Vrushali (00:03)
So here we are. Thank you so much for tuning into this conversation. I'm Roshali. I lead social efforts here at Rio.dev. And today we're discussing a very interesting topic. So the way developers discover, evaluate, and buy software is changing. And at the pace of, let's say, every time we blink, the anthropics and Geminis of the world dropping updates almost every day. Developers changing how they work because of it. AI providers building even more for that behavior. It's a loop that's honestly
it doesn't look like it's going to slow down anytime soon. And the companies who are selling to developers and technical buyers in the middle of all of this definitely need to adapt their ways of going to market. And how do you exactly do this is a big part of our conversation today. So two founders with me who are both building for this new reality, Zach, co-founder and CEO of Gravity AI, a platform that surfaces contextually relevant ads inside AI assistance and coding environments, and Achantya, CEO and co-founder of Reo.dev.
a GTM intelligence platform that helps developer-focused companies find engineering teams who are looking for a solution exactly like theirs. So, welcome, Zach and Achunthia, and thank you so much for taking your time for this.
Zach Oldham (01:12)
Thank you. Hi. Yeah, sure. mean, you summed it up at a high level. think ⁓ really our goal is to make AI more accessible. ⁓
Achintya Gupta (01:12)
Pleasure.
Vrushali (01:14)
Yeah. So Zach, let's start with you, right? So can you tell us a little bit about Gravity AI?
Zach Oldham (01:32)
⁓ Ultimately, think accessibility means free. ⁓ And we've ⁓ seen information become free, the internet be free. And most of this has been powered by advertising. ⁓ But lots of the advertising isn't that great, in my opinion. It's not very tasteful. ⁓ So I think ⁓ especially if we just talk about ⁓ coding with AI today, ⁓
⁓ It can get quite expensive quite quickly, ⁓ especially if you want to use tools outside of like the cloud codes that are heavily subsidized. ⁓ And that really means that it's really only usable by companies, startups, and rich people. ⁓ And so as a result, don't think nearly as many people can use it. So our kind of approach to solving this is...
Now, A, we think we need to maybe pay around a trillion dollars a year ⁓ by 2035-ish to AI companies to actually make AI truly accessible, similar to the way that Google Search was and the internet was. And B, we solve that by building actual infrastructure itself. So I think our approach is focused on the infrastructure that is available to companies, whether they are advertisers or platforms. ⁓
And yeah. ⁓
Vrushali (03:02)
got it very interesting and like how long have you guys been building this out and can you talk us can you talk us through a little bit about how did it happen for you guys how did you even come up with the idea what what was going on yeah
Zach Oldham (03:13)
Mm hmm.
So I think we're eight and a half months in, we started in July of 25. ⁓ How it happened is, ⁓ I come from the world of like performance marketing. And I, the last company I built was like ⁓ tech focused performance marketing. ⁓ We delivered performance marketing services to like these like consumer brands, you think like online retailers. ⁓
And I think, I was, I ran that business for six years or so grew it to 50 people. like a, I was, I had spent a couple hundred million dollars on like meta Google ads and these platforms hadn't changed very much in a while. and, and so, ⁓ I was getting frustrated with them and I felt like there was a lack of innovation there. ⁓ and, and, and at the same time, there was like the tsunami of AI that was arriving.
And it was clear that it was coming for advertising and I didn't know exactly what that looked like at the time. So I started like spending time in San Francisco and ⁓ meeting people and just tinkering at the same time, like through my last company, tinkering with different ideas. ⁓ And it eventually became clear that like, this is like a really big problem with AI itself. And there are not many people in the world that actually know how ad tech, ⁓ ad networks work.
deeply and especially how they actually deliver results for brands, which is the thing that's going to make advertising sustainable. It's not just like, hey, let people run brand campaigns. ⁓ Everybody ⁓ with the ⁓ evolution of accountable advertising, you can spend money and see return and you can track that return. ⁓ And that's really important to brands and that's increasingly important to all companies. ⁓
it became clear that there weren't many people that could actually solve this problem well. ⁓ And my co-founder and I decided to kind of go all in, start a new company to do this. And that was July.
Vrushali (05:23)
Got it, very interesting. And what kind of people are you seeing, what kind of coding assistants are you seeing working with Gravity today? How has their experience been so far? What are you seeing in terms of how are developers reacting when they see ads inside their coding workflows? Do you want to comment a little bit around that?
Zach Oldham (05:32)
Mm-hmm.
Yeah, I'll give an example of a couple. ⁓ There are several, but one, the first ad program in coding ⁓ assistance that really existed was AMP code. ⁓ And we initially supported their program alongside their direct advertiser relationships and eventually ended up being responsible for 100 % of it. ⁓
Another example is like CodeBuff. ⁓ I think ⁓ that their free product is launched or maybe not yet, but will be. ⁓ And it's really cool. It's like genuinely you can have free coding supported by ⁓ advertising ⁓ on the topic of like how people are reacting. ⁓ I don't think everyone in the world likes ads, ⁓ but they still use Google search, they still use Instagram. So it's like, think people
people ultimately like the product more than ⁓ significantly enough to use a product even with advertising in it. ⁓ But I've been very surprised at the response. There's a couple of things that we've seen. A, there's not a single platform right now that has lost users because of advertising. So retention has not gone down. In some platforms cases, it's increased, which is kind of weird to me, but it's happened. ⁓ And then the second thing is ⁓
We've seen several posts on social media, dozens of posts actually, with people screenshotting the ads and going, this was actually really helpful. I didn't know Cerebris had all of these products. ⁓ Because the ads are designed to be relevant to what you're actually working on in the time. ⁓ through some clever deployments, we've been able to have ⁓ the coding assistant kind of like.
give us context on what the code base is, what the tech stack being used is. And then on top of that, an understanding of the conversation. The ads can actually show you companies or solutions like Rio, for example, that are actually relevant to what's happening. And if you think of it from the context of a coding agent, you're installing stuff a lot. ⁓ And there's often solutions that exist to your active problem that you were not aware of.
So still facing those solutions has had a really good response from users.
Vrushali (08:10)
Got it. There are a lot of things that you mentioned that I want to double click on, but I'm putting a mental pin on that for now and we'll come back to those in a minute. ⁓ Also, I'm really curious, right? how did this, ⁓ how was this received when you pitched this idea to your investors? Do you want to tell us that story? Because it sounds very exciting because you're building this out and it seems like
It would be such an amazing opportunity for people who want to advertise in just the relevant space that would ⁓ see a lot of positive reception. So I just want to understand. Yeah.
Zach Oldham (08:46)
So actually, before doing this, my ⁓ co-founder and I were working. We had gone pretty deep into trying to essentially automate advertising. ⁓ And I was new to the world of fundraising, because I bootstrapped the last company. ⁓ And ⁓ I didn't even know what a seed round was. Seriously, I understood no terminology at all. ⁓
So I took a few meetings for that last company and ⁓ for that last idea and everybody thought it was a really horrible idea. I was like, interesting. And, you know, like, you know, some of them had some great insights, some of them had some terrible insight. ⁓ But nevertheless, like, it was definitely eye opening for me about how the world worked.
⁓ And then ironically, actually, when we started working on this idea, I didn't go to like fundraise or anything. I had invested into the company. So we had funds and we were, we really just focused on like integrating with platforms, kind of like starting the network itself. ⁓ And so ⁓ to this day, guess like tech, I've spoken to investors, but I haven't really pitched many. ⁓ Our seed round, we got preempted for.
So a fund called us and basically said, hey, we'd like to invest. And that kind of triggered lots of other funds looking to meet with us over the next week, like the preceding weeks. ⁓ And so it's been a very interesting dynamic where it's like, think from an investor's perspective, ⁓ especially a venture investor's perspective, there's a lot of room for this to grow. ⁓
And there's parts of it that are very easy to understand. ⁓ So I've been surprised at the response. It's definitely being overwhelmingly positive. ⁓ I've had dozens of firms ⁓ ask to invest or try to get meetings with me to invest. ⁓ ultimately, think people understand that it's a difficult cold start.
It's a really big problem to solve if it works. ⁓ And there is no limit to the amount of capital that we could bring on and effectively deploy. ⁓ To date with what we've spent and been able to return ⁓ from a strictly business revenue perspective ⁓ or revenue opportunity perspective, ⁓ the economics of this model have been really strong. Yeah.
No. No.
Vrushali (11:38)
got it. Very interesting to see
how this plays out. It's really interesting to see. ⁓ I love that. It's amazing. ⁓ And now to you, Chhantya, right? So for those who don't know, can you give us a little bit into ⁓ InsideViewofYou.dev and how you got here?
Achintya Gupta (11:54)
⁓ Sure. I mean, Rio.dev is kind of like a GTM platform for companies which have technical mind journeys or which focus on developers. ⁓
So in a way, what we do is we use AI to interpret millions of activities and signals across open source, first party data, product trials, third parties, and so many sort of, you signals and convert that into a sales pipeline for customers to tell them that, okay.
These are the organizations accounts, these are the people, these are the teams who are evaluating you and this is the right time to get in. So we sort of add a lot of context to buying journeys and tell our customers this is the right time to actually focus on and that's what makes all the difference. ⁓ Yeah and I come from a background where I have built dev tools, built in for software products and sort of I mean always saw this problem that technical buying journeys are fundamentally different and
the realization was that whenever the buying journeys change the GTM software changes right and yeah that's that's that's kind of the change or the wave that we are trying to write
Vrushali (13:06)
Right, pretty cool. It's amazing to hear both of your origin stories, especially yours, Zach, and yours as well, Chantya. It's emerged out of a personal edge in a way. Right, so ⁓ I'm excited to take this to the next part, right, where we dig this a little deeper into how you guys seeing the developer behavior shift happen.
Zach Oldham (13:07)
Okay.
Okay.
Vrushali (13:26)
So my first question is to you, Achanthia. ⁓ I'm directly going to address the elephant in the room, right? So the rate of change of
how AI has been influencing developer behavior is absolutely insane. And so what's your comment on this, right? Like, how are you seeing? How have the ways and channels through which developers discover and buy software evolved, in your opinion?
Zach Oldham (13:36)
Okay.
Achintya Gupta (13:47)
Yeah, so I think developers are at the forefront of most of the modern software buying journeys, right? Like how developers behave is probably
how others behave couple of.
months years later right so if you see fundamentally the the search and discovery of tooling has changed a lot. mean the journeys that you start from communities or from search tools or search engines is now coming to sort of you know
Zach Oldham (14:08)
Okay.
Achintya Gupta (14:19)
agents, chatbot and those interfaces where people are discovering the tools. So essentially, mean, know, a lot part of the top of the funnel of mid of the funnel of an evaluation is moving into into chat based interfaces or AI tooling. yeah, I mean, it's it's it's a fundamental
shift. And I think I mean, today, starting with DevTool companies with tomorrow.
Zach Oldham (14:37)
⁓
Achintya Gupta (14:42)
any software company will have to think and consider that change because the other buyers will also start ⁓ behaving similarly.
Vrushali (14:52)
Got it, makes sense. Zach, is there anything else you want to add to this? Are you seeing the shift in your perspective in a different way or?
Zach Oldham (15:02)
No, I mean, ⁓
I think that's pretty spot on. It's definitely shifting quickly. I'm actually really curious, like, ⁓ for you, Achintya, like how you guys have kind of thought about your roadmap, like over the coming years. Like lots of the CEOs I know, including myself, of like even in this year alone, like with the improvements that have come with Opus and Codex, ⁓ like... ⁓
just the speed of being able to ship has really changed things. like lots of people have been overhauling their roadmaps entirely. you know, we've done the same thing in several ways. So I'm curious.
Achintya Gupta (15:46)
Yeah, so for us, for us actually there are two things. One, there's a change in the, on the signaling side a lot, right? So ⁓ because, you know, like Rio is built on the thesis that dev buying journeys are different. And that is like every month it is becoming more and more different, right? So that's, that's great, right? I mean, there are new places of sources of intent and we are capturing that like, so, so that
Zach Oldham (15:55)
Listen.
Mm-hmm.
Achintya Gupta (16:15)
that thesis becomes very strong almost I mean like how this change is happening. The second thing is that GTM is so much impacted by this whole sort of AI right now and that is great because we, one of the biggest things that we wanted to do is look at these patterns or developer activities and predict what's the right time to get in and probably four years back Rio was not even possible.
Zach Oldham (16:26)
Mm-hmm.
Mm-hmm.
Achintya Gupta (16:45)
there are so many signals out I mean if you just try to logically do that without AI right you will not be able to sort of predict or you might be able to predict it with with with less certainty but now sort of so for example
you know, we analyze if a team is evaluating a customer, we analyze what are the CLIs they are, they are sort of, you know, tracking or they're trying out, right, you know, and just understanding the patterns of how these CLIs are doing and how people are logging in into the product. It's a very strong evaluation intent that we give without the customers or customers knowing that, somebody was actually evaluating like, like five engineers at Tesla's were actually evaluating them. So, so now with AI, a lot of these things are possible. So it's a great
Zach Oldham (17:24)
Mm-hmm.
Achintya Gupta (17:28)
time to be building something like this where we are predicting the right time for our customers to get in right because in DevTooling it's a big timing is all that matters in DevTooling right if you do it too early the buyer has not learned about that you know the teams are evaluating the product if you enter too late then probably the developers have even built something on the open source or they've gone to competition and now prediction of timing using the AI and patterns that's what we do has become so much more possible.
Zach Oldham (17:32)
Thanks.
⁓
Makes sense. Interesting.
Achintya Gupta (18:02)
Yeah, I and I think, mean, in fact, I was
also sort of, sort of curious in terms of like, how do you see, like, I mean, in fact, coming from sort of, you know, the...
⁓ the brand side or even from the developer side I still see there is a lot of sort of you usually but you have a you're a three-sided platform because on the third side there is also the the code or the IDE right and you know how
Zach Oldham (18:30)
Yeah.
Achintya Gupta (18:30)
do you how do you work with them and how do you convince them of something like this right yeah I mean I was also curious about that
Zach Oldham (18:39)
so, ⁓
I think from a convincing standpoint, it's not so much as convincing versus really just a logistical question of where do your opinions lie? Is this something that you believe in for your company, or do you take a hard stance ⁓ against it? Because it really seems that like,
Achintya Gupta (19:17)
sense.
Zach Oldham (19:17)
There's kind of like two camps. either like, you're really against ads or you're kind of indifferent. ⁓ And it's the folk that are like, indifferent, where they're like, you know, I'm not strongly against ads. ⁓ They're like, okay, cool. there's actually probably opportunity here. And it's easy to look at like the history of advertising and go, cool. The model is proven. It's really obvious. It really exists. And then when we start looking at know, Cospo queries.
Achintya Gupta (19:23)
Yeah.
Zach Oldham (19:46)
and things like model costs dropping, but ⁓ the amount of tokens per task increasing, actually see like ⁓ cost per task is still increasing as harnesses improve. And so there are some of these underlying problems that the companies are looking at and going, OK, cool. ⁓ We can differentiate by being opinionated or like,
Achintya Gupta (20:00)
I agree. Yeah.
Zach Oldham (20:12)
having a strong thesis and really just building tooling that we like internally. And as a result, think you're seeing lots of beautiful outcomes across the, from anywhere from the anti-gravity ID to cursor to AMP code to factory to codebuff and to all of these different parts of this market. And it's a really, really massive market because we're not talking about just developers anymore. We're talking about people, like regular.
like people in adjacent industries or other professions being able to start using these tools and really benefit from them, ⁓ which is cool. ⁓ So ⁓ in terms of like integrating with them, it's really just like logistically, it's like, okay, cool. Does this make sense for your roadmap? If so, when and where? ⁓ And ⁓ do you do a free plan only? And is that free plan? Does it have capped usage or does it have unlimited usage? ⁓ Do you have... ⁓
like subsidized usage in a paid plan ⁓ or is it free plan where you pay some usage but it's discounted usage. It's like there are lots of different ways to kind of like split it up. ⁓ And so I think ⁓ that's like the more technical stuff but even then it's relatively easy. So beyond that, it's really just like what do the ads look like? Where do they fit? Does it feel good for the platform? Is it actually helpful? Because if it's not,
Like I don't want to work with those companies either. So there is like a bit of a, you know, a joint decision where it's like, we have to believe it makes sense for us to invest in and like put resources into. Um, um, because, uh, you know, in the short term, was a lot, like, there was a lot of work for us to do for every platform. Um, and, and especially when it comes to code bases, like there's a lot of like safety and like privacy considerations that we have to bake into the product. Um, so we have to be very like mindful about like how, how, how we handling data and like.
obfuscating enough data so that there's no risk to a developer using the product in the first place. ⁓ So it's been interesting, but overall, it is relatively simple. It's like, we have an SDK. You integrate it. ⁓ You make money. ⁓ From there, you can choose what to do with it. Yeah.
Achintya Gupta (22:27)
Of course.
Vrushali (22:32)
It's interesting how do you bring that up, because I was going to ask you, ⁓ can you give us a little bit of insight into how does this really work? If I'm, let's say, building a payment integration or something in my coding assistant as a developer, does Gravity AI enable the partner to show them a StriPAD in that moment or how is it actually set up? Can you give us a little bit of a view into that?
Zach Oldham (22:43)
Mm-hmm.
Mm-hmm.
Correct.
Yeah, so we're using ⁓ conversational signals to identify ⁓ what would be helpful in this conversation. ⁓ then ⁓ we look at active advertisers. And it's like, OK, who is most relevant? And do they meet a certain relevancy threshold? have. ⁓ And if they do, if that's met, I'm giving away some of our system here. ⁓
⁓ If it's met, then we're selecting the advertiser and we're injecting it. ⁓ like, we do all of the verification on our side ⁓ and we meet any platform requirements that exist, like whether there's a relevancy as a consideration or like how they want ads worded or structured. We handle all of that. And then this is all happening in real time. So it's happening in like half a second or so. ⁓ We've actually had to build a very
complex and very fast optimized system that scales, especially with the amount of conversations that we're processing. ⁓ the first few months, that was a very difficult problem for us to solve, was how do we get incredibly low latency when we're running a system that has several models ⁓ all talking to each other? ⁓
Yeah, I mean, like a high level, essentially, like we we're selecting the advertiser, we're determining how much the ad unit is worth and ⁓ and then paying that to a platform.
Vrushali (24:36)
got it. Aachhanta, do you have any questions there? Otherwise, I have something that I want to ask. All right. Cool. So Zach, you mentioned before, right? Like, when it comes to especially developers, first of all, there's very less people who want to see ads. But when it comes to developers, they are especially a little more allergic. But the kind of environment that you have facilitated, it's relatively, from what I understand, a little more contextual, right? Because it's learning from conversational
Achintya Gupta (24:40)
No, I think yes, let's go ahead.
Zach Oldham (24:56)
Mm-hmm.
It's all like.
Vrushali (25:06)
and the system is kind of picking up an advertiser and making it a little more personalized for the experience. ⁓ So, but how do you see it though? Are you what I mean, how are you seeing this work? Right? So, ⁓ mediums of advertisements that that would that developers would see ⁓ versus this, how are you seeing a difference?
Zach Oldham (25:19)
Mm-hmm.
⁓
So.
I mean, I haven't run lots of ads specifically for developers on other platforms. ⁓ mean, technically everybody is a consumer. So I've run ads to like pretty much, I think like everybody in the world has seen ads that we've ran at least like five to 10 times. ⁓ If I do the math on the spend. ⁓ So we've shown ads to lots of different people. think.
Just for developers specifically, I understand why they're particularly allergic. ⁓ objectively, most ads really do suck. They don't look good. They're not even very intelligent. ⁓ And that's a very painful thing to anyone that's inside of a co-base. A, you usually have quite high standards.
you really care about accuracy and when you see ads that aren't relevant or are just ugly, it hurts the soul a little bit. I think that this context is interesting because it is inside of the tool which you're using. And then at the same time, these aren't big disruptive ads, not like banners or something that'll pop up, they're not particularly interrupting you, they really are designed to be like...
⁓ we saw that you were working on this. Not sure if you're aware, but Cerebris has a solution to this. ⁓ Or you can debug race conditions with PowerFlare. ⁓ Whatever problem you're having, of course, an agent might solve it, or you might need to build your own solution, but there are pre-built solutions. And I think that's going to continue to exist for a long time, especially as ⁓
There are things that matter outside of just being able to strictly write lines of code. And there are reasons to continue using vendors in other companies. Maybe some SaaS apocalypse. Some SaaS things don't exist, but many will continue to. Most will.
Vrushali (27:48)
Got it. No, makes sense. Very interesting. And, Achanthya, do you have any thoughts on how do you see? Because you work with more than 200 plus developer companies, And DevTool companies. And in your experience, how do you see this playing out, especially the advertising ecosystem inside Coding Assistance? Do you have any thoughts on that?
Achintya Gupta (28:07)
Yeah, think, I mean, there are two or three ways of sort of looking at this. I mean, first of all, like as a brand marketer or say a dev tool marketer, I think I'm just...
contextual advertising is a great opportunity at least sort of in my opinion right because because hey there is nothing so contextual like that right I mean if somebody is in the at a place where they are trying I mean they are writing code or they are trying out things or discovering things and and and by the end of the day while doing this
probably if there is one place where say a developer needs more support is while they are building things and at that stage you know if they get something contextual actually you know there's something like that right so it sort of you know just fits the part of the puzzle. ⁓ So that's one way of looking at it of course you know the other is that yeah I mean ⁓ what we are seeing is that ⁓ the second thing as DevTool marketers is you know DevTool marketers are now understanding
Zach Oldham (28:48)
Mm-hmm.
Mm-hmm.
Achintya Gupta (29:11)
they are buying journeys much more than what was six months back, right? And they know that
How is it that developers are evaluating them? So for example, look at so many of them have actually taken their docs and exposed that as an MCP. So many of them are taking their products and expose this as an MCP. In fact, we at Rio have now launched an MCP gateway for their product and for their documentation so that we can extract the context and start sort of, know, using that as an intent signal. But ⁓ yes, I mean, there's a lot of awareness now coming in the brand or the
marketing or the revenue teams in terms of how the journeys are sort of know originating and they are trying to be a part of that. ⁓
Vrushali (29:58)
Right, makes sense, no. So I was gonna ask you these questions in the upcoming part, right? Like the next big thing that ⁓ even I was actually curious about was with all of this shift happening, especially ⁓ with the behavior and so many updates that we are seeing in new models and how are ⁓ AI providers kind of changing their ways around it. ⁓ If I'm a DevTool company, especially somebody who's sitting at the GTM, right? ⁓ How do I kind of...
Zach Oldham (30:23)
Okay.
Vrushali (30:28)
prepare for this change. And one of the ways like how you have talked about it, Achanthi, is that, you know, with more and more queries kind of moving into IDE from docs and all, one way of
Zach Oldham (30:40)
you
Vrushali (30:41)
it is definitely like understanding where is this agentic discovery and agentic evaluation kind of coming from. But if you had to like, you know,
Zach Oldham (30:41)
Okay.
Vrushali (30:50)
Okay, so let me think about it this way. So you have sold to developers in technical bias before. If you were in that position today around the similar landscape that we are seeing today, like where would you start? Because it is really overwhelming with the pace and all, right? So how would you look at that?
Zach Oldham (31:01)
Thank
Achintya Gupta (31:05)
Right, so I think for it,
it has changed every part of it. So for example, if it's coming to sort of the discovery part of it, then as a marketer, they focus on AEO or sort of at the places where sort of if my product is being recommended in the chat or in the AI agents becomes so much more important, right? So my content strategy changes
the discovery journeys are changing. So for example like earlier the developers it was the humans or the developers who were reading documentation right. Yes that still happens right it happens a lot but that has moved towards the sort of you know bottom of the sort of evaluation funnel on the discovery part it's an actual agent that is reading a documentation or sort of you know trying out things in the code. So a lot of intent has moved from a human to an agent right and
Zach Oldham (31:42)
Okay. Thank you.
Achintya Gupta (32:00)
and and as as as a revenue person or as a marketer in a in a dev tool company yesterday I had tooling to understand intent from humans I have to now understand intent from agents right I mean because by the end of the day ⁓ sales needs to happen right so so yeah I mean it it just changes the perspective of things and and it sort of you know makes
Zach Oldham (32:22)
Okay.
Achintya Gupta (32:27)
makes so much more sense that once they start understanding this journey they start asking the right questions in terms of how do we solve for that.
Vrushali (32:39)
recorded. make sense. Yeah, please.
Zach Oldham (32:39)
If I might add, if I might add
as well, it's like the world of like consumer online retailers, like DTC advertising marketing is incredibly optimized. Like mostly because the channels for marketing have existed for a long time for those companies.
And they've really performed well for those companies. there's like hundreds of billions of dollars of ad spend that's like marketing spend that is poured into those, like various marketing channels for these companies. ⁓ so it's incredibly competitive. There's tons of tooling that exists. There's lots of different options on the market, but what's interesting is like dev tooling historically has, has, has not had as many great, like, like great ways to.
to advertise, to do marketing. ⁓ you know, companies like Rio, like you guys have existed. ⁓ But I think, like, it simply has been much harder for them to spend money and see it ⁓ return well. But I think that's starting to change. you know, like, with everything from, like, GEO to, like, what we're doing ⁓ with advertising ⁓ and just the way that AI is, helping discoverability.
Achintya Gupta (33:46)
Yeah, I agree.
Zach Oldham (34:00)
⁓ It's cool because I think the world of dev pooling is going to see this optimized performance marketing world start blending into dev tools, which ⁓ I guess positions you guys at RIO really, really well for this. yeah, there's going to be so much more that you, just simply because you can spend more money and actually get good return from it, there's going to be so much more focus. There's going to be more people working on it. There's going to be more companies needing
are signals like the ones that you guys provide to actually ⁓ understand their marketing journeys, ⁓ to actually really deeply improve these things as well. So I think we'll see that the world of online marketing and digital advertising ⁓ blend ⁓ really well with DevTools as a category.
Achintya Gupta (34:53)
Yeah, and I see that, mean, like these companies, these teams actually want more revenues to be able to market, right? And sort of, so there are two very interesting things that's happening. from, I mean, from on one end, the very definition of a developer has changed in the last six months, like, know, so of course, like how people who were not sort of, you know.
coders themselves, perhaps now can use actually these tools to understand schema, build things, right? And that's where sort of there is a massive expansion of time for a lot of dev tool companies, which earlier would not have existed. A lot of non-dev tool companies have started behaving like data pipes because, or pass because now their customers have started sort of using any of these tools like plot code, et cetera, to write simple lines of code and sort of
Zach Oldham (35:23)
Mm-hmm.
Achintya Gupta (35:46)
extract data from them as a simple API. So the definitions have sort of expanded a lot for these companies. At the other end, sort of the avenues of marketing have sort of reduced. Like for example, you look at communities, right, I mean, look at a lot of places where these developers were actually reading and learning content. Now they are just on an agent or on a tool, like an AI tool and learning from there. So it's a very exciting time for everyone.
Zach Oldham (36:15)
Very. Thousand percent.
Vrushali (36:17)
Right, ⁓
Sorry, were you saying something, Zach?
Zach Oldham (36:22)
No, no, no. Go ahead.
Vrushali (36:24)
Got it, got it, no. So Ajinta, you mentioned something very interesting as well, right? So the definition of developer, et cetera, that reminded me of, ⁓ you know, how the decision-making processes also influenced from role to role. I remember you had shared like, ⁓ and a very interesting thing about how the role of developers is kind of ⁓ changing in enterprise tech sales, right? And how that might have certain implications as well. Do you wanna talk a little bit about that?
Zach Oldham (36:39)
you
Achintya Gupta (36:54)
Yeah so I mean what's happening is that if you look at dev tooling or infrastructures or software essentially what you what you are selling is like a nut and bolt of an engine which is your software right and and that nut and bolt if gone wrong can actually crash that entire engine so so that nut or nut that bolt is so important that
Zach Oldham (37:04)
Excuse me.
Achintya Gupta (37:21)
the evaluation of that often goes to the practitioner who knows how that works right and that practitioner will have a lot of saying which tool would be used for building ⁓ that piece and it can be like AI infra or data infra or any kind of know tooling like you
DevOps and so many other places security right you know these are so critical and so underlying for software and and ⁓
And at the other end, it's, mean, what has happened in the last couple of years is that the software has become so composable and so complex in every industry that even the CTOs of the VPF in general have not been knowing how many tools are being used to actually compose their software. And hence that's the reason why even the largest of the enterprises or even the smallest of the companies, like, know, it's the decision of a tooling is going to the practitioner and that will happen more and more. I mean, it's,
a behavior that's going to get any lesser in fact looking at how complex software and how composable now it is even more becoming it's only going to increase.
Vrushali (38:32)
You got it, makes sense, no? Also, when you spoke about the avenues of marketing, right? I was thinking about this and ⁓ the question is a little unstructured, but ⁓ we are seeing agents being built for, mean, tools being built for AI agents, right? Like we've seen ⁓ AgentMill kind of build a tool where, ⁓ you know, ⁓ it's a software for basically, I mean.
tool for AI, it's an email tooling software for AI agents. We have social media platform for AI agents, which is taking us to, you know, what can you do to get on the radar of AI tools to kind of recommend you better. And, and this is a black box and we're still optimizing it.
as we go and I just want to understand if you had some thoughts there and how can kind of because there are so many articles that are being shipped right now on how cloud code chooses or how ⁓ there is a lot of recency bias that's happening. So there's multiple people are talking about different different things. So if there's any thoughts that you guys have that you would like to share.
Achintya Gupta (39:36)
Yeah, Zach, was, mean, what are marketers sort of thinking from that and like, does their product get recommended? I mean, that's a very interesting question.
Zach Oldham (39:47)
⁓
in the context of just being recommended in general by AI.
Achintya Gupta (39:53)
Bye-bye.
Zach Oldham (39:55)
⁓ I, I'm not inside all of their heads, but like from, from, from the conversations that I've had, ⁓ I mean, I think it's become, it's, it's become clear to companies, to, to at least like the top 10 to 20 % of marketers that, that, ⁓ AI will fundamentally, it's fundamentally changing the way that
that marketing works. ⁓ It's removing some of the advantages that previously existed ⁓ on ⁓ Google search, the classic SEOs, even the more optimized Reddit comment sections and Stack Overflow branded posts. ⁓
⁓ I mean, I think everyone kind of knows that, like, ⁓ people have started to see GEO and like really focus in on GEO. ⁓ But I think what's a good and a bad thing, but ultimately, I believe it's a good thing for the world is that ⁓ AI, in theory, will get way better at recommending the best possible product ⁓ and like specific for your need itself. ⁓
And so that's really great for the companies that actually do care about product. And it's really bad for the ones that don't. But ultimately that's a good thing for the world. Historically, one thing that I found a lot when I was working even with consumer brands is that I came across many really great products, but they were like small companies. And oftentimes they were led by people that did not know how to ⁓ increase discoverability of their products, or they did not understand marketing at a high level, or did not want to involve themselves in it. And they were higher firms like mine. ⁓
But ⁓ I just think that ⁓ distribution is still such a skill. ⁓ But I do think this kind of levels the playing fields ⁓ in distribution a little bit, where as more people use tools, ⁓ these tools are embedded into everything. Where you have agents and video editors and design tools and your Gmail and absolutely everything. Discoverability is only going to increase. ⁓
which means ⁓ it's a really positive thing for the people that have helpful solutions to their customers. ⁓ And so hopefully the world switches to the companies that didn't put as much of a priority on product will start increasing their priorities ⁓ with product. And we should see ⁓ far more innovation in general, ⁓ which is ultimately great for all of us.
Achintya Gupta (42:45)
I agree, yeah.
Zach Oldham (42:46)
Yeah. Yeah. Awesome.
Vrushali (42:48)
Very interesting.
Cool. So we're nearing the end and I just have a very interesting set of questions. ⁓ So yeah, I mean, with so much of wipe coding and AI system coding like spiking up right now, do you think it still makes sense to learn coding for young people, you know, whatever your answer is? Why?
Zach Oldham (43:10)
⁓ I'm curious what Achencha says. No, no, no, go, go. I want to hear your response.
Achintya Gupta (43:12)
Yeah, Zach, go ahead. Okay.
Vrushali (43:12)
Yeah.
Achintya Gupta (43:17)
So, so, I mean, okay, I think, I think the short answer is yes, I mean, and there are two reasons for it. I mean, let's, let's look.
Let's look at it, probably even sort of what's happening currently, right? What AI is doing is it made top 5 % developers like 500 % more productive, right? So yes, I mean, that means that people who are actually good at code or good at their job, right, are actually using it. It's helping them become better, right? So in a way, sort of the need for being good at this is not going away, at least from what the trends are showing right now.
The second is, I mean if you just look at historically right, I mean it's like yes you have calculators but then would you stop teaching multiplication to kids in schools right, I mean you know and that's that's a fundamental question I mean the answer to that if it's the answer is yes, the answer is yes, if the answer is there is no, then the answer is there is no. But yeah that's it's a fundamental correlation in that way.
Vrushali (44:19)
Okay, thoughts, Zach?
Zach Oldham (44:19)
Thousand percent.
So ⁓ I've tried to like play out in my mind like five, 10 years down the line, 20 years down the line, what happens with AI. ⁓
And ultimately, think maybe the paradigm will shift again if everybody has a Neuralink chip in their brain and it knows exactly what you're thinking at all times. ⁓ But up until that happens, communicating with AI is still a problem and it's still a really important problem. ⁓ And I think ⁓ if you are not specific about things, then an LLM or an agent is always going to do its best to infer what you mean. ⁓
But you can, we've seen it over the last couple of years and it will, this will improve a lot. But even over the last couple of years, you've seen it where it's like consistently people download Cloud Code or Open Cloud for the first time and they get it to make software and it works for them locally, but it doesn't, there's bugs in production. And now again, that's going to get better and better. Like that's going to, that will go away, but there will still be like a difference in like the quality of output that an LLM creates.
⁓ in all domains, like beyond coding, apply it to insurance, apply it to healthcare. ⁓ so knowledge in those areas is still ⁓ going to be important. ⁓ However, an LLM can help you get that knowledge too. I don't see a world in which there is not an advantage by ⁓ still being a domain expert, but I do think LLMs and AI can help you become a domain expert much faster.
⁓ But ultimately, if you can communicate what's in your mind better to an agent, then your outputs will get better. And that communication ⁓ seemingly involves better understanding. ⁓ If you understand how a system works better, then you're going to be able to, you can be more specific about the design and the architecture of it. ⁓ And those things are advantages in at least the near term, at least like the next two to three years. ⁓ Watching the evolution beyond that will be interesting. ⁓
But yeah, ultimately, even then, like I don't see a world in which AI knows exactly what you're thinking ⁓ without being embedded in your brain. ⁓ So I think there's like that paradigm will probably continue to exist until it gets flipped on its head by chips.
Vrushali (46:50)
Got it. And the next one's a fun, quick fun one, right? Like what were some of the tools or products that you've discovered or validated through an AI assistant? I'm just curious, because ⁓ I had done it the last time and I remember my, ⁓ just the process from when I started discovering a tool to when I executed it was just a couple of minutes. I was looking for a voice assistant and that's when I came across. I mean, I think in all of the...
AI tools that do amazing research. I personally feel that Poplexity is the best out there because of the way its algorithm is designed or whatever. don't know. But yeah, I I literally remember discovering
Zach Oldham (47:20)
Thanks.
Vrushali (47:29)
a voice assistant in a couple of minutes and implementing the use case I wanted to do and have a project ready by a couple of hours. So I just wanted to understand from you guys, like what was the last tool or an aha moment that you had that you validated something through an AI assistant or an AI chatbot? Anything interesting to share there?
Zach Oldham (47:50)
There's so many, I don't even know where to start. think, yeah, like, it's like, there's something new every day. ⁓ Like, more interesting examples, like getting stuck with, ⁓ you know, like, we were really trying to evaluate, like, a variety of different embedding models. ⁓
Achintya Gupta (47:50)
Zach.
Zach Oldham (48:10)
And ⁓ AI helped us ⁓ discover embedding models that we didn't realize existed, or embedding models that performed really well for our specific use case. So I think that's a ⁓ more niche example. And then helped us build benchmarking off the back of it and use a good benchmarking suite. So ⁓ that's one micro example of something relatively technical that actually really improved the output that we were getting.
from even just recommending different models from providers that we were already working with like Voyage.
Achintya Gupta (48:47)
No, mean for me, like as a founder or as tech founder, the last thing you want to do is to look into contracts many times, right, legal contracts and I mean that's, or this financial paperwork and that's, I mean, and yet you have to do because nobody else in the company will do, right. But yes, that has been the biggest help, like a lot of things where I wanted to know whether sort of this contract covers this, covers that, I think, you know, it has been really, really, ⁓ you know.
Zach Oldham (48:56)
haha
Achintya Gupta (49:16)
Apart from so many cases, ⁓ this is something that one I can remember. It has changed the way, as a tech founder, how I derive productivity.
Zach Oldham (49:22)
Mm-hmm.
Vrushali (49:30)
Got it.
Zach Oldham (49:30)
One of the interesting
discoveries from AI that I've had is books and stuff as well. I do still find it a helpful way to expand understanding. ⁓ And book recommendations from AI has been an interesting one.
Achintya Gupta (49:37)
Okay.
Cool, very cool.
Vrushali (49:48)
That has happened to me as well, ⁓ especially with the music recommendations. I recently found out that I like lot of rhythm and blues music and if I shared one of my favourite playlists is called Ion and Wine. And when I shared it with Claude, it was able to blurt out a lot of other artists who I had not discovered and was just so amazed to hear that out. So yeah, that was pretty cool.
And it's, it's, it would, you would expect it to come up in your Spotify recommendations, but it, has like never done. So it's crazy. So yeah, that anyway, so yeah, I had a last one. So ⁓ yes.
Zach Oldham (50:22)
Can I, Rushele, do you mind if we click this
and I interrupt you really quickly? I gotta go in five minutes, my laptop's about to die, so I need to plug it in. Let me grab the charger. I'll be back in five seconds. Sorry.
Vrushali (50:27)
Yeah.
Alright.
Alright. Yeah, no worries.
Zach Oldham (50:36)
⁓ alright.
Sorry, I thought I'd take advantage of you mentioning that we could clip that part. ⁓
Vrushali (50:50)
Yeah, I know. Please
feel free to do that. fine.
Zach Oldham (50:53)
Awesome. All right.
⁓ one second. Let me just see. All right. It should be good. I'm just going to monitor it. I'm not using like an actual MacBook charger. I'm using a phone charger. So I might interrupt you once more. Sorry. I'm good.
Vrushali (51:02)
All right. Sure.
Got it, No, that's all right.
Yeah, yeah. No, I think we'll also wrap up in another five minutes. I just have a last couple of questions. Yeah, great. All right.
Zach Oldham (51:22)
Okay, perfect. Thank you.
Vrushali (51:27)
So I have another last one, right? So ⁓ where do you see this space going in the next couple of years, right? Like with everything, AI native, discovery, GTM, way evaluation, buying is changing. How does it look like? And ⁓ if I may, like, I don't know if this is a good way to think about it, but somebody told me that when there were a lot of... ⁓
Zach Oldham (51:48)
you
Vrushali (51:52)
digital marketing changes happening, especially in the landscape when Amazon was building it out. ⁓ I'm not sure if it was their CEO or management had commented about like, you know, there were a lot of changes that were happening, but ⁓ their kind of principles were aligned on what are the things that they are seeing are gonna remain constant. So I think ⁓ if in the...
Zach Oldham (51:54)
which is not the right time to do this.
Vrushali (52:15)
especially in the AI native GTM space if we had to comment on what are some of the things that are going to remain constant which we can
Zach Oldham (52:21)
Thank Thank
Vrushali (52:22)
kind of rely on even with all of the changes around like how would we think about that if you were to comment on that? ⁓ Achinte, maybe you can go first and...
Achintya Gupta (52:32)
Yeah, I think a couple of things here. ⁓ One is, know, sort of, I agree, mean, know, products and startups should be built on things that don't change rather than sort of, you just things that might change temporarily. And...
One of the things that we fundamentally believe is that the buying journeys of software are going to become more and more technical in nature. And one of the very initial thesis on this actually Rio was built is that...
Zach Oldham (53:07)
Let's do it.
Achintya Gupta (53:16)
software become more and more composable and more and more people will behave like software developers right and that sort of seems to be coming true right so it's it's that we have very exciting time for a company ⁓ like us ⁓
Yeah.
Vrushali (53:36)
say anything you had to.
Zach Oldham (53:36)
Excuse
me. I agree with the chintio and the like, you know, we've touched on it, but I think we're seeing, ⁓ we'll slowly see more and more people become developers. And I actually think that will only increase. It's like with the irony of like, AI.
⁓ being the first replacing the people that built it, the engineers and the devs, ⁓ I think it only makes software more powerful. And it does make it more accessible to more people. So yeah, to kind of echo ⁓ Atin's point, that we will see way more people becoming developers. And in that sense, again, completely correct, ⁓ where
the buying journeys will start at the development level. ⁓ And so it's definitely an evolution of marketing for all software.
Vrushali (54:43)
Got it, makes sense. So Zach, what's it looking like? What's next for Gravity? What are some of the things that are coming up exciting for you?
Zach Oldham (54:53)
I mean, I'm very simple person. ⁓ We buy ads and we sell ads basically. ⁓ I mean, we will buy more ads, we will sell more ads. What that means is, we'll accept more platforms as advertisers. ⁓ We'll let more people off the wait list. ⁓
This is the same on the platform side. It's really like integrate with more platforms, let them start using this as a monetization channel. And beyond that, there's a lot of different ad units that I think will be very exciting that we'll start seeing. Ads that you can click on that go and have an agent buy the product for you or install something automatically for you. Yeah.
Vrushali (55:42)
Make sense. Attend the same question.
Achintya Gupta (55:44)
Yeah, mean, so ⁓ far it's a very exciting time. We sort of want to capture new sources of intent signals. That's one of the most exciting things that we're working on. And a lot more ⁓ is in pipeline today that's going to come out soon in terms of what are the places where we are. ⁓ Sort of, you know.
going to capture intent. We are working a lot on partnerships right now with specialized providers of sort of AI tooling where we can sort of capture this. ⁓ But yes, mean, the one thing we would really want to focus on is are we the best provider of GTM signals to anybody who has technical buying journeys, right? And yeah, we'll...
just keep on working more and more on it.
Zach Oldham (56:44)
⁓ yeah. Sure, in terms of, There you go. I'm at Zach the old ham on all platforms and try gravity.ai.
Vrushali (56:44)
Got it. Makes sense. ⁓ With that, is there anything else you want to say to anybody watching this, if you want to drop a line about where can people find you, et cetera?
Achintya Gupta (56:57)
Yeah, people generally find me on LinkedIn. ⁓
Yeah, yeah, ⁓ I'm on acente at theratria.dev, also on LinkedIn, where most of the people find me.
Vrushali (57:08)
But I think we.
Zach Oldham (57:10)
Beautiful.
Vrushali (57:17)
Got it. No, I think with that, we are at the end of this conversation. Thank you so much for taking out time for today. I absolutely had an amazing time. Yeah, that's all.
Zach Oldham (57:27)
Yeah,
this was fun. Awesome. Thanks guys.
Achintya Gupta (57:28)
Yeah, thanks. Thanks, Rishali.
Vrushali (57:30)
And for anybody watching
as well, thank you for tuning in. Yeah, thank you.
Achintya Gupta (57:34)
Take care. Bye. Thanks, Zach. Great meeting you. I mean, I finally got a chance to meet you.
