If Notion beats HubSpot, will they still lose to Claude?

[00:00.600] Eric: [upbeat music] This is different. [00:11.240] John: This is different. Yeah. [00:11.960] Eric: Yeah. It's gonna take me a minute to get used to it. [00:14.800] John: [laughs] Well, we're looking right at each other. For new listeners, we are, especially those on audio, we are in a new studio, [lips smack] uh, which is just a couple rooms down from the other one that we had. Previously, it was a small room, so we kinda crammed it into a corner and had a fake plant-... to make it look nice. And, uh, we're now in a larger room with way more comfortable chairs, and we're looking straight at each other-... which is awesome. Yes. Less decorations. [00:17.579] Eric: Yeah. [00:33.079] Eric: Yeah We did have a fake plant. [00:40.380] Eric: Less Less decorations, though. We have some decorating to do. [00:46.039] John: But it's cool. This is kinda like-... startup garage vibe. All right. You wanted to talk about Notion and HubSpot again. So for- [00:47.380] Eric: Yeah Yeah. [00:54.299] Eric: I need to rescind the last episode that we did on Notion and HubSpot. No. That's probably a little strong. We haven't talked about this. [00:57.179] John: You need to rescind. We didn't talk about this. We have not talked about this. We usually do way more prep, but you brought this up, and I said, "Let's just hit record-"... "because I wanna hear this live and-"... "see my own visceral reaction later." So- [01:07.019] Eric: Yeah Right Yeah. [chuckles] Can you do a re, do a recap for us? 'Cause y- I think you kinda brought up the, the topic initially. [01:17.939] John: I did, I did bring up the topic initially. So there is a lot of interesting counterintuitive competition-... in the market-... being created by AI. So one example actually that came up recently, someone wrote a post called OpenAI Should Build Slack. Right? I... Your average person, including me, is not going to n- wake up one day and say, "That's a really natural thing for OpenAI, you know, to do." But if you step back and think about it, it makes a ton of sense-... for something like that to happen. So a couple episodes ago, I essentially tried to convince you that Notion, even though most people view it as maybe a cleaner document management system-... you know, better than Google Drive, for example, I think is sort of the main use case-... uh, that actually they have the ability to go after a gigantic incumbent like HubSpot, and essentially be the entire connected system for operating like a small to medium sized business. And the basic thesis was, under the hood, Notion is a database, so you can build CRMs, you can build inventory systems, you know, you can really sort of build whatever you want. You have, uh, individual apps, so they acquired a calendar company. Uh, they built an email product, and all of those things are connected. So for example, you can see a database in your calendar if there are dates on it, right? Which again, s- that feels a little bit counterintuitive, but actually makes a lot of sense when you think about a connected system. And then the kill shot for my thesis, which you're about to dismantle-... was that all of those things individually, that's great, but really the AI agents make it all work together in a way that seems like magic. And especially for small to medium sized businesses, you'll be able to do things and execute workflows that are still actually very difficult for most businesses today. So the example that I gave was, you know, you get an email from a customer or a prospect, you need to like update the CRM, schedule a meeting, you know-... draft a response to the email, all that sort of stuff. Possible in Notion today. Um, and their new agent technology is really incredible. So we talked about a couple gaps. You know, uh, HubSpot has analytics. They have much more sophisticated website publishing, and then of course the big one is marketing automation. But Notion's so flush with cash, those are an acquisition-... or, you know, a smart engineering team, you know, a couple months away, both... And they have both options. And so I think that's really interesting. You know, I'm not on their board, um-... which may surprise some listeners. [laughs] [01:26.900] Eric: Yes Right [01:37.340] Eric: Okay. Yeah. [01:49.340] Eric: Right. [01:53.140] Eric: Right [02:09.038] Eric: Mm-hmm [02:14.180] Eric: Right [02:30.090] Eric: Right. [02:49.460] Eric: Yep. [03:11.439] Eric: Okay. [laughs] [03:22.520] Eric: Yes. Right. [03:44.300] Eric: Right [03:57.229] Eric: Right. [04:07.060] Eric: Right [04:14.199] Eric: Right. [04:18.399] Eric: [laughs] Not yet anyways [04:22.259] Eric: Where's your poster? [04:25.319] John: Yeah. [laughs] I know. I do need my poster. The, the CEO anchor poster we're gonna do. Um, so that was my thesis. So you wanna rescind... You mostly agreed with me, but you wanna rescind your previous perspective. [04:25.589] Eric: Yeah. Yeah. Exactly. Yeah. [04:38.069] Eric: [laughs] I think so. I think, I think there's a couple of things that I've been thinking about since we released the episode, and some of them from a, are from a technology side. Some of them are after actually working with Notion AI with somebody on something they wanted it to do. Um- Yeah. And then, and then maybe a general perspective. So I'll, I'll start specifically. So somebody, um, um, s- somebody came to me and was like, "Hey, I'm trying to do this thing in Notion, like with Notion AI. Like, can you help me? Do you have any thoughts?" And they had used Cloud Code to produce a bunch of markdown files, were happy with the results, were, um... I think they were planning a content calendar for a marketing team. And they felt like Notion was a better, um, UI for the marketers to use. So they were comfortable, like Cloud Code, markdown files. But let's put this in Notion. So they're essentially trying to recreate something they already know works, uh, like via this other method in Notion. So b- and like the short of it is they're having trouble. [04:53.060] John: Okay. Ooh, this is gonna be good. [05:09.459] John: Mm-hmm. [05:19.459] John: Okay. [05:25.170] John: 100%. Yep. [05:42.779] John: Hmm. [05:43.569] Eric: Um, and I think they had started with Auto, which, which in all of these, uh, AI tools, it's like let the... [laughs] There's this like sales pitch of like, "Oh, the tool will just pick the best model for you." It's like-... which is probably a cost optimization thing. So I told them like, "No, no, no. Like force it to use like Opus, you know? 4.6." [05:56.300] John: Mm-hmm [06:02.480] John: Right. 4.6. Something heavy duty. [06:04.300] Eric: Um, something heavy duty. Um, still had troubleAnd... and it was a relatively complex. Like, you can imagine, like, it's creating multiple entries on a table-... and then kind of stubbing out some content and, you know, things like that. Um, but here's what got me thinking. Because of that, like, particular struggle, I just got to thinking, like, "I don't know if their current architecture is gonna work right." Like, work how they would want it to. Um- Notion Notion. No, no, no. No. Because they, they replicated exactly what they want. They're just trying to improve the user experience with Notion-... to make it more accessible to non-technical people. So there's been a lot of, like, growing buzz around Obsidian, which is kind of a Notion competitor, but more, like, kind of open source core, like, l- more techy people, I think, in general use it. [06:07.959] John: Interesting [06:13.579] John: Mm-hmm [06:17.279] John: Mm-hmm. [06:28.160] John: Mm. You mean-... the individual or Notion? Okay. [06:39.199] John: Okay [06:42.540] John: Mm-hmm. [06:56.399] John: Yeah. I mean, it's kind of funny because probably the way that s- you know, people would think about Notion as, "Okay, well, it's like a clean document editor." I think about Obsidian that way because the main use cases from the people... I, I don't use Obsidian. But the people who use it, I... It's kind of like, okay, just a note taker. It's a markdown editor-... essentially. Which is how they started, I think. [07:06.199] Eric: Mm-hmm. [07:12.538] Eric: Right. [07:17.420] Eric: Mm-hmm Right. Yeah. Yeah. It, th- that's been the core of the platform. Um, and they've just greatly benefited from AI, which is just phenomenal with text files. Searching text files, editing text files. It's just, like, very comfortable. Um, and, and I'm not ev- even saying, like, for them to take over a lot of Notion's market, like, is... Like, the personas are wrong. Like, they'd have to redo it. Like, I don't think that's probably gonna happen. Um, although Obsidian has a massive, like, marketplace store, and you could basically rebuild a lot of Notion-... with all the right plugins and stuff, but it's just, the personas aren't right-... for, for, for Obsidian to take over Notion. So I'm actually not arguing that. I'm arguing their architecture is correct, and I think, I wonder if- Yeah. For the AI stuff. So I'm wondering if Notion's current APIs, nested pages and structures and stuff-... is going to struggle-... if they're gonna struggle with AI because of that, whereas Obsidian is just, it's already on... It's landed on disk, and it's a bunch of markdown files. Like, the AI will crush at that architecture. Um, and I don't know. Like, maybe they work it out. Maybe it's just a model release ahead, and it'll be fine. But if they had to re-architect a bunch of stuff, they've got cash, they could do it. Um, and, and again, I don't know enough about their backend to know what they would have to do, but-... that- Um, okay. Yeah. So- [07:27.300] John: Mm-hmm. Totally. Mm-hmm. [07:41.860] John: Totally. Right. [07:50.040] John: Right [07:53.899] John: Hmm Mm-hmm. [08:02.470] John: That Obsidian's architecture is correct. Hmm. [08:11.139] John: Hmm Interesting [08:20.660] John: Okay. [08:24.100] John: Yep. [08:39.779] John: Right I think it's Postgres. I think it's just all Postgres under the hood. And they... I think they... There was a point at which they al- the company almost died because they waited until... They waited as long as they possibly could to shard [chuckles] Postgres. Uh, which, you know-... for the, for the non-technical... You, you're a database guy. Explain sharding and why-... that was potentially going to bring Notion's business to its knees. [laughs] [08:44.730] Eric: Yeah. [08:56.519] Eric: Oh, okay. Yeah. Like- [09:03.519] Eric: Yeah. Sure [09:09.750] Eric: [laughs] Yeah. I mean, I think the, the easy... Here's my best analogy for this. Imagine you have an old computer, and like, every time you boot it up, it says, like, "You're, you're low on disk space." And you're like, "I'll deal with that later." And then you boot it up again, it's like, "You're really low on disk space." Then you boot it up again. And then you, and then, like, you go to boot it up, and it, like, doesn't start. Like-... that's essentially... Like, you can imagine, like, the database partition, shard, whatever you wanna call it, as getting to capacity of what it can actually, like, do. And you need to have multiple of them. And I, and I guess the story is that they just waited, like, way too long [chuckles] to- [09:22.700] John: Right. [09:26.370] John: Right. Yep [09:36.750] John: Yep. [09:42.139] John: I think it... I mean, apparently the story is that every engineer quit everything they were doing, and that that's the only thing the company worked on. Which is wild. Yeah. Yeah. [09:48.980] Eric: Crazy. I did not know that. Okay. Well, I guess they have a history here then. But I guess my point is, interestingly enough, my prediction is not that, like, Obsidian's gonna, like, dominate them in the market. My prediction is they may have to do some major re-engineering-... um, based off what I'm seeing, like, how these things work really well and fast. [10:02.399] John: Interesting [10:08.820] John: Did the person you were helping, though, did they actually get their project working in Notion? Like, did they get to a point where- [10:17.799] Eric: Not really. I don't, I don't think they were happy with it. Um, but I do think they kind of, like, developed a workaround where we can kind of break this down into more steps and, like-... you know, like, 'cause anything with AI, like, if you... The idea was like, "Hey, I wanna, like, give somebody this thing," and they run it, and it works. But most things, if you broke it down into more steps, like any of the AIs will-... AIs will be fine. So I think that was the end goal-... which wasn't necessarily bad, and they'll probably still use it. And again, Notion AI is probably, is not gonna get worse, I don't think. Um, but I, but I do wonder if, if the architecture is going to work. Not... And this is extending beyond Notion. The people that don't, that have the traditional... It, it seems like it's a fair argument to say, "We might need to re-architect a lot of things-... because we have this new technology that works in a completely novel way." [10:18.990] John: Oh. Okay. [10:28.879] John: Hmm [10:36.460] John: Right. [10:40.120] John: Hmm [10:43.539] John: Interesting [10:46.590] John: Right. [10:50.720] John: Yeah, totally. [10:57.460] John: Mm-hmm. [11:06.000] John: Hmm Mm-hmm. [11:10.740] Eric: And when people start new, they're not doing it this other way typically. [11:14.600] John: Yes. Yeah, yeah, for sure. For sure. Which is wild because that's changed in the last, like, three to six months. But h- this would be my question. As I think about the use case that you're describing, I think it's really atypical-... for the archetype-... Notion as HubSpot-... company that I'm thinking of. So generating a bunch of markdown files and creating that taxonomy outside of Notion-... and then trying to import it, I think would probably be challenging-... because of Notion's architecture, which is your point. But what if you created all of that within Notion to start? [11:16.620] Eric: Yeah. Yeah. Yeah. [11:29.559] Eric: Okay For the small business Right [11:46.919] Eric: Right [11:51.159] Eric: Right [11:54.620] Eric: Right. [11:58.919] Eric: No, no. That's what they did. They, like, prototyped it and then went and tried to do it native in Notion. So they weren't, like, importing files. Yeah. WhichI mean, I guess the other thing, and this is, this is like a bigger picture thing related to the Notion. So that's like one thing, but I do think the use case is, um, it was a pretty aggressive use case. Like, I wanna like kinda one-shot plan, like a full month of- You know, of content, and then I have all the... Like, it was fairly advanced. Um, but I think the other use case or the other thing that, that's interesting with Notion is it's, it's a simple tool. It, it has an awesome, like, user base that, that like a small business can sign up for an account or like an enterprise and, and it just spans like a big group of- Lot-- It spans like one persona but across a bunch of businesses. They didn't-- They somehow bridged the enterprise like free user gap- Which is really hard. Um. [12:05.440] John: Oh, okay. Interesting. [12:24.840] John: Mm. [12:28.879] John: Yep. [12:50.309] John: Huge. [12:53.708] John: Yep. [12:59.480] John: Totally. Yeah. In, yeah, I think in particular, for particular personas. So for example, in, you know, in product development or like a team wiki- Or other things like that, it's, it's used. I mean, I actually, right before we started recording, I was researching MCP stuff and came across an integration for Notions MCP. And I believe on their site it said ninety-eight percent of the Fortune Cloud one hundred, which I don't know what the Cloud one hundred is. [13:06.080] Eric: Yeah. That's what I mean. Yeah. [13:12.139] Eric: Mm. [13:27.769] Eric: Mm. [13:36.320] Eric: What's the Fortune Cloud one hundred? I guess that like- No ExxonMobil in that list. [both chuckling] [13:37.809] John: There's- Right. Right. Yes. But I mean, that is, that is pretty interesting. Um- [13:42.799] Eric: Right. Yeah. That is interesting. I-- Okay. So here's my other like more meta theory, is I, I think, cool, more advanced use case, whatever. Like, a lot of people can... A lot of people will be able to use Notion AI for their email, their Slack, to do some planning, to do, like generate some docs, like it'll be great. I think the other thing though is, you know how there's this principle in, in like technology where like everything eventually becomes JavaScript? Have you heard that one? [13:57.279] John: Yep. Yep. [14:06.169] John: Totally. [14:14.620] John: [chuckles] Yeah. Yes. [14:16.049] Eric: Yeah. Yeah. Which, which the point of that is, is everything eventually like becomes the most, I don't know if simple is the right word, but the most like condensed or simplified- Like, like I wonder, I wonder if that happens with technology too. And that, and, and what I mean- Beyond even Notion, where I live my personal life and I use text messages, like Apple Reminders occasionally- Apple Notes occasionally. Email, a Google Doc very occasionally. A Google Sheet very occasionally. [14:25.740] John: Yeah. [clears throat] [14:29.710] John: Hmm. Do you mean beyond Notion, like- [14:42.360] John: Mm-hmm. Mm-hmm. [14:47.000] John: Mm-hmm. Mm-hmm. [14:51.159] Eric: That's kind of it, you know? And where essentially those are the form factors that are so embedded that like a lot of stuff, a lot of other stuff goes away. And Notion you could say is a better version of Google Docs, but, but, but if, if the AI, like these AIs can either equip people, act as employees or equip, you know, people to work in a new way, I think the ways people prefer to think and work are basically along those lines. And if you can just dele- 'cause I don't-- We're like right on the line of like being able to delegate work or responsibility versus tasks. I think. You can-- There's a lot of tasks you can delegate and it can do- Responsibilities of like, "Hey, you're responsible for marketing." Like, people like do that on YouTube, but it's not like- That's not quite a thing. But, but if you could do that, and if we are headed there, then like how do you wanna interact with the thing- That like can do your marketing or whatever? I mean, probably like some kinda doc, some kinda spreadsheet, like text mes- you know, some kinda messaging thing. Like, I don't know, I mean, if it works how, you know, a lot of people think it's gonna work. I th- I think that it might revert that direction versus the additional complexity of other tools. Notion being kind of in the middle. Like it really is a better version of Google Docs in a lot of ways. So that might be an exception, but I think a lot of tools that are essentially just complexity, like may go away if that form factor becomes like, well, you just communicate with it the way you would want to, like personally, which is like the most easy, simple way to communicate. [14:52.399] John: Right. Right. [15:00.759] John: Hmm. [15:19.960] John: Hmm. [15:28.139] John: Interesting. [15:31.840] John: Mm-hmm. [15:37.710] John: Right. Right. Right. [15:45.840] John: Hmm. Right. [15:57.110] John: Mm-hmm. [16:12.480] John: Yeah. [16:27.259] John: So what's an example of something that you think would go away specifically? Yeah. [16:31.240] Eric: Oh, like a tool? Um, I mean, I hope Jira. [both laughing] A lot of project management tools. [16:39.279] John: Isn't Linear... Do you think Linear's gonna go away? [16:41.279] Eric: I don't know. I mean, that one's really good, and they're trying to lean into be almost like an orchestration tool. Um- [16:49.500] John: Right. Which was the only, uh, I mean, it's easy to say this in hindsight, but that's the only way that they can survive, I think. [16:59.620] Eric: Yeah. I think you, I think honestly, the only net new tooling, aside from what I like just described with like the basic tools like people use in their personal lives too- Is probably some kind of orchestration tool. But I mean, that's really broad. Like- Like, that's, like that'll be all the way from like one side Linear to another side, um... What's that com- [chuckles] What was the company that was on the list that, that you shared with me the other- Um, Monday, like Monday for sure. Um, I think, I think Linear will, will win out there. Um- [17:10.519] John: Mm-hmm. Hmm. Yeah. [17:25.539] John: Oh, Monday? [17:32.880] John: Oh, for sure. [17:34.500] Eric: Uh, anyway, there's several of the more technical like agent orchestration platforms that are like v- like LangChain. That's what I was thinking of. Like one of those is like way or n8n, like way over here, like- Technical whatever. Like will start competing with Linear potentially. You know? So that- [17:40.980] John: Oh, LangChain. Right, right, right, right. Yeah, yeah. [17:45.779] John: Right. Yep. Right. Okay. This- This goes back to counterintuitive- Counterintuitive- Competitive- Forces, right? Like thinking about n8n competing with Linear sounds, would've sounded so weird even a year ago. And now, you know, it's like, okay, wow. Like- [17:53.359] Eric: Mm. Yeah. Yeah. Yeah. [18:04.759] Eric: Right. Right. [18:09.400] Eric: But, but I think it gets drawn off of like agent orchestration, which has this like really broad like-Think of a bucket of like project management tools and then a bucket of like these really like kinda complicated workflow AI tools, like could all consolidate into like agent orchestration. And then the way people are working, it's like a Google Docs or maybe a Notion. It's like some kinda email tool, some kinda messaging tool, whether it's like oh, everybody just uses Claude or OpenAI and chats with people and agents in-... the same place. Like that might happen. Probably will happen. [18:22.971] John: Right. [18:35.922] John: Right [18:39.011] John: Right. [18:40.192] Eric: And then it just... Because the main complaint, I mean, I've been doing technology stuff for a long time, especially when I was like in technology leadership. It's like, "Oh, we have too, we have too many tools. Too many tools. We want them all to talk together." Like. [18:51.612] John: Yeah, exactly. Well, okay, so this is interesting if, as we think about these, these competitive forces. So I think what's compelling about Notion for a particular subset of users is that the closed connected system, let's call it-... [19:13.251] Eric: Right [19:15.872] John: it is better for... It's a better environment for AI to operate in. And I would argue it's probably a lot easier for the Notion team to build agents that do interesting things-... on behalf of their users-... because it's all connected, right? Linear is interesting because they have done a really good job of excellent integrations-... and MCP integrations-... and, you know, so it's orchestration, but it's actually, it's, it's they've done an incredible job of sort of the main, the main set of tools that their target users-... are operating in every single day. You can actually connect those very well, like as an orchestration tool. And so that's... I think that's what's interesting, like the integration problem is, I agree with you, like maybe the tool set itself actually shrinks, but that only becomes desirable if the subset of tools that you have are really deeply integrated. And you can sort of easily work across them. And I think for a lot of people, if, if there's friction there, like the limit of what they can experience in like a closed connected system, it will just be such a better experience that they won't... You know, it won't matter. [19:31.592] Eric: Right Right [19:46.711] Eric: Right Right [20:03.652] Eric: Right [20:26.951] Eric: Right. [20:30.402] Eric: Yeah. [20:46.820] Eric: Right. Well, and then, yeah, I think I agree with that. But the other direction this could go would be a bunch of competing ERPs, basically. [20:56.892] John: [laughs] [20:58.571] Eric: I mean, that's kinda sad, but like No- Notion-... just decides, we're gonna add chat. We're, we're gonna buy or build chat. We're gonna have calendar-... notes [21:00.592] John: Yeah [21:06.811] John: Oh, you mean like-... like they're gonna build Slack? [21:10.071] Eric: Yeah. Yeah, sure. Yeah. Just like literally you buy Notion, and you open Notion in the morning, and it has your email in it, it has your calendar in it, you take notes on it during the day, the AI takes notes for you, you talk to it for things you wanna do, you do things, you produce, [lip smacks] um, a- assets, whatever you want. And then maybe there's some boundaries where they decide not to get into email marketing-... and they have like a couple integrations. But I mean, especially with how fast you can, you'll be able to build software, and you already can, like why not do that? Even, even same for Linear. Like why not just, just move toward, you know... I mean, sure, maybe agent orchestration is the first thing, but eventually, maybe you just have a bunch of competing brands where you're like, "Oh yeah, we're a Notion company," and like it's just like full stack Notion, or, "We're a Linear company"-... it's like full stack Linear. Um, and then it, and then it's just a brand thing. I mean, I'm looking at a picture of a car, like your poster over there. Like what's the difference between like the, like this car versus another car? And, and people will be really passionate about the differences. But to run a business, just like to get around, like the answers you could use either. Yeah. [21:11.432] John: Mm. [21:29.332] John: Yep. [21:34.471] John: Mm-hmm Mm-hmm. [21:42.692] John: Yeah. [21:58.951] John: Yeah Right. [22:06.412] John: Right. [22:17.432] John: Right. Totally. Totally. Yeah, it's the bundling. I- it's actually like a really... It's, it's an interesting form of bundling because it's creating, uh, bundling opportunities that just didn't previously exist-... or would have been prohib- i- it would have been prohibited from pursuing them in some way, right? Whether that's cost or, you know, the going after multiple markets. But it is fascinating that, you know, these things are sort of combining to create-... potential, like significant bundling. Um, you know, which that is kind of a doomsday thinking about major [laughs] all these tools turning into major ERPs-... competing [laughs] with each other [22:21.432] Eric: Yeah. [22:35.332] Eric: Right [22:52.211] Eric: Right [23:03.912] Eric: Well-... well, if you, I mean, if you, if you watch any of the like... I can't think of a specific movie, but a lot of the futuristic movies that like they're set like 200 years from now. I can think of like there's this idea that everything's gonna be combined. Like, you know, like-... you know. L- like, and the like silly things like drug companies and, and like restaurant chains and-... you know what I mean? Like, like there's-... silly combinations of mergers that's- [23:15.271] John: Mm-hmm. [23:20.811] John: Mm-hmm. Yeah, for sure [23:29.201] John: Mm-hmm Yeah, yeah By and large from Wall-E, right? [23:34.352] Eric: Yeah. [laughs] Yeah, exactly. Yeah. But, but, but I kinda see that happening-... here. Um, and I think one other, one other thought. So, so maybe I'm not that... I think I, I definitely agree with the premise, that there's gonna be weird competition like HubSpot and Notion. I think, I think my, my biggest thing that I had to rethink since we last talked with Notion was, will, [23:36.231] John: That's, that's the [laughs]... [23:41.031] John: Yeah Yeah. [24:04.092] Eric: will, will people just not use, like not want to have all of that and basically- [24:10.451] John: Mm [24:11.875] Eric: Go with a model, with one of the frontier models and just... And it depends on the business. But there's a lot of businesses that are just very operational, like heavy-... where they're marketing light or like sales light even. Those may just go with a frontier model, and they just like chat and talk to the thing all day, and then they're in group chats with some people. And it's a little chaotic, but it kinda doesn't matter because you-... can just ask the model what happened. And, and it essentially like, has enough context to reconstruct things that normally would've been documents. But it just reconstructs-... on the fly for you. I think that's the other possibility that would be a big thing-... for Notion. [24:16.615] John: Mm-hmm. [24:20.536] John: Right [24:24.175] John: Mm-hmm. Mm-hmm. [24:34.316] John: Yeah [24:43.076] John: Right. Right Right. I agree And what's, what's interesting for Notion, I just-- we, we are just armchair quarterbacking Notion's future so hard. [laughs] [24:55.115] Eric: Yeah. They should probably have us, have us to talk about their future. [25:00.536] John: But what, you know what's interesting is [lip smack] the-- there is tension because Notion's MCP... Like, Notion's MCP setup is outstanding. Like, it is really good. I wired it up to Claude Code to-... you know, as part of my like- [25:20.996] Eric: Yeah Did they take the limitations off of it? The- Like, you can-- Like, I mean, you can like write pages and stuff. It was like basically read-only for a long time. [25:25.155] John: I don't know. [25:30.655] John: That's interesting. I don't know. [25:31.536] Eric: Which I don't know. Anyways. [25:33.915] John: I'm pretty sure you can write, but I don't know. [25:35.726] Eric: 'Cause that's-- 'cause there's a lot of walled garden MCP stuff happening right now. [25:39.155] John: There is a lot of walled garden MCP. [25:40.655] Eric: Where, where it's like we have one and it's great, and then-... you try to do something, you're like, "Oh, well, you can't do that." You're like, "Well..." [25:42.526] John: Right Yeah. But Notion's APIs are good. But also-... I am speaking from-... like we're, you know-... Vercel, we're on the whatever probably the top-... like enterprise tier is, right? But it works really well, and so increasingly if I'm, you know, r- if I'm working on a project, a lot of times I will, you know, have files locally in a folder, um, and use Claude Code or Claude Cowork- [25:46.855] Eric: Yeah Yeah Right Oh, true Yeah. That's a good point. Yeah. [26:12.215] Eric: Mm-hmm [26:13.756] John: ... and then just connect to the MCPs that I need, including Notion, which is, you know, a huge repository of information, um, you know, the company. And so that is a major way that I interface with Notion, right? Just through a different, just through an entirely different tool. Which is tension for them because- You know? [26:26.476] Eric: Question. [26:29.925] Eric: Yeah. Of course. Yeah. But, but is the content that you want stuff that you pr- previously produced and is in Notion, or is it team members' content that you want or, or both? Like, what's the mix? [26:45.756] John: What makes it really valuable is that I can ask it to search for content because the, there's so much in there. And, you know, it's just... I mean, it's-- teams try to keep stuff organized, but you're just a fast-moving company, and so-... you know, it's, you know, just hunting and pecking and trying to find stuff is really hard. Notion's AI search is, is a starting point for a lot of things-... that I do. 'Cause it's just like, is-- show me a list of all of the information related to this topic. Right? Which is really, really helpful. But then you can do the same thing, you know, through the MCP, uh, through the MCP integration, which is awesome. Right? So it's like, oh, this is interesting, like what information do we have? There's a bunch of, you know, product information, there's a bunch of marketing information, there's, you know, just, there's a wealth of knowledge, which is great. But it is kinda interesting because that essentially turns Notion into a database. Um, I mean, it is a database, but practically, right, the interface changes. And so, I don't know, it's interesting because that's much more of a, that's much more of the, the Fortune Cloud 100 play-... where someone like me who's working at Vercel interfaces with Notion in a bunch of different ways, but the API and MCP stuff is like really critical-... for this audience. But I think, you know, that world, even though probably like extremely lucrative for Notion-... like isn't most of the world. The wider opportunity-... is for people who don't know that Notion even has an API and don't-... care. Right? But it's all connected. [26:51.516] Eric: Yeah. [26:57.955] Eric: Yeah. Yeah [27:10.286] Eric: Mm Right. [27:16.556] Eric: Right. [27:24.855] Eric: Right. [27:35.036] Eric: Yeah. That makes sense. [27:40.695] Eric: Yes. [27:52.986] Eric: Mm-hmm [28:04.816] Eric: Yes Right. [28:15.276] Eric: Yeah [28:18.935] Eric: For sure [28:23.246] Eric: Right Yeah. Right. Right. Yeah, it'll be interesting because, yeah, I, I do think especially for, especially for marketing, especially for marketing or, or content, you know, any of those workflows, like you have to have a place to author and to store and to build up, like, information. Um, what I, what I'd be curious, 'cause I just haven't seen the use case as much, is like operational-heavy companies. Like what are, what do they do now-... and what would they do? Where it's, they just, for whatever reason, they just don't have... Maybe they like use a marketing agency, like-... so they don't really have much internal marketing. And, like, are they going to... 'Cause this is what I would imagine. They're gonna like have meetings, like keep the records in the meetings in like whatever frontier model they use, and then just like ask questions like, "What did we talk about in that meeting?" Or, and then same, like if they're gonna have meetings, they're gonna have email, and they're gonna have some kind of messaging. And then they want the AI connected to all three, and they'll just like ask questions. Um, and it'll be interesting, like what, are they okay with transient assets or do they want a lot of like produced assets-... that, that are physically stored somewhere? And, and to start with, like I think everybody's like, "Oh, I, I wanna know where that PDF is." But if it's good enough-... to produce it on demand-... which it will be. It'll be interesting if they care. Which I think they won't. [28:54.036] John: Mm [29:00.516] John: Yep Yep. [29:21.756] John: Mm-hmm. [29:26.016] John: Yep. [29:34.955] John: Yeah Mm. [29:41.155] John: Right. Right. It doesn't matter Yeah Yep. Yeah. Totally. Do you think Notion is threatened by Anthropic, you know, or OpenAI or whoever? [29:57.476] Eric: For that use case, yeah. For the... If you're actually producing content, like authoring content, and you're not just gonna like throw AI. Like I, I mean, of course, some companies will be like, "Oh yeah, we just like have it-... published to the website and we don't even look," you know? [chuckles] There'll be some version of that, but I-... I don't thinkI don't think there's a long-term content play for anybody of fully generated like AI that no humans look at at all. [30:09.536] John: Mm-hmm Mm-hmm. Right [30:24.544] Eric: Except for maybe... Ex-except for like, I don't know, maybe like user's manual. Like-... there are certain kinds of content, sure, or technical writing maybe. But the purpo- but like thought leadership or like things like that, I, I don't, I don't know. I don't... I think you need a place to author and to revise and to like-... edit [30:29.364] John: Right [30:32.953] John: Right. [30:41.604] John: I agree-... but I mean, I think Clau- like Claude's desktop app-... to me is increasingly feeling like a natural place for that to happen. [30:46.683] Eric: Mm-hmm [30:52.644] Eric: Yeah, I mean, I guess they could add collaboration features and, and make it-... Notion-y. Yeah. Yeah. Sure. [30:56.084] John: Document editing You know? I mean, that, that starts to get really interesting, right? Because then to your point, you d- you know, you open the path of least resistance for someone using these tools, you know, let's just say your average, you know, knowledge worker or s- you know, someone using some sort of business software. It's like, well, I mean, the thing that's attractive about, you know, the integrations that you can add to GPT-... or the integrations that you can add to Notion or Anthropic or whatever, is that you just sort of have a single interface to access and work with this different information. And so it's like, okay, well, why wouldn't Anthropic just build a document editor like Notion-... and then just start storing your stuff? You know? I mean- [31:01.763] Eric: Yeah. To your point. [31:24.104] Eric: Mm-hmm [31:40.963] Eric: Right. Right Yeah. Sure. Yeah. I mean, they already have the, whatever it's called, the artifacts thing. It's just not like well built out. But they could- [31:51.183] John: Yeah, totally. I mean, the, it is kinda funny 'cause the way that you would do an editor inside of Claude, and they have a template for this-... where they say, "Build an artifact that is like an AI-enabled Markdown editor," and you build-... an app [chuckles] and run it locally. [31:58.743] Eric: Yeah [32:04.703] Eric: Sure Yeah. Right. [outro music]