Stay ahead of AI. Get our free weekly briefing.
The playbook for launching new AI products is being rewritten in real-time. Coordinated PR blitzes feel like a bygone era, with more companies following Lulu Cheng Meservy's manifesto of "going direct."
Today's most successful AI products spread through a different pattern – one that starts with developer enthusiasm, gains momentum through tech media, and ultimately reaches mainstream professionals through a powerful combination of authentic discovery and strategic distribution. This new model is transforming how AI labs increase developer adoption and awareness, acquire users, and build sustainable businesses. You can see it at work in the most unexpected places – like my family's group chat.
A few weeks ago, my younger brother sent an unexpected message: "Have you tried Deepseek? Just downloaded it - pretty impressive."
My family group chat is my litmus test for most AI products. I usually hop in to ask my mother and brother their thoughts on different products I'm playing around with, or new features from ChatGPT.
So when my younger brother posted about Deepseek I was pretty surprised.
My family's interest in Deepseek highlights a crucial shift in how AI products spread – particularly in how they reach what the tech industry calls 'prosumers.' These aren't your typical consumers or enterprise customers, but rather professionals who use software to solve work-related problems. While their technical aptitude and expectations follow regular consumers, their needs remain business-focused.
This market has exploded beyond its origins in the startup world – today, it includes virtually any knowledge worker using AI to enhance their productivity.
This evolution in the prosumer market coincides with the obsolescence of traditional customer acquisition strategies. Almost all customer acquisition strategies from the SaaS era of software—both in enterprise and consumer—are outdated now. Competing on paid ads on platforms like Google and Facebook is an expensive game to play. Relying on social media referrals for growth is hard when there's no more authenticity in referrals and influencers are selling out for a few bucks. Even in enterprise, referrals are getting hard—any smart startup knows well networked people that can intro them to a potential customer. From the customer side there are only so many vendors you're going to use; even if your best friend started a new competing vendor, the significant switching costs often outweigh the benefits of trying a new vendor.
The first stage of this flywheel is on X, which has became a massive information source for most of the world.
Usually what happens is that there's something really interesting/cool/exciting/groundbreaking that goes somewhat viral on X. In the Deepseek example, it was primarily the inaccurate information around Deepseek's AI model cost $5.6 million to train, and that its reasoning model was on par or better than OpenAI's O1, at a fraction of the cost. But the key part here is that most of the virality comes from genuine and authentic posts from developers. Everyone is now keenly aware that developers are the ones that are building any software ecosystem, and because of that smart observers closely track what's trending with that group.
X is becoming a major information source for media publication. It always was, to be honest; there used to be journalists that would just write up stuff that was trending back in the SEO days, trying to tap into the curiosity around trending topics.
But all this got accelerated dramatically after Elon's headfirst dive into politics. Between his posts and Trump's history of using social media as an announcement platform, the mainstream press started to cite X as a source more and more. Now, if something's trending massively on X and capturing the X zeitgeist, you can be sure that media outlets will pick it up. It might not be the CNN's and Wall Street Journal's of the world, but the TechCrunch's and The Verge's certainly will. Not to mention the hundreds of creators with their own audiences following suit too.
Not everyone is going to learn how to code or be salivating at the mouth to try out the newest AI tools. If mainstream, normal, people were to have that curiosity they'd just be in tech. Most people want to only learn about the cream of the crop—the top 1-5% of products that are actually making a difference. The assumption is, if its going viral on X and a publication is writing about it, its worth checking out.
Developers quickly capitalize on the growing consumer interest in different technological innovations. The open-source nature of AI makes this easier too; companies can easily add new functionality and swap out models quickly. While brand new products are a bit more difficult to conceptualize and build (we still haven't see much around Anthropic's Computer Use API, though it came out months before OpenAI's Operator), there's an initial wave of developer adoption.
ChatGPT's initial launch and GPT-4 is really the prime example of this. It generated a ton of hype on X, garnered media interest, captured consumer mindshare, and OpenAI was able to use that to catapult ChatGPT to the top of the App Store and millions of users. It worked so well because a) what they released was novel b) it was really impressive for developers and the average prosumer c) they had a massive first mover advantage.
This is a really interesting case study. Anthropic's Computer Use API felt like a massive innovation and a push for Anthropic to take the lead on the AI Agent revolution. It came out months before OpenAI released anything and, while it was launched as an API for developers, it felt like this would inevitably be added to Claude.
OpenAI's $200-a-month tier is an interesting wrinkle in the flywheel. Operator and (especially) Deep Research both captured a lot of attention on social media and media publications but it hasn't translated to ChatGPT rocketing up the App Store.
In the AI era, the Developer-Prosumer Flywheel isn't just another marketing tactic. Unlike the old days, when consumer and enterprise software traveled down separate channels, AI is rewriting the rules. Here, genuine developer passion and real-world prosumer needs combine to create a self-sustaining cycle.
The most successful AI labs know that you don't win by throwing money at ads. Instead, it's about building products that naturally spark excitement and drive organic growth. As this flywheel gains momentum, developer enthusiasm becomes the leading indicator of mainstream adoption, blurring the lines between consumer and professional tools.
In conclusion, the Developer-Prosumer Flywheel isn't simply a new distribution strategy—it's a fundamental realignment of how breakthrough technologies penetrate the market. For AI labs, this means rethinking everything from product development to go-to-market strategy. Embrace this model, and you're not just capturing attention; you're reshaping the very landscape of technology adoption.