The greatest opportunity for ascension in the 2020s lies in AI integration and figuring out how AI will be used in the coming decades. AI models are experiencing extreme innovation, but the true profiteers of this time will be the people who find out how to use these models in the most useful manner. It wasn't Internet Service Providers or Broadband companies that consolidated the most incredible wealth in perpetuity; it was the internet entrepreneur. Mark Zuckerberg, Jeff Bezos, Larry Ellison to name a few. Sure, ISPs are successful businesses, but what ISP founder is one of the richest people in the world?
It is with this backdrop that I have migrated my thoughts on startups to an AI-first perspective. Big companies cannot and aren't willing to overhaul their existing business models and services to become fully engulfed in useful AI integrations. There is plenty of fluff and lackluster AI uses on the internet built out by these tech giants. Comment summarization comes to mind. People enjoy reading comments, so an AI integration that replaces an enjoyable, community-building experience is worse than unnecessary. Many of the current uses of AI among the tech giants are like this. They can't rebuild their services so they're left with small breadcrumbs of AI usage everywhere. This era may prove to be the most overflowed with opportunity since the early days of the internet or the social media boom.
As of now, I'm a technical consultant at Oracle, but I have many projects that I'm building that will take advantage of this opportunity. PLATO5 is an AI-first social app, designed to create online connections that evolve into IRL friendships. This is the first and most fleshed-out project that I've been working on. Kommin is an offshoot of PLATO5 that grew into something that no longer had a direct bearing on PLATO5's value proposition. After many iterations and pivots, I've set on turning it into a mobile-first, minimal oratory board that hopes to one day stand toe-to-toe with the entirety of the podcasting industry through small, digestible, auto-play enabled voice clips that can be easily thrown on during work or transit. Goldstein Simulator is a big one. This project is dedicated to optimizing existing LLM responses for simulations that maintain the logic of their respective realities and could one day be the narrative backbone of AI-generated video games, choose-your-adventure films, and more. My personal favorite theoretical application of Goldstein Simulator is my idea of Simulation Lounges that encourage social connection.
PLATO5 was initially not supposed to be an AI app. In fact, I started building it before LLMs were fully introduced to the wider public through ChatGPT. Unfortunately, early on I fell into the same trap as the tech giants and built AI Chat Suggestors, instead of my current solution: AI Chat Managers that passively and actively aid users with managing their conversations and help them brainstorm new things to discuss, making IRL connections more likely and conversations more engaging. The chat suggestions were peripheral; the chat managers are direct, constant, and easy to utilize. This is the future of ai-chat integration in my opinion.
The chat suggestions weren't necessarily bad, but they distracted from the main purpose of PLATO5: building those IRL friendships for users. I even gave the Chat Suggestors 5 different personalities for different types of chat suggestions. Zen was a neurotic creative, Steward was a historical savant butler with a penchant for uniformity, Plato was a mirror of the Ancient Greek philosopher of the same name, but with little quirks like the propensity to get flustered and call you 'malaka', Cyba was a cyborg with a cybernetic startup and glitchy hardware, and Gertrude the Conqueror was a sweet, docile grandmother who doubled as a warlord, ever-defending the PLATO5 realm. While humorous, this got in the way of PLATO5's value proposition, as you might have guessed. Also, it's worth noting to anyone interested in AIs with personalities: it's one of the things that LLMs lag behind in. Not believable whatsoever.
Kommin will use AIs in a more limited capacity for now, but maybe later bigger integrations could be built. AI Elaboration is an actually useful version of summarization. Instead of having to tap a user's profile or dig through old komms, you can feed that data to AIs and they can provide insight and intriguing connections on why the user may have recorded this komm and who they are. This is a good example of a useful AI integration that stands in stark contrast with comment summarization. Also, all algos will likely be AI enhanced when I have enough data to train an algorithm on.
Goldstein Simulator has an obvious AI application. Its business model hinges on AIs. It's actually the multi-layered approach, turn-based collaboration features, on-the-spot fine-tuning, and backend coordination that makes Goldstein Simulator's process unique. I'll expound on this more in the future. I'm about a month away from starting the build.
So, to wrap things up here: this is my big play. If I'm to make it in the world of entrepreneurship, all or some of these projects will have to propel me. I'm not naive or reluctant to get my name out there and stump for my visions of these companies and I'll be dedicating my money, time, and spirit to these projects until I'm either repeatedly and forcefully denied or… I win.