Andy Singleton

I told you I pivoted to AI. It's going extremely well. I raised $4 million. The pitch deck had 22 slides. The word "AI" appeared on every single one. Fourteen slides had the word "AI" in the title. Three slides were just the letters "AI" in large font over a stock photo of a brain with circuits on it. One slide said "THE FUTURE IS NOW." That was the financial projections slide. There were no financial projections. There was the brain. My startup is called Synthetica. We are an "AI-native intelligence platform." That means we built a website that calls the ChatGPT API and displays the response in our own font. The font is called Satoshi. I chose it because it sounded like crypto and technology at the same time. Two birds. Neither of them real. We have a whitepaper. It's the same whitepaper from my crypto startup with "blockchain" replaced by "neural network." Find and replace. That's our IP. Our product costs $29 a month. ChatGPT costs $20 a month. The difference is our logo and a loading screen that says "Thinking deeply..." while it waits for the API to respond. It's the same API. The customer is paying $9 a month for the loading screen. I call that the moat. My technical co-founder built the entire product in a weekend. We spent the next eleven months "iterating on the brand." That means we changed the color of the loading screen four times. He quit in month seven. I replaced him with a contractor in the Philippines who charges $15 an hour. I told investors we have "a globally distributed engineering team." That's two people. One of them is me. I don't code. We raised the $4 million in February. The lead investor asked what our moat was. I said "proprietary AI infrastructure." He asked what that meant. I said "we've built a custom orchestration layer on top of foundation models." He asked if that was an API key. I said "it's significantly more sophisticated than that." It's an API key. He invested $2 million. His fund has "AI" in its name. It was a crypto fund until 2023. They changed the name. They changed the website. They did not change the partners or the strategy. The strategy is to invest in things they don't understand and exit before anyone notices. I respect that. It's the same strategy as mine. The partners all have the same LinkedIn arc. Crypto evangelist from 2020 to 2022. "Building in stealth" from 2022 to 2023. AI visionary since January 2024. The conviction was always there. The noun changed. OpenAI just raised at a $730 billion valuation. That's more than the GDP of Switzerland. Anthropic is at $380 billion. In January and February alone, $220 billion went into AI companies. 83% of all venture capital in February went to three companies. Three. The other 17% went to four thousand startups like mine. API wrappers with pitch decks. Loading screens with brand identities. $29-a-month products built on $20-a-month products. A company with fewer than 100 employees is now worth $12 billion. I don't know what they do. Neither does the company. But they have a whitepaper. And the whitepaper has a diagram. And the diagram has arrows. Arrows mean progress. We're early. I launched a token. SYN. The Synthetica utility token. It powers the "decentralized AI marketplace" we haven't built yet. Someone asked what the token does. I said it "facilitates value exchange within the Synthetica ecosystem." He asked what that meant in plain English. I said "you can buy it and it might go up." He bought $12,000 worth. The total market cap of SYN is $340,000. I own 40% of the supply. My Discord owns another 30%. My Discord has 1,200 members. Eight hundred of them are bots I bought on Fiverr. We have a Telegram too. The Telegram has a price bot. The price bot posts the SYN price every hour. The price has not changed in three weeks. Nobody has traded it. The bot keeps posting. That's community engagement. I used the same Discord. I just changed the banner. The crypto community became the AI community overnight. Nobody noticed. The conversations are identical. Just replace "to the moon" with "to AGI." I also run a Polymarket bot. An AI-powered prediction market trading agent. It bets on real-world events using an algorithm I don't understand, funded by money I don't have, on a platform a U.S. Senator is trying to ban. Fourteen of Polymarket's top twenty traders are bots. Bots made $40 million last year exploiting pricing gaps. One bot made $115,000 in a single week. Mine lost $4,200 in eleven days. But I made a course about it. "AI-Powered Prediction Markets: The $115K Playbook." It costs $497. The playbook is a PDF. The PDF has screenshots of someone else's bot. I added my logo. Thirty-one people bought it. I made more from the course than the bot made from the market. That's the real alpha. The CFTC put out a warning that said "fraudsters are exploiting public interest in artificial intelligence to tout automated trading algorithms" that "promise unreasonably high or guaranteed returns." I screenshotted that too. I posted it in my Discord. I said "they're trying to shut us out." That got forty-seven rocket emojis. The bots sent thirty of them. A veteran VC said this week that AI valuations are "overheated." He said "buy high, sell higher only works in a bubble." I screenshotted that quote. I posted it in my Discord. I said "this is what they said about the internet." They also said it about the metaverse. They were right about the metaverse. I was there. I owned eleven properties. They're worth $6,400 now. Combined. My Bored Ape went from $189,000 to $14,000. The Gucci store is still empty. My beachfront villa is a mobile app. I learned a lot from that experience. I learned that if something goes to zero, you should pivot to the next thing and do it again faster. The metaverse taught me timing. Crypto taught me language. AI taught me that the language doesn't have to mean anything as long as the timing is right. We're early. I hosted a demo day. Fourteen investors came. I showed them the product. I typed a question into Synthetica. The loading screen said "Thinking deeply..." for eight seconds. Then it gave the same answer ChatGPT gives. One investor asked "is this just ChatGPT?" I said "we leverage GPT-4 as one component of our multi-model reasoning stack." He asked what the other components were. I said "proprietary." He asked to see them. I said they were "in stealth." Stealth means they don't exist. He invested $400,000. My mom called. She asked how the AI company was going. I said "we just closed a $4 million round." She said "is this like the metaverse thing?" I said "this is completely different." She said "you said that about the NFTs." I said "the NFTs were digital art. This is artificial intelligence." She said "is the monkey still your profile picture?" I changed the subject. She asked if I was eating enough. I am not eating enough. I spent my grocery budget on GPU credits. I don't know what a GPU does. But you need them for AI. Everyone says you need them. I have $7,000 in GPU credits on a platform I've logged into twice. That's infrastructure. My accountant called. The same one. He asked about the startup. I said "we're pre-revenue." He said "you've been pre-revenue at every company you've ever started." I said "this time we have product-market fit." He asked what our product was. I said "an AI-native intelligence platform." He asked what it did. I said "it thinks deeply." He said "so it's a loading screen." I hung up. He's not a visionary. We're early. I know we're early because I've been early my entire life. I was early to the metaverse. I was early to NFTs. I was early to the DAO. I was early to the token. Every single time, I was early. I have never once been on time. But that's the thing about being early. You don't have to be right. You just have to be first. And then when it collapses, you say you were "too early." And when the next thing comes, you say "this time is different." This time is different. The AI bubble is not a bubble. It's a paradigm shift. A fundamental restructuring of how value is created and captured in the digital economy. I read that in a pitch deck. It might have been mine. They all look the same. I have a folder on my desktop called "Pitch Decks." There are forty-seven files in it. I opened one from 2021. It said "THE METAVERSE IS A PARADIGM SHIFT." I opened one from 2024. It said "AI IS A PARADIGM SHIFT." Same font. Same brain. Same slide. I didn't delete the metaverse one. I might need it again. We're early. As long as the graph goes up and to the right.

Reminds me of Peter Naur's classic 1985 essay "Programming as Theory Building" which argues that a program is not its source code. A program is a shared mental construct (he uses the word theory) that lives in the minds of the people who work on it. If you lose the people, you lose the program. The code is merely a written representation of the program, and it's lossy, so you can't reconstruct a program from its code. If you think of total software debt as technical debt + cognitive debt, then previously, we mostly had technical debt. Now with AI we have both. Previously, when you built something, you accumulated technical debt but relatively little cognitive debt because you had to understand what you were building in order to build it. In other words: the theory came for free as a byproduct of the work. AI breaks that coupling. Now you can produce code without building the theory. So you're now able to accumulate both kinds of debt simultaneously - technical debt in the code and cognitive debt in yourself. And cognitive debt is arguably worse because you can fool yourself into believing it doesn't exist. Technical debt tends to show up in semi-obvious ways that we understand well as an industry. Cognitive debt is more insidious - it means you're unable to even reason about the program (because you possess no theory of it) - which is what Naur describes as the "death" of a program.

Even the best developer tools mostly still don't let you sign up for an account via API. This is a big miss in the claude code age because it means that claude can't sign up on its own. Putting all your account management functions in your API should be tablestakes now.

As agents become the highest volume users of software in the future, a lot is going to become critical to support for them to be effective. Agents need to be able to signup for your tool on their own, have their own scoped access controls, be able to use your entire system through API/CLI, be able to be billed for their usage, need a computer and filesystem to use, and much more. We’re going to evolve from building primarily for the human user, with APIs as a means to get that data or tool in another platform, to a world where the API becomes the core source of truth actions. Any software that can’t support this basically won’t exist to agents.

Today we’re open sourcing Evolver, a near-universal optimizer for code and text. While benchmarking we achieved SOTA (95%) on ARC-AGI-2 (last week that is 😆) and 3x’d performance of the best open model, reaching GPT-5.2-level performance.