis Maybe there's gonna be missing payment like that's just disaster like you can't be like trying to track every payment and then our payouts like from lemon squeezy to us were just like Ian Landsman Random it was like oh you know it's coming you know april 14th oh no and now it's april 28th and you're like why tell me why yeah and then they sent me the wrong tax form they sent me you know They attributed all of try hard's revenue to aaron francis the person and i'm like guys guys guys guys guys this is wrong so lots of things where it's like, I think you guys pulled the Johnny Depp where you sail into port as the boat is sinking and then you step off onto the dock and you're like, I'm here. I did it. Wow. Oh, man. Yeah. Aaron Francis Well, that's good news, though. It's shipped. I love the kind of new platform. See, a little AI with a little guidance. Bug-free new platform. I love it. So that's all good. Yeah, the attribution. I mean, it's where like the brand advertising, which I've never done enough of. And I have some ideas to change that in the future here. But like, because it's just, yeah, you just need to be everywhere, kind of. And eventually people are finding you. And the thing is that that's like even more important than ever because now with like sort of the dying slash death of seo it's like there's not the free easy marketing channel for small Companies is kind of gone and so you need some other way to generate whether it's word of mouth or just be visible to people. ([Time 0:53:00](https://share.snipd.com/snip/1012d4d0-50e1-4e3c-970e-906de15d436e))