Welcome to Words with Wynn! If this is your first time perusing my content and you’d like more of my weekly musings, subscribe below:
The longer I work with startups, the more acutely I focus on distribution. It isn’t enough to have a great product if you can’t get it into the hands of your customers. The same goes for ideas. A lone genius in a village has little impact if her ideas can’t spread.
Some of humanity’s most impactful leaps forward have been accelerants for the distribution of information. The printing press and the internet (by extension, social media) were major upheavals for the interchange of ideas. If we think in terms of information as water and distribution as channels, these two inventions were equivalent to the smashing of dams. Each accelerated the flow of information in unimaginable ways.
Extending the metaphor, distribution widens the channels, but in order for rivers of information to flow there needs to be volumes of content pushing the current. So, what taps the source?
Creativity is the wellspring of ideas, and there have been a few brilliant technologies that have unlocked it at scale.
The best examples recently bend what I think of as the ‘arc of accessibility’. It’s the idea that consumer technology broadly arcs toward democratizing access, and some of the most successful startups accomplish this aim. For example, AirBnB’s and Uber’s breakout successes were in part attributable to their ability to capitalize on demand for premium experiences (accessible travel lodging, cheap personal drivers), while tapping latent resources (extra rooms, individuals’ cars). As it pertains to information, there have been a number of technology, gaming, and media startups that have risen to prominence by unlocking latent human creativity to empower individuals while creating content at scale. I believe that the explosive emergence of agentic coding platforms is the next generation of these disruptors.
Unlocking Creativity at Scale
Before speculating on where we are going, it’s important to consider where we’ve come from. The printing press and the internet fueled broad macro trends, unlocking myriad technological advancements. But as we approach the present bleeding edge, let’s consider some micro examples of what I really mean. A couple technologies that sit within my framework for unlocking creativity at scale are digital audio workstations (DAWs) and user-generated content (UGC) gaming platforms (i.e. Roblox).
Digital Audio Workstations (DAWs)
Digital Audio Workstations (DAWs) are programs used to record, edit, and produce music. These pieces of software are incredibly complex and computationally intensive. For context, producing a single track may require a number of individually taxing workstreams to be running in parallel (virtual instruments, data-heavy audio, effects, plug-ins, sub tracks, etc.). Hence, these tools weren’t really commercially accessible until the late nineties and early two thousands as the power of personal computers increased exponentially.
DAWs’ mass market viability rode the wave of Moore’s law, and their proliferation changed the game for musicians. These pieces of software significantly decreased the cost of music production while simultaneously expanding access to studio-grade tooling. It’s hard to quantify directly, given the numerous inputs in producing and selling an album (recording, mastering, mixing, etc.), but I think it’s directionally informative to note that Wikipedia’s list of most expensive albums are largely clustered around the time period right before digital music tooling was taking off (‘90s-’00s). The popularization of programs like Ableton, Logic Pro, and FL Studio empowered independent producers by dramatically reducing the cost to create commercially-viable sounds.
Historically, an emerging band had a minefield of costs to navigate in creating their first radio-ready tracks. These costs could have included limited and expensive studio time, audio engineering, mixing/mastering, and other nickel and diming across the value chain, not to mention the pains of physical distribution. With the advent of DAWs, a single skilled creator could produce a (roughly) commercially viable track in his bedroom at a fraction of the cost, a phenomenon especially potent in more digital-native genres like House music. This widened the aperture for aspiring creatives, and was supercharged by the internet’s distribution channels.
DAWs developed in tandem with the birth of the internet, which upended music discovery, sharing, and listening. Permissionless distribution like Soundcloud and early music blogs transformed virality in a new way, as emerging artists like Odd Future were far less beholden to label strangleholds on physical distribution. This has continued to evolve today as startups push the bounds of distribution as well as real-time collaboration (which has historically been a challenge with DAWs). A great recent example of this is [untitled], which Packy McCormick dissected at length.
UGC Gaming Platforms
If you’re outside gaming, it’s easy to overlook Roblox and Minecraft as simplistic toys for childhood entertainment. It’s underappreciated just how massive these businesses are and their importance in profitably pioneering UGC in gaming. For scope, Minecraft is the best selling game of all time with over 300M copies sold (DAU undisclosed) and Roblox reportedly clears over 70M daily active users. But what’s driven their virality and how did they unlock creativity?
Both platforms were launched in the early ‘00s and center around lego-like functionality where the ‘games’ themselves are sandbox style play environments. Players enter the worlds, and then have free reign to collect resources and build to their hearts’ content. Where most traditional games follow pre-ordained stories with defined checkpoints and goals, sandbox games like Minecraft simply drop players into a blank slate world and provide them with the tools and resources to entertain themselves. The open-endedness of the platforms leaves it to the user to decide how shallow or deeply they would like to engage.
For the novice player, often younger demographics, these games can be played as simply as “gather wood, build house, kill monster” and offer basic play patterns to mete out a meager existence mining or farming. But where they really struck virality is in the complexity that can be built thereafter. Beyond basic resources, the games offer deeper mechanics, enabling players to create mechanical and electrical components like pistons and circuits to build increasingly inventive contraptions:
The early inventiveness of their users, in combination with burgeoning video-sharing platforms like Youtube, helped to drive the viral adoption of Roblox and Minecraft. These blank slate game worlds provided simple, interactive environments in which players could express their own creativity, building everything from villages to assembly lines.
Roblox took this a step further with the introduction of the Roblox Developer’s Environment (now Roblox Studio) in 2006. This developer program allowed players to go beyond ‘in-game’ creation within pre-programmed sandbox levels or simply playing through content created by the Roblox team, and empowered independent developers to actually create their own games for the broader Roblox ecosystem.
But where it gets really interesting is in the monetizability of the games themselves. In 2013, Roblox introduced DevEx, which allowed developers to exchange the in-game currency (Robux) for real-world dollars, thus monetizing their on-platform creations. This supercharged developer adoption by aligning incentives between Roblox and its creator community, allowing independent developers to capture a slice of the value they contributed to the game’s ecosystem. Adept monetization created a massive flywheel effect, which was kickstarted by the already-captive audience on the platform, and has been highly lucrative for both developers and Roblox:
These two platforms have been trojan horses for unlocking gamer creativity at scale. Gamers of the early aughts would have needed coding expertise and heavy-duty computers in order to program games. Minecraft and Roblox empower amateurs to learn and build in a vastly more accessible way. Where a bedroom game developer would previously have had to code an entire game and figure out distribution, today they can learn in a seamless, self-contained platform like Roblox which also contains built-in distribution at massive scale (that 70M DAU we mentioned earlier). This spins a powerful flywheel for emerging and professional developers alike to hone their skills on a captive audience.
Interestingly, the gamified engagement that these platforms created has captured attention beyond simply recreation:
These two games have made learning basic programming, chemistry, and physics more enjoyable experiences in a way that educators have been attempting to crack for years.
Simply, these technologies both bent the arc of accessibility for everyone from hobbyists to professionals. The same leap forward is happening within software development today.
AI and Vibe Coding
For those of you that haven’t followed the AI arms race, things are getting pretty wild. ChatGPT and other LLMs were initially exciting to the average joe (read: me) for their robust search and chat functions. But where things are really exploding is the exponential improvement in agentic coding (coined ✨vibe coding✨). That is, the ability for AI to turn the text of your ideas into iterations of software.
Platforms like Replit, Lovable, and Cursor enable the utterly untechnical to go from vision to prototype with a mere penning of ideas. They have burst open the bastion of software development in a way that is hard to fathom. Historically, devs have been highly sought after and handsomely rewarded precisely for the scarcity of their expertise. The launch of these vibe coding platforms has created a reality where an unskilled troglodyte like myself can code with the efficacy of a first year comp sci student, all from the riffing of a few ideas in a text box. Sure, I can’t magic away the nuance and systems skill of a battle-hardened dev (yet), but I can whip up a shitty MVP of almost any idea in the span of a few hours. This technology has enabled the non-technical to become 1x engineers, and the highly adept to become 100x engineers. And, at risk of sounding over-exuberant here, you have to realize that these tools today are the worst they will ever be. It only gets better. Exponentially.
For example, I’ve had a passing interest in learning rhetoric. I minored in English Literature in undergrad, and one of my biggest lamentations is that I was never able to fit a Philosophy 101 class into my overburdened schedule. Hence, I’ve never had any formal education on the various schools of thought, and my love for Greco-Roman history always made me wonder what it would be like to learn the art of argumentation in a structured way.
Anyway, I’ve had that idea cooking as the ChatGPTs of the world began to improve their agents, and their personalities. But what really urged me to take the leap and build something for myself was Twitter. The longer you swim in that cess-pool of ideas, the more idiotic and vapid you realize the absolute clownery of discourse is there. Flawed reasoning and fallacies abound, which made me wonder: “Could I create an app or agent to teach me rhetoric like Cicero?”
A perfect little use case for testing these tools, and they did not disappoint. In few hours one Sunday, I was able to whip up a janky ChatGPT wrapper that would prompt me with debate topics, and provide a sparring partner and coach in a fairly robust way:
For the flow itself, Cicero’s a fairly simplistic little web app. It creates some randomized debate topics, then it’s effectively just an agent tapping ChatGPT’s API as it ingests your thoughts and itself responds (as prompted) like a rhetoric teacher of old. Like anything LLM related, it took a decent bit of prompt-crafting to refine exactly how Cicero surfaces debate topics (“Make them less dry, make it more topical, funnier, etc.”) as well as how it goes through the motions of debating/coaching you:
But what’s equally magical to me are the little things. The devil’s in the details around my little chat bot. Adding a timer with a corresponding visual bar at the bottom? Half a second. Generate a pixelated pantheon wallpaper? Half a half a second. Completely rework the flow to include back-and-forth with the Agent as well as better coaching characteristics? An abysmal three seconds.
Then you start cross-coordinating your AI’s..
“Hey ChatGPT, provide me with a detailed, structured outline of how to build ABC project. Below is my vision including functionality, aesthetic, and user flow. Please structure your response to be pasted directly into Replit.”
I know it’s a meme– but you really are limited only by your imagination and your agency. Where there’s a will, there used to be a way. Now, where there’s a half-baked stoner idea at 3AM, there’s an MVP. The cost to ship is collapsing, and it’s empowering an entirely new generation of solopreneurs and indie hackers. What the printing press did for the proliferation of reading and writing, vibe coding might just accomplish for software development.
This all converges on a thesis I stole from my chats with Brexton, the founder of Ohara: software will become content. The channels of distribution remain established, and they will soon be flooded with software not unlike the proliferation of video from smart phones or books from the printing press. Personalized apps, software as a meme, and any number of inventive applications may come out of this, with value accruing to a Youtube-like player (or perhaps more closely a Roblox-like behemoth) that can aggregate the creation within the confines of their own platform.
Conclusion
History doesn’t repeat itself, but it does rhyme. Making that which was rare accessible to the masses is both a means of fomenting tectonic cultural change (think: literacy, internet) and accruing unimaginable personal riches (think: YouTube, Roblox). The incentives are ripe for vibe coding to follow a similar path to previous creativity breakthroughs.
The dam has cracked. What was once the domain of highly specialized experts is rapidly becoming a playground for anyone with an idea and a keyboard. As software becomes content, the ability to create will slowly shift from a privilege to an expectation. In the words of William Gibson, “The future is already here– it’s just not evenly distributed.” and the opportunity is immense for those who seize it.
- 🍋
Addendum:
I wrote the first half of this article on DAWs and UGC a few months back, then the back half as I played with Replit. It’s been a couple weeks since I got around to publishing the piece, and it’s already outdated (of course).
A few bad ass examples of precisely what I mean:
1) We may be seeing the very evolution of DAWs themselves right before our eyes. The birth of Model Context Protocols (MCPs – a coordination layer between LLMs and other software) has empowered vibe production akin to vibe coding:
2) About a month ago, Levelsio (a famous indie hacker/solopreneur) went giga-viral with one of the first vibe coded video games: fly.pieter.com (thread)
Twitter lost its mind in the ensuing mayhem as he raked in cash, monetizing on various purchasable plane types as well as ad space, and exceeding $90K in the first month:
The thread goes deep, and was a fascinating saga.
Suffice to say, these tools are going exponential. Software will become content.
Soon, the only real excuse for not creating will be a lack of agency or imagination. Start playing around.