The Last 90 Days in AI Compressed an Entire Decade. Here's What Actually Changed

**Cody Schneider** (0:00)
I think the thing that people don't realize is like, if you can do it on a computer, an agent is going to be able to do it.

**LGD Set** (0:06)
What's up, everybody? It's LGD Set here, and welcome to Milk Road AI, the almost daily AI show that checks their LinkedIn account, but honestly, can't tell who's real and who's a bot on there. Today is June 22nd, 2026 We're recording on June 17th. We all rave about these AI tools, but I also feel like we all use them in a basic way, and don't actually push them to their limits. Well, our guest today has gone viral for his use of agents to the point where he's been able to help turn months of marketing work into just a few minutes of effort using a lot of agents. Cody Schneider, the head of graph.com, is here to tell us which tools and companies matter the most in the AI race, and give a little insight into how to use these things to their Macs, and are companies actually going to replace their marketing teams with AI sometime soon? It kind of sounds like they will. And also, if you're listening for the first time, or you're interested in getting a little bit more helpless, maybe some of your AI scope-age, some of your investment ideas maybe, our Milk Road Pro team is doing 30 trades a month right now. They're on absolute fire. They're doing a lot of selling, doing a lot of buying, really invested into the AI build-out, so you can go see what they're up to, their portfolios, or watch us all in Milk Road Pro. The link is below in the description. Today's episode is brought to you by Nexo, earn interest, borrow and trade crypto in Bitget Stocks 2 with real liquidity, real dividends. Cody, what's up, man? Welcome to the show.

**Cody Schneider** (1:24)
Yo, thanks for having me, LG. Stuck to be here. Been looking forward to this all week.

**LGD Set** (1:27)
Okay, dude, who are you, man? Okay, listen, Mr. Vyral Marketing Guru, telling people how to do this stuff. I feel like you're such a busy man. Phone ringing off the hook. Everybody wants a piece of what you guys are doing and trying to figure out how to do their marketing using agents. Tell me a little bit more about that.

**Cody Schneider** (1:43)
Yeah, so high level, just quick background. I was early at a company called Rupa Health, where a lot of what we were doing was basically how do I build out every channel simultaneously and also build out media arms of the company. We ended up taking that from a $20 million valuation to 110 in about six months, and then that exited in 2024 My co-founder and I, Max of Graph, we met at Rupa and basically had been, like since whatever it was, July of 2022, had been experimenting with, like how do you build AI tools for marketing? So had been working on these products for the last four years. And then, yeah, ended up working on Graph. So Graph basically, we forward deploy engineers to implement marketing agents for companies.
When I say a marketing agent, imagine an agent that entirely runs your Google Ads, your Facebook Ads, your cold email, your LinkedIn DMs. We're doing TikTok real farms. We're doing Instagram DMs at scale. Like you name a channel and we probably have built something at this point. What this looks like and how companies are using us is we're a data pipeline, a data warehouse, and then also the agent infrastructure.
So we are basically connected to the live business data. The agents are making their decisions based off of this. And when you think about like, compute is going crazy, all of this is like the physical infrastructure that's necessary. This is like what boots on the ground looks like. It's like actually getting these virtual employees into the wild making decisions for these organizations. And in contrast, it's like say, you know, historically, you would try to go and ship whatever, 10 channels over the next six months, we can get that same amount of channels or distribution out the door in weeks. So anyway, yeah, happy to answer any questions about that or dive deeper on anything.

**LGD Set** (3:33)
Why don't you flex a little bit for us, man? What's your token bill every month?

**Cody Schneider** (3:37)
Yeah, I don't even know, man. We have like part of it subsidized by one of our seed investors. We have like a deal with them in the US. But personally, I think I'm doing, it's nothing crazy actually, it's only around like 5K or something like that. But yeah, it's really interesting. If people are flexing that with state-of-the-art models, there's this part of me that's like, do you actually know what you're doing? Because more and more what we're doing is we're building the agents with state-of-the-art models, and then we're using open-source models as the daily marketing operations agents. So for example, it's like, hey, I'm going to actually go and build out the code. What is an agent to? I mean, just to go high level real fast. Everybody's talking about this. What does this actually mean in practice? An agent to me is basically like code under the hood, with some type of thinking loop that's involved, with live data that's connected to it, where it has a decision loop that's happening on top of it. So in my world, I've been head of growths and startups, etc. What we would do historically is we would go and we test a bunch of things, we'd look at what works, we'd do less of what doesn't and more on what does. We're just taking that same strategy, that same process. But instead of having humans be the operations components, we're having code and inference basically be the decision makers on top of that.