Pre-Black Hat crunch time hits Miscreants hard as brand audits pile up and Webflow goes down for half a day. Andrew's new MetaMonster grid shows AI's bipolar nature - brilliant one moment, baffling the next. 🤖 Plus: positioning challenges, category creation in the AI era, and why Poboy the cat makes a terrible hide-and-seek player.

Links:

For more information about the podcast, check out https://www.smalleffortspod.com/.

Transcript:
00:01.03
Sean
What's up?

00:02.39
Andrew
Your energy is different after we hit record than it was before.

00:06.79
Sean
I know, I gotta turn it on I gotta turn on for the podcast.

00:10.99
Andrew
What's going on, man?

00:12.23
Sean
I can't believe we both just went like double peace signs on an audio podcast, by the way.

00:12.40
Andrew
you're a

00:17.63
Andrew
We have video now, Sean.

00:19.07
Sean
We do, we do, and do. We have great intros.

00:20.91
Andrew
If if somebody would ever approve the latest videos to go live, then we would have even more video.

00:27.34
Sean
Yeah, yeah, yeah, yeah. that sounds like That sounds like a job for someone.

00:33.60
Andrew
Oh, hey, we have a new episode up.

00:34.09
Sean
and

00:36.53
Andrew
Someone finally approved something.

00:36.57
Sean
Yeah. Yeah. I'm working on it.

00:40.65
Andrew
All right.

00:40.64
Sean
I'm doing it. It just sounds like a job for someone who doesn't have 30 hours of meetings a week right now. So...

00:46.70
Andrew
Jesus Christ. Yeah, I'm not going to lie I've been trying to meet with you and JJ at the same time lately and trying to find overlap where neither of you has a meeting is almost impossible.

01:00.43
Sean
it is It is pretty much impossible.

01:00.69
Andrew
Oh, boy, you sure you want to go out? Can we pause real quick?

01:05.95
Sean
Yeah, sure. OK, we'll come back. No problem.

01:09.12
Andrew
Thank you, thank you. Po'boy wanted out of the office, but also Wally, guy who works on our house, over. And so then Po'boy was immediately terrified and regretted his choice to leave the office and ran and hid under the covers of the bed.

01:26.62
Sean
Gotcha.

01:27.14
Andrew
It's pretty funny and kind of adorable. Like when our cat wants to hide from something, he just hides under the covers and it's like the most obvious lump in the entire world. So it's a completely ineffective hiding place, but he feels safe because he can't see anything.

01:42.68
Sean
Look, man, if I can't see you, you can't see me. That's how it works.

01:45.72
Andrew
He's an adorable, dumb little fuckhead and I appreciate him.

01:46.52
Sean
That's

01:52.19
Sean
the...

01:53.48
Andrew
Yeah. Yeah, yeah. Trying to get time on your calendar right now is tough.

01:58.27
Sean
yeah Yeah, same. Me too. It's like you start the day with yesterday's to-do list, you have a bunch of meetings, and you end the day with a new compounded to-do list, and then

02:11.16
Andrew
So what's going on? I thought i thought you were like removing yourself from a lot of client work. And yeah, what are all these meetings?

02:18.30
Sean
ah so i am ish but i also am not because it's pre-blackhat crunch time and i really have like i think i need to be there like to help you just get things across the finish line we launched a site last week we just launched a site that today we launched a site like the week before we're another monthly sprint to launch like two more sites

02:45.30
Sean
Yeah. Plus every single client wants a brand audit because, you know, they work with us at the early stage. We built, we built them a brand. It works really well for the first phase and they all reach a point where like they need to advance it, which is fine. It's normal. It happens with every single client. They just all consolidated onto the same last couple of weeks.

03:07.59
Andrew
Interesting.

03:07.84
Sean
So

03:08.03
Andrew
So you're actually doing a brand audit of your own branding work?

03:11.43
Sean
yes.

03:12.20
Andrew
Huh. That's kind of cool.

03:13.24
Sean
Yes. Yeah. Well, it's because like...

03:18.53
Andrew
Sort of seems like a racket, but...

03:22.74
Sean
Fair. Fair. I mean, it's not like we're charging for them. So it's part of the retainer.

03:26.12
Andrew
Wait, why are you not charging for them?

03:27.78
Sean
It's part of the retainer. It's part of...

03:29.03
Andrew
Oh, okay, okay, okay, okay.

03:29.02
Sean
It's, its you know...

03:30.39
Andrew
okay

03:30.35
Sean
Yeah, yeah, yeah.

03:30.90
Andrew
Yeah, yeah, yeah.

03:32.56
Sean
It's like it's time to evolve.

03:33.04
Andrew
You're charging for them, but you're not charging separately for them.

03:36.84
Sean
Yeah, yeah, exactly, exactly.

03:37.60
Andrew
Okay. Okay.

03:38.75
Sean
So it's like half a racket. No, it's, it's ah a you know, when you are in early stage company, you are asking for very specific things to get you up and running, right?

03:48.30
Andrew
Yeah.

03:49.09
Sean
and And over, you know, a year or so as you grow, like, lore gets built about your company more meaning.

03:56.03
Andrew
Hmm.

03:56.18
Sean
Your product changes, your content strategy changes, and all this sort of stuff. And, like, it at some point it evolves past that. what the branding that we did in the beginning can support and honestly what the branding they paid for can support, which is totally fine because it's like, you know, there's no reason that we start off and you pay us a bunch of money to do all this thinking about your brand for you to like possibly never even use it because you are still in that PMF stage, like finding PMF stage.

04:09.70
Andrew
Mm-hmm.

04:25.49
Sean
But yeah, a bunch of brand and and also competitors exist and the competitors copy you and it's like, and now, you know, At this point, we should just build invoices to some security companies because every time we do something, shows up on their site.

04:38.04
Sean
I'm just like, dude.

04:38.96
Andrew
Dang. Shots fired.

04:41.55
Sean
Yeah, yeah, yeah.

04:42.03
Andrew
You to call names?

04:43.42
Sean
No, absolutely not. No, security companies, we love you all oh equally. and But if you would like someone to just do, if you like work from the source, you might just, you know, hit us up.

04:54.89
Andrew
We love the ones that pay us money more. I think that's fair to say.

04:58.44
Sean
Yeah, that's fair. That's fair. We do love the ones that pay us money more.

05:03.82
Andrew
You can buy our love.

05:03.86
Sean
Yeah.

05:04.68
Andrew
Yeah.

05:05.48
Sean
You can. it is It is for sale.

05:07.42
Andrew
Yeah.

05:07.78
Sean
Mine and yours currently are...

05:09.29
Andrew
Kind of expensive, actually, but you can buy it.

05:13.47
Sean
It's... it's It's an investment. OK. It's an investment piece. ah Like, yeah, I think I'm just helping as much as I can get things across the finish line.

05:27.20
Andrew
Nice.

05:28.51
Sean
And it's just like, you know, everyone's got something.

05:31.15
Andrew
So we had we had a client ask us this this week, which is like, how do you balance the future vision for where you want the product to go in your like positioning statement and brand versus like the shit you can actually do right now? Because you don't want to over promise, under deliver, then you...

05:51.59
Sean
Yeah.

05:52.19
Andrew
have high churn and people like try out your product and they're like this is fucking bullshit but you also don't want to like like you know you want to hint at where you're going and you know give yourself that room to change and evolve and stuff so how do you how do you guys think about that

05:56.83
Sean
Like Devin.

06:04.11
Sean
Yeah.

06:07.59
Sean
Yeah.

06:13.61
Sean
So... I know which client you're talking about. I think like, like, like product roadmap is a thing, like knowing what your actual wedge is, is a big part of it.

06:24.07
Sean
You know, to me, brand strategy is business strategy is product strategy is all of it.

06:24.39
Andrew
Thank you.

06:28.62
Sean
at least all of it informs each other. So there's that, you know, it's like, like, yeah you know, at some point, every plenty of companies, their vision grows or, and and even if the company,

06:44.36
Sean
Their vision grows or they've always had the vision of like becoming this like full multi-product platform like a Palo, right? and like or Or they get acquired into and this sort of thing. But for the most part, you know if you want to survive, you really like there's so few companies that can like reach Palo size without becoming multi-product.

07:06.42
Sean
Actually, I would say it's probably impossible to become like a Palo, CrowdStrike-sized company without being a multi-product. company at some point i think all all of these sas companies become like software manufacturers and build things anyway

07:17.88
Andrew
Yeah. At the same time, I feel like a lot of them try to do that, make that change way too early, but that's, yeah.

07:24.61
Sean
definitely definitely so i think i think it's like you know what it like usually there's one product usually the product has a limited set of capabilities but there are opinions about said why those capabilities exist and that's what like messaging and brand is built off of and and and and you take on like you take in like context from like what the current market is what competitors exist garner categories how you actually sort of like are different that way and then we just try to start there before getting into like uh the dream state a lot of times we also like

08:02.53
Sean
like with brand, that the that the creative director at Collins, I'm blanking on his name, Brian Collins.

08:14.06
Sean
Yeah, Brian Collins has a thing about how designers are time travelers. They bring the future of what you are to like to the present, to the client.

08:22.58
Andrew
So

08:23.66
Sean
And I've always kind of had that mental model. And, you know, even if we present to you like a future, future, future state, there are things to cut back on because like businesses are fundamentally promises at its core and you don't want to not deliver on said promise because brand brand and reputation over time is built on keeping that promise alive or delivering on that promise consistently, which is also why.

08:47.78
Andrew
so

08:49.09
Sean
Yeah, go ahead.

08:50.37
Andrew
So let's get concrete.

08:52.36
Sean
OK.

08:54.06
Andrew
you ah ah You've got hindsight and you have you know too much about this, so it might not be the perfect example, but like my vision when we were starting MetaMonster was AI-powered Screaming Frog.

09:05.73
Sean
Yeah.

09:06.64
Andrew
What we delivered, V1, was generate page titles and meta descriptions. How would you have like thought about our brand and specifically like our positioning?

09:13.01
Sean
Yeah.

09:17.17
Andrew
What would you have made the H1 of Metamonster knowing that you like wanted to be AI-powered Screaming Frog? And like you and I have seen AI-powered Screaming Frog gets clicks, but like at the same time,

09:31.57
Andrew
you know when i tried to show people videos after calling us an ai-powered screaming frog they were like this is not screaming frog bro and i was like yeah

09:44.10
Sean
I think a couple things. one like One is that you are a bootstrap company and you operate very differently than a VC funded company.

09:54.99
Sean
And I know that's a cop out. I know that's a cop out. just But just to put that out there compared to the like one of our clients, right?

10:03.17
Andrew
I don't know. I think maybe VC backed companies, especially in today's climate, should act a little bit more like bootstrap companies. Like, they still need to make sales. They still need to, yeah.

10:14.24
Sean
Agreed. ish. But yes, agreed. Agreed.

10:17.22
Andrew
Yeah, I think more of them should try to make sales.

10:17.35
Sean
to unlock the next round funding.

10:20.21
Andrew
I'm becoming increasingly convinced that the that like being like, we're a VC-backed company, we don't give a fuck about sales, is like not a winning strategy.

10:20.98
Sean
Yes.

10:27.44
Andrew
because it just like Sales are like a valid way to like validate whether what you're building is actually valuable or not.

10:29.39
Sean
I don't think...

10:38.23
Sean
I know there's a couple clients that we've had that have had that mentality. I agree. Fundamentally, you need sales because sales is how you unlock.

10:46.74
Sean
Like ARR is how you unlock rev like the next stage of funding, right?

10:50.95
Andrew
And it's also how you prove that you've built something useful.

10:54.78
Sean
Yeah, same same difference. but yes you're right agreed i honestly think most seed company like seed stage companies should operate half like a services company and like a tech enabled service to prove out their product but i think you'll learn a lot about how not great part of your product is once like Because it's I think it's easier to sometimes to buy a service because it's cheaper and you get to use it and you get to see the client's reaction and it becomes like a paid trial.

11:24.28
Andrew
Yeah.

11:24.98
Sean
And for the people that we've kind of pushed to ah try that, we've seen it, you know, we've seen it work. Mm-hmm. Mm-hmm.

11:31.08
Andrew
Alex recommended that we try that with Metamonster. And I could never figure out like what the offering was. I didn't spend a whole lot of time thinking about it. Maybe if I thought a little bit harder, I could have found something. But like I could figure out the offering for you know like like Mike Privet.

11:50.06
Andrew
It was like, okay, cool. i'll do I'll do an SEO audit for you. And I'll give you like recommendations and everything. But but For SEO agencies, I don't think they want to buy an SEO audit. maybe they want to Maybe I could have tested, like, hey, do you want to buy someone creating optimized page titles and meta descriptions for you?

12:16.34
Andrew
Maybe?

12:18.18
Sean
Yeah, i don't I don't know how I would sell to an SEO agency off the top of my head. Well, to to go back, to the the reason I brought it to go back to your question is like, I think like the thing with Metamonster was that, the thing with like V1 of Metamonster was that you approached it as like, let's build this tool and just like focus on like getting meta descriptions right or meta titles and descriptions right, at this point is like a tiny module in the grand scheme of,

12:43.47
Andrew
Mm-hmm.

12:49.90
Sean
that so to me like

12:58.14
Sean
like it it it feels like it's two separate situations a little bit in my head because to me it's like the difference between like cool tool that got maybe misrepresented as AI powered screaming frog versus like platform vision that you can scale back into like specific little tiny modules and present that vision as, I don't know, something, maybe it's AI powered screaming frog, even if though the future vision is like all these other sort of clay for for SEO agencies, right?

13:34.29
Sean
I think,

13:36.03
Sean
that,

13:37.73
Andrew
It's hard. It's hard to figure out, like, how concrete to be in the early days versus how much to sell on the future.

13:40.20
Sean
Hmm.

13:43.64
Sean
Mm-hmm.

13:48.09
Sean
Yeah.

13:48.76
Andrew
And, like, ideally, your product is evolving constantly, like you were talking about earlier. And so you're, like, too stuck in the present, it's going to be outdated really fast.

13:53.34
Sean
Mm-hmm.

13:58.88
Andrew
But...

13:59.43
Sean
For sure. I think, I mean, I think one way you can approach it in terms of like the objection that like, Hey, this is an AI power streaming frog. that like, like, yes, but that is what we're getting to. Do you want to be an early design partner of our product?

14:15.59
Sean
And they kind of push that way. That's one thing that we see clients do you as well as like, you know, they like, At Enterprise, like, they are selling this vision, but it's to lock in future, revenue like, potential revenue if they accomplish said things.

14:28.15
Andrew
Yeah.

14:28.40
Sean
Again, right, very different because they have all this VC money and... Mm-hmm.

14:32.03
Andrew
Well, interestingly... Obviously, VC and Bootstrapped, you're going to take different strategies. Although, again, I kind of am coming to believe that like more VC companies should like act like Bootstrapped companies with more resources, which is kind of what they are. like They should still try to like build a fundamentally sound business, but then be willing to like set money on fire to test things way faster.

14:59.65
Andrew
So like you still want to build something that people want to... buy adopt use it's just you can like run more experiments and run them faster with experts instead of having to do it all yourself But the enterprise aspect is pretty different because like we are not building an enterprise product. And so, you know, if you're going into an enterprise, then, you know, convincing an enterprise to work with a startup, maybe you do have to sell the future vision to convince them that it's worth their time.

15:37.72
Andrew
In addition to being like, here's what we can do for you today and balance those two things.

15:43.29
Sean
and do you want I do want to touch on the

15:49.58
Sean
the... That thing about like... services for seo agencies i don't know one of the 22 subscribers that we have on youtube if you want to leave ideas in the notes that'd be very helpful but i don't know what what service so if if i'm an a seo agency what service would i buy i would buy yeah go ahead uh

15:54.50
Andrew
Sure. We can riff on it.

16:15.03
Andrew
Okay, so here's an idea.

16:18.11
Andrew
We could...

16:20.100
Andrew
I only and don't think most agencies would buy this, but top-tier agencies might pay us to build custom like rag pipelines for them.

16:24.57
Sean
Yeah. Yeah.

16:33.01
Andrew
Build... build like build some AI tooling for them.

16:38.60
Sean
yeah

16:38.73
Andrew
And it wouldn't have to be totally custom. It could be build it with off the shelf stuff. But like I could see and forward, I mean, literally my friend who runs an AI agency, like, or runs an SEO agency has been hiring people to, is trying to hire someone to do that for him.

16:56.71
Sean
Yeah.

16:58.26
Andrew
It's like write prompts and and build tooling.

16:58.57
Sean
Yeah.

17:04.16
Sean
Yeah, I think i think that's a, I mean, I think that's a great one. Whether it's like, I think there's like a couple of different things you can do, right? I think like one is like build tooling, whether it's like N8N or like custom agents or a rep like RAG for them.

17:19.35
Sean
But also like, uh doing like prompting prompt engineering workshops like i think that's also but like internal training at larger agencies that's also another thing and then they can see like that at play for like how it's used the meta monster

17:37.35
Andrew
that's a really cool idea yeah there's something really cool there like which again kind of comes back to like maybe the value maybe meta monster is less of an ai powered screaming frog and more like

17:38.18
Sean
v one

17:49.87
Andrew
I keep thinking more and more that I want to position it against like chat GPT rather than screaming frog. Like it's like when you've exhausted the limits of what you can get out of chat GPT and you need a better interface, like Metamonster is the interface for SEOs, the like AI interface for SEOs, something like that.

18:04.43
Sean
yeah

18:10.43
Sean
yeah i think so

18:14.54
Andrew
But that's like, that scares me because it feels like

18:17.55
Andrew
Category creation, which is generally a losing battle for bootstrappers. So I'm trying to figure out like how how to position it.

18:26.44
Sean
I'm starting to... So so this is like this is like an interesting thing. I'm starting to wonder how good that advice is in an AI world.

18:40.19
Andrew
Don't create a category.

18:42.00
Sean
Yeah, yeah. Like category category creation is expensive. I agree category creation is expensive. but uh i think one of the reasons it's been expensive is because of like the willingness to try new things and ai inherently is so so cool that there is a willingness like i think that's why ai's proud streaming frog got clicks and and across a bunch of our clients like when they're outbound just says ai blank like ai tool alternative or ai uh ai and then job role

19:04.56
Andrew
Mm-hmm. Mm-hmm.

19:14.96
Sean
Like that has the most open rates across the board.

19:18.99
Andrew
I think we're also seeing, because it's like a massive wave, we're seeing categories get created really fast because it's just like new technology, new like wave, categories are going to be created out of that.

19:35.90
Andrew
And so maybe it's not that we have to create a category ourselves, but maybe we can find one of these new categories to like slot into and grow with.

19:44.55
Sean
Yeah.

19:46.13
Andrew
like we don't have to be the very first you know like you know like we're not clay we're not having to invent the like agentic you know uh workflow builder sort of thing but like we can be we can sort of fit into one of these at new kind of ai native categories and maybe those categories are going to disappear in a couple of years and

19:45.63
Sean
Yeah, I agree.

19:54.64
Sean
Yeah.

19:58.28
Sean
Yeah.

20:01.95
Sean
Yeah.

20:09.57
Sean
yeah yeah

20:13.87
Andrew
AI will truly be integrated into everything, but I'm not so sure.

20:18.65
Sean
I mean, I also think it's like not worth, I think, I think you're right in terms of like, but like because you're building an ai product, this is the time to just try to catch waves slash latch onto to different things.

20:33.76
Sean
because there's a high willingness to try. i think that, uh, like, like the concern of like, Oh, will AI just be like, I, I, I think,

20:44.79
Sean
I think six months from now, yeah like the product will be yeah like you'll much like the product and thinking will be matured and you'll see more out of it, I think a year or two years from now, like, like the the strategy needs to change it's not like you know it's not like you can do the latching on forever, but it makes sense right now.

20:56.95
Andrew
yeah

21:05.96
Andrew
on the flip side like i can't help but feel like we help you like even the the workshop thing like we help you learn how to use ai better it's not the end goal it's not like the real progress they want in their lives like maybe they want to be able to use ai better but like why like what's the causation there like they they want to use ai better because they want to continue to compete with other agencies and they want to use ai better because they want to be seen as forward thinking they want to use ai better because they and so i don't want to get too i don't know

21:50.29
Sean
Well, I think,

21:50.72
Andrew
trying to do a better job of tying things back to like the jobs to be done, the progress people want to see in their lives.

21:59.40
Sean
i think I think in terms of like the prompt engineering training for SEO agencies, the the the reason I like so like offering like a sidecar service is because it just gives you much more insight, right?

22:12.20
Sean
If you are coming in as a prompt engineering consultant,

22:12.71
Andrew
Yeah.

22:15.70
Sean
you will see the things they want to automate and therefore it informs your product more because you'll see a trend across a bunch of things.

22:18.06
Andrew
Mm-hmm.

22:23.78
Sean
I think that's the value of that. Yeah, I mean, and I think in terms of like, I think all all the reasons you said of like why someone wants to use AI, i think that's all, I think that's all true.

22:36.54
Andrew
So anyway, back to the original question. If you've got big grand vision, but you can only do a little bit now, how do you what's the right way to position yourself?

22:39.28
Sean
Yeah.

22:46.54
Andrew
Or what's the right way to create a brand?

22:52.98
Sean
I think you latch onto the wedge. i think I think there's like ways to differentiate, right? You can differentiate by the folks you're targeting. You can differentiate by like a key unique feature that others don't have.

23:09.08
Sean
You can also differentiate by like a degree of like where X thing went way better and or whatever, which I always hate as ah ah as a general messaging thing and and for brand.

23:20.71
Sean
I think so so, like in my brain, I think there's a big grand vision for the current state of MetaMonster.

23:31.87
Sean
i would just, I think you could scale the vision back and and I think you can scale the vision back into like what other core modules you'll want to sort of release with and what core use cases you want to target.

23:50.86
Sean
I just think that just meta descriptions and meta titles was a weaker use case, but yeah, you you know that, or we've, we've spoken about that.

23:55.14
Andrew
Oh, totally. Yeah.

23:57.99
Andrew
Yeah. We picked a bad wedge. We picked a bad starting point. Yeah.

24:02.59
Sean
I mean, I think you picked a great starting point for, and think you picked a great starting point for like, let's, get this out so that we can think like like this and figure out all the problems that exist when when building these things.

24:16.38
Andrew
Yeah.

24:18.28
Sean
Yeah.

24:18.63
Andrew
I guess if the two starting points we were considering were like internal linking or like meta descriptions, internal linking is its own piece that is kind of...

24:27.71
Sean
Mm-hmm.

24:28.77
Andrew
The tooling we're building now, we're still not exactly sure how we're going to want to do internal linking within it because internal linking is just a totally different model.

24:35.79
Sean
Mm-hmm.

24:39.37
Andrew
But anyway.

24:40.62
Sean
Yeah.

24:42.25
Andrew
So, okay, I think what you're saying is, like, when you're an early stage startup and you've got, like, big grand vision, you want to build a brand that, like, hints at the vision but is focused on the differentiators now.

24:42.64
Sean
You would almost.

24:58.18
Sean
Yeah. And I think you can think about like the, like we did this exercise with the client where like we kind of just did like a now, like a, like a current, current state,

25:08.93
Sean
like near future and then like a future future state and it was easy yeah yeah yeah thanks yeah thanks um um um yeah i mean we we kind of just did that with them and like you know they had a now which was ai something and then a lot of the work was to work on and they they had a now and they basically had a later so the goal was to figure to like bridge that gap of the in-between and it was the call out like specific like

25:10.86
Andrew
Now, next leader, your product manager. Yeah.

25:37.20
Sean
there were specific things about their product that worked better for a type of infrastructure. so that was a big part of it. and And then there were specific work like workflows and stuff to call out that would map to like jobs to be done or use cases.

25:54.25
Sean
And I think that's where we got to with it of like what outcomes can you what business outcomes can you deliver on right now was one of the drivers of the message.

26:01.84
Andrew
Mm-hmm.

26:05.61
Sean
with the goal that like as they build a product they will fall into this larger you know gartner category yeah and i mean i think you can apply it the same way one of the i think one of the things that came to mind about like why vc and bootstrap is different is because uh with bootstrap it is permissionless and you get to just make the thing whereas with VC startups you have pitched a vision to a bunch of people they've given you money for said thing and now it's the work back plan of like how do we get to

26:40.54
Andrew
That's true. That's interesting. Yeah. Because you do need the grand vision to sell VCs. I think this is the hard thing about being a VC-backed CEO in particular is you've got to constantly be holding the future vision and the now in

26:54.80
Sean
yeah

26:56.78
Andrew
in your hand and the future vision's got to be big or else you don't get investment and you know for a bootstrapper my future vision can be like a couple steps from where we are now and that's fine but like for a vc back company you need this like big vision and being able to like rein yourself in and get concrete on what you can do now is hard

27:00.32
Sean
Yeah.

27:03.61
Sean
Yeah.

27:21.87
Sean
yeah Yeah. Challenges. I think challenges to both. mean, regardless, like I, i I think the current future state vision of metamonster makes a lot of sense to me.

27:32.10
Sean
one One of the things that one of our clients pushes back on or asks a lot about is like how much should marketing be ahead of product. And so, for example, like we did a solutions brief for them. Right.

27:44.05
Sean
And we did list integrations that were coming soon in the next quarter as part of their market texture. ah hundred percent 100 we did because like i don't see a reason not to do that you know and and and i think yeah i think but we didn't go as far as to like you know i like list ones that we know they're not going to be working on until you know a year from now or we're not listing out like

28:13.53
Sean
Like it's it's like 10% bump on what their current capabilities are because they're working towards it already. Like it's on their near your term roadmap versus like, you know, they are now the AI CISO, for example, which is not even close to what they can do yet, but could do in the future.

28:22.25
Andrew
I

28:33.60
Sean
Yeah.

28:34.26
Andrew
like that as a guideline, like a 10% bump on what you can do now.

28:36.32
Sean
Mm-hmm. yeah

28:38.92
Andrew
Anything that you could realistically say is coming soon, and then you can designate this is coming soon.

28:45.03
Sean
Yeah, two quarters ahead. Yeah. Yeah.

28:50.22
Andrew
10% bump. Cool.

28:52.37
Sean
Yeah, that that's a number I pulled out ass. I don't actually, I don't even know how can measure that, to be honest with you. I think two quarters ahead. I think you can sell. Well, okay, sorry. I think you can sell two quarters ahead in enterprise because your sales cycles are so fucking long anyway.

29:04.00
Andrew
Yeah.

29:05.33
Sean
So maybe in this case with like MetaMonster, maybe like a quarter ahead, three months ahead.

29:10.29
Andrew
A month ahead.

29:11.73
Sean
Yeah, two months. Yeah, I mean the middle.

29:14.75
Andrew
Maybe. We'll see.

29:16.60
Sean
How are things going?

29:18.67
Andrew
I've been playing around with the the new grid a lot. You have access, by the way, if you want to if you want to poke at it. I can send you my prompt library if you want to.

29:29.55
Sean
Cool.

29:30.35
Andrew
I think I'm going to turn this into a piece of content at some point, by the way, is just be like, here's all our prompts. you want to use them in chat, GBT, fucking go for it. It's going to be a pain in the ass. Have fun.

29:42.48
Sean
Sweet.

29:43.04
Andrew
Like, make it a downloadable or something like that.

29:46.26
Sean
Nice. Yeah.

29:47.79
Andrew
or just give it a, maybe I won't even make it a downloadable. I should make it a downloadable.

29:52.82
Sean
I think you should make it a downloadable, but then give it away on social one at a time.

29:57.54
Andrew
Oh yeah, yeah, yeah, 100%, totally. um God, then I gotta go rate all of those.

30:03.56
Sean
Post on social.

30:04.45
Andrew
Yeah, fuck.

30:05.22
Sean
Yeah, sorry.

30:06.89
Andrew
i have not done my task by the way i have not uh recorded my video part of that is because one i was procrastinating in until right before this call and then i started the same thing happened to me a couple videos ago where i was like trying to show off a new feature that we were building and then i started testing the feature and it was just like a rabbit hole time suck of like

30:18.94
Sean
Nice.

30:30.91
Andrew
figuring out how to get the most out of the new feature ai is so fucking weird in that you you can do anything but like getting it to do what you want it to do is just like takes massaging and and practice and tweaking and and stuff and so

30:43.79
Sean
yeah

30:48.89
Sean
Yeah. Yeah.

30:49.99
Andrew
basically what i'm saying is product's not fucking working right now it it sometimes you'll have the experience where you'll like run a prompt and you'll be like holy fuck that was cool like that took so little time like you know simple stuff like you know if i want to extract the h1 from a page and i

30:49.44
Sean
Mm-hmm.

31:00.10
Sean
yeah

31:11.30
Andrew
So before this call, I was trying to extract the H1s from the page. And like before, adding an H1 to our product would have taken ah writing some code and yeah and and then like publishing that code and having new buttons and designing new features.

31:28.70
Andrew
Now it's just, you know, enrich, slap a prompt in here, run it, test it, see if it works, tweak it. And now we've got all the H1s extracted.

31:40.50
Andrew
And then, like, at first it wasn't working because the we had a bug where our we weren't storing the HTML content in the grid.

31:49.55
Andrew
Like, it's stored on the page crawl, but it's not converted over to, like, the grid structure yet. and But then I was like, all right, I'll just adapt the prompt for right now to go off of the markdown content that is being stored in the grid.

31:54.49
Sean
Yeah. Yeah.

32:03.95
Andrew
And so, boom, I had H1s. And like, again, it's it's like coding with words. it's

32:12.09
Sean
yeah

32:12.32
Andrew
When it works, it feels fucking magical. And you're like, holy shit, I see the future. This is so cool.

32:17.76
Sean
yeah

32:18.23
Andrew
And then I try to get it to check to see if keywords are in the H1. And it's like, nope, no keywords in this H1. And I'm like, the H1 is literally the keyword. That's all it is.

32:30.90
Sean
Yeah.

32:31.01
Andrew
And it's like, nope, I don't see it. Not here.

32:33.43
Sean
it he Only two R's in sh strawberry.

32:36.03
Andrew
I'm like, what the fuck, man?

32:38.50
Sean
Yeah.

32:39.81
Andrew
So, yeah, I am currently, like, testing. i've I've got this big library of prompts that I spent, you know, several hours working with Claude.

32:50.77
Andrew
this week to create.

32:52.67
Sean
Mm-hmm.

32:53.41
Andrew
And they're super thorough. And they're not all working. So I got to figure out why they're not working. Figure out you know where we need to build tools and expose those tools to the AI.

33:06.60
Sean
yeah

33:07.25
Andrew
And where I just need a better prompt or need to try it with a different model or give it better input or something like that. So.

33:16.64
Sean
so i learned something the other day i i'm like hounding this founder to let me put an angel check into at this point uh they're doing they're doing a lot of stuff i'll tell you off the call one of the things i learned is like

33:25.36
Andrew
Cool. Mm-hmm.

33:31.17
Sean
I think Olama, you can you can test prompts with Olama because it's cheaper. It's locally hosted, or you can locally host it And then after you get more confidence in the prompt, then ship to production to like OpenAI.

33:47.16
Sean
And that's like ah way that you can reduce costs when fine-tuning all of your... Mm-hmm.

33:52.74
Andrew
Yeah, from what I understand, like prompts aren't always one-to-one between different models.

33:55.01
Sean
Mm-hmm.

33:57.86
Andrew
Like I saw this when I was testing our like original title and meta description prompts. Like Claude had different biases than ChatGPT.

34:04.27
Sean
Yeah. Right. Right. Fair.

34:07.87
Andrew
And so like the word list that I had for ChatGPT, like saying don't use these words, didn't apply to Claude.

34:08.54
Sean
yeah

34:14.61
Andrew
And there were all new words that I started noticing it repeating all over the place.

34:16.89
Sean
right

34:18.47
Andrew
So prompts aren't always one-to-one between models. I can totally see that being effective. Also, like, the kind of testing we're doing is pretty fucking cheap. So I don't care.

34:28.69
Sean
there

34:29.26
Andrew
I'm happy to test it on Broad.

34:31.26
Sean
Fair. Specifically, Olama and ChatubiT being like, they're not one to one, but they're similar enough.

34:38.52
Andrew
Oh, interesting.

34:39.67
Sean
That's what I mean.

34:40.21
Andrew
Hmm.

34:41.14
Sean
Yeah. Yeah. Yeah.

34:43.52
Andrew
That's fascinating.

34:45.82
Andrew
Yeah.

34:46.80
Sean
yeah I think it's really interesting that you were, because of this, like the side effect is that you become like a power user of the product. I mean, I i think I meet plenty of CEOs that are not, you know, so I think.

35:00.77
Andrew
that's That's one of the things that has me hyped about this direction is like ideally the marketing is me using the product and we use the product to build the product. like We want a new feature.

35:13.37
Andrew
We use the prompt the product to like test the prompts until we get them working. And then we roll out that prompt as a template. And that's a new like feature.

35:21.62
Sean
Yeah.

35:23.59
Sean
Right. and you've And you've experienced all the edge cases that could happen or a lot of the edge cases that can happen and all the different nuances. So you you are now the best user of the product.

35:34.79
Andrew
That's the goal.

35:35.02
Sean
Yeah.

35:35.59
Andrew
So, I mean, it is, yeah, it, it's already cool.

35:35.92
Sean
Yeah. It's pretty cool.

35:40.51
Andrew
You can, you know, so you can choose your context from different, you know You can choose choose your context. You can write a prompt. You can define your output and as one or more columns.

35:57.43
Andrew
So you can also create multiple columns at once.

36:00.89
Sean
Oh, cool.

36:01.69
Andrew
You can you know set your output to be text, HTML, markdown, URLs, numbers, single selects, multi selects even.

36:01.91
Sean
Nice.

36:11.27
Andrew
like You can do like all sorts of different...

36:12.31
Sean
Cool. Oh, you guys have been cooking. didn't...

36:14.78
Andrew
Yeah, Austin has been flying.

36:15.51
Sean
Yeah.

36:18.11
Andrew
Claude Code, he said, is like superpowers.

36:18.63
Sean
Uh-huh.

36:21.56
Andrew
He's like, it is ah ah level up from even Cursor.

36:24.49
Sean
Really? Nice. Nice. Mm-hmm.

36:27.53
Andrew
I've been hearing this from other people, too. that like and this is By the way, this is something that's been really interesting and weird.

36:36.23
Andrew
I feel like with so many tech waves before, there's like a hype cycle. And I hear the really hype-y, the hype-beast talking about like, oh, this is the next thing. This is the next thing. And I'm like, you guys are full of shit.

36:51.23
Andrew
And it all dies out.

36:52.55
Sean
Yeah, at some point your camera doesn't need more megapixels. You know you don't need it.

36:55.10
Andrew
yeah Yeah, or like NFTs aren't going to suddenly make crypto work, right?

36:56.87
Sean
I don't care.

37:01.22
Sean
Yeah. Yeah.

37:03.02
Andrew
With AI, I've had that same experience of like, You know, people have been hyping up agents for, like, in if you're in nerdy tech circles, people have been hyping up agents for, like, almost the last year.

37:17.79
Sean
Yeah.

37:18.27
Andrew
And I was kind of like, okay. And then it's like, oh, no. Agents, the ability to combine LLMs with code that does things reliably is actually really fucking powerful.

37:24.42
Sean
Mm-hmm.

37:32.20
Andrew
It's still a little wonky, but, like, it can be really fucking powerful. And then, like, the new hotness right now... that I'm seeing in developer circles is like Cloud Code and there's a you know a couple of others that that basically like they'll generate tasks for you.

37:51.05
Andrew
They're essentially generating like PRDs or like you know very detailed you know detailed context on what you want the AI to build and how you want it to build it.

38:03.78
Andrew
And then you can go in and edit that context and everything, and then feed that to an agent, and then like work with the agent task by task. And like this is the the workflow that I'm hearing people say is really changing how they build.

38:12.82
Sean
yeah

38:21.96
Andrew
And Claire Lu, do you know her she's a She's head of product at, like, a venture-backed startup.

38:26.65
Sean
whos

38:30.04
Andrew
But she built this thing, like, a year ago or more, more than a year ago, called ChatPRD.

38:38.80
Sean
Oh, I have seen that website.

38:41.16
Andrew
And she's been saying for ages that, like,

38:41.26
Sean
I've referenced that website.

38:45.60
Andrew
You know if you use AI to generate a PRG and then feed that to something like cursor, it's more effective. And I was always, know, again, I'm a little bit of a skeptic at heart. And so I was kind of like.

39:00.14
Andrew
I don't know that this is really going to be that powerful. And now this is the thing that like people are swearing by.

39:03.45
Sean
Mm-hmm.

39:07.00
Andrew
so it's, it's kind of challenging some of my skepticism radars um to be part of a tech wave that's maybe based in some amount of reality. But then again, at the same time, like,

39:20.77
Andrew
you know people swear ai is going to change the world and right now i can't get it to tell that building the ultimate seo sidekick contains seo sidekick

39:31.41
Sean
Nice. That's,

39:34.31
Andrew
so all this is to say like there's so much fucking hype and i think some of it is real and also this stuff is still hard and weird

39:42.05
Sean
what do you remember the last time you like, uh, this became like a bro about something like, you know, like, like a crypto bro or like ah you know, they're like bought in, bought in.

39:53.42
Andrew
i

39:55.85
Andrew
I think I'll always be skeptical to a certain degree.

39:59.22
Sean
Mm-hmm. Mm-hmm.

39:59.66
Andrew
I mean, i was pretty excited about crypto in like 20...

40:05.82
Andrew
16 2017 when i first that's when i first started really learning about it and i was excited about it i wanted to build a like a transparent voting platform to like try to make online voting a thing and i got really hyped on it and then got kind of turned off by all the bro-iness uh i was

40:28.93
Sean
Damn.

40:30.89
Andrew
I worked for my first job in tech was working for a company when like, you know, mobile apps were like the current hotness and we were just building,

40:45.38
Andrew
We were going we were trying to do the Peter Levels thing. We were going to build an app a week for a year and see what stuck. And we built six apps in like two months.

40:55.69
Andrew
And then we were like, this isn't really working. Maybe we should spend a little bit more time on something. and So I think that was like my first introduction to like hype waves or maybe bullshit.

41:08.74
Sean
Gotcha. Gotcha. speaking Speaking of mobile, just to digress a little bit, have you seen like Cal AI and all these 18-year-olds becoming like millionaires off of...

41:19.16
Andrew
Yeah, I'm skeptical. i this Going back to my skepticism.

41:22.68
Sean
Uh-huh. Uh-huh.

41:23.97
Andrew
Maybe this is real. how

41:27.74
Sean
know these kids are crazy smart they like yeah

41:31.88
Andrew
Are these is the kids who are like to have been talking for ages about like building consumer-facing apps and like like hacking TikTok to the TikTok algorithm to like sell the shit out of them?

41:40.73
Sean
yeah yeah yeah I think I think so I think it's like and don't remember their names but yeah they're all like 18 to 21 ish year olds and full-on like the new consumer app like

41:43.31
Andrew
and This is that kind of group?

41:58.25
Sean
you know, they did Kali AI, like some looks maxing thing. They're just building stuff. And I think that, yeah, I think with the Kali AI, like a lot of it was TikTok growth hacks to get, to get signups.

42:11.22
Andrew
Interesting.

42:12.34
Sean
Yeah. Yeah.

42:14.77
Andrew
Yeah, I've seen them.

42:16.55
Sean
I wish I had 18 year old levels of energy and carefreeness, but,

42:24.90
Andrew
Dude, most days it feels like you have way more energy than I do.

42:29.39
Sean
It's the caffeine. Yeah.

42:36.18
Andrew
yeah so i'm trying to get metamonster to work it's fucking cool at times and infuriating at times aka the experience of working with ai yeah

42:38.44
Sean
Yeah.

42:49.69
Sean
Checks out. Makes sense.

42:51.34
Andrew
yeah

42:51.72
Sean
Good luck.

42:51.89
Andrew
but austin has been cooking and this shit is like

42:54.21
Sean
Yeah.

42:57.60
Andrew
You can see the power. We make it more reliable.

43:02.77
Sean
I mean, like, love the multi-column stuff. Cool. i will I will try to give it a try by the next time we have a call.

43:15.22
Andrew
Sweet. That be great.

43:17.15
Sean
Yeah.

43:18.18
Andrew
Anything else going on with you? Or you want to rap early?

43:22.02
Sean
uh i want bitch and moan for a second tuesday webflow went down for like nine plus hours and oh my god my life like

43:25.06
Andrew
Do it.

43:34.88
Andrew
Damn, that's a this's the first I've heard of them going down.

43:38.10
Sean
it's it's the worst incident i think in all of my experience with webflow and like how many years has it been

43:45.06
Andrew
What happened?

43:47.18
Sean
I don't know. Nobody has said anything. My guess, to be honest, is the fact that they rolled out like an auto saving feature for see him for the CMS thing, which is really cool and making people's lives like like people love it from what I can see on Twitter.

44:03.55
Sean
but i'm I'm just guessing, like, I don't know, some auto-scaling fucked up or something. Because in the past, like, you know, saving to the CMS, like, felt very rate-limited.

44:15.02
Andrew
Yeah.

44:15.68
Sean
I'm guessing if it's auto-saving, it's, you know... saving as you as you type and and whatnot so anyway yeah tuesday was rough uh a lot of current clients had to send out like a message have this idea in my head of like for for the longest time of doing like a customer emailing list so that we can tell them things about you know, Webflow's had this update, this has this up update, and kind of doing that, and even for older clients, and and it would have been really useful for this, but I had to go through every single Slack, send like a Webflow service thing.

44:42.56
Andrew
yeah

44:48.79
Sean
If your thing isn't loading, this is this is why. Please do not burn our retainer thinking we fucked something up. And the worst part is, as was starting come back online a little bit,

45:01.69
Sean
and people's sites came back online the home page was offline was 404 everything else would work but the home page so yeah the home page people's sites dude gradeways was down for for a while but then i had a i had like a suspicion that if you just republished it you like it would come back online and then later on they sent down the thing like just hit republish so when i had the

45:09.36
Andrew
The homepage of people's sites? Damn.

45:16.50
Andrew
Damn.

45:27.09
Andrew
Maybe I don't want to move to Webflow. Maybe I want to stay with my weird-ass Astro on Netlify build.

45:32.98
Sean
Do it. I mean,

45:38.25
Sean
it makes me, it makes me like, we've thought about like trying to put together like backup contingencies for some clients, especially like our larger enterprise clients that we really don't want their stuff to go down and like build out like fit.

45:48.22
Andrew
Mm-hmm. Mm-hmm.

45:50.02
Sean
This is, this is where and actually just need like a CTO to go figure stuff out. it'd be really nice to to do something like that where there's like an automatic web flow flow over, web flail over.

46:07.18
Sean
Anyway, sorry. That was fun. Old clients that we've never even spoken to hit us up or haven't spoken to in a while. Like, they were texting me like throughout the entire day.

46:18.08
Andrew
Were they blaming you or were they just like, hey, what's up?

46:20.34
Sean
and theyre what Yeah, they were like, hey, our site isn't loading. What's up? What's happening? i was like, I need everyone to chill for a second.

46:30.13
Andrew
Yeah.

46:31.48
Sean
Yeah, I came out, i came back online at the end of the day. Yeah, yeah first first internet snow day. also It also meant that we couldn't, like both the backend and the front end went down. So like we couldn't make any changes for anything the longest time.

46:45.49
Sean
So that was fun. That was really exciting. don't know, most of my week since last week has just been like hours and hours and hours of meetings. But yeah, it's okay.

46:56.89
Andrew
Damn, sorry.

46:58.99
Sean
It's fine. Yeah.

47:01.12
Andrew
Not a fun week.

47:02.67
Sean
Yeah. Some of it is sales, which is good. You know, it's still a good amount of sales work.

47:06.25
Andrew
Cool.

47:10.85
Sean
Yeah. That's all I got. I didn't do my, I didn't do my, the thing I promised I would do, but for good reason, for good reason.

47:13.69
Andrew
Cool. to Because you're a cheapskate.

47:19.24
Sean
One, was way more expensive. Yes. I'm a cheapskate. It was way more expensive than I thought. And two, I could not for the life of me book another meeting onto my calendar.

47:25.48
Andrew
Yeah, yeah, yeah, yeah.

47:26.68
Sean
This, this, yeah.

47:26.85
Andrew
It doesn't seem like a good week for you to be booking more meetings.

47:30.16
Sean
Yeah.

47:30.69
Andrew
I'm a little worried all my, like, clients, like, my client stuff and my Metamonster stuff, the, like, user interviews I'm hoping to do for both clients and Metamonster are going to stack all on top of each other. And I'm going to have, like, one Sean week and want kill myself.

47:47.84
Sean
Godspeed.

47:49.90
Andrew
Yeah.

47:51.42
Sean
Cool.

47:52.36
Andrew
All right, man. Well, I'll catch you later.

47:54.30
Sean
I'll see you later.

47:56.87
Sean
Bye.

47:57.69
Andrew
Bye.