Get live statistics and analysis of Michael Mignano's profile on X / Twitter

While we’re all talking about Nathan Fielder, just want to go on the record and say that The Rehearsal S2 finale is one of the greatest episodes of television I’ve ever seen.

Introducing Oboe. Out now. Human beings are getting stupider. We all know it.  Meanwhile, we’ve spent trillions to make the most incredible technology in the history of humanity, and we’ve outsourced all of our individual agency in the process. (1/3)

People are underestimating what will happen to software as a result of AI. Most are focused on how it'll make creating software incrementally easier, faster, cheaper. Few are discussing just how vast the universe of software becomes when this happens. We will have 100x more software (at a minimum!) as a result of AI. There will be a gigantic long tail of every type of application you could ever imagine. Most will have few users, if any. But a lot of it will be super niche and customized to very specific use cases. Some huge hits will break out. People who make software (without ever actually knowing how to code) will emerge as "creators". Most of these people will be dismissed by purists for not making "real software". And some of these people will become very wealthy. And an entire ecosystem will spring up around it. Marketplaces, tools, entertainment, media, etc. The same thing that has happened to every other form of media is about to happen to software. It'll look like a joke at first. But then it will be far too big to ignore.

This might be the most pivotal moment for the music industry since Spotify launched in the US.

Deep Research marks the first time I've been convinced that AGI is already here. The discussion is over.

Technology is not a moat. Never has been. There are so many examples of incumbents copying startup’s technology and leveraging their distribution to crush competition. And AI coding models are quickly making technology even less defensible. With coding models like Claude, assistants like Cursor, and “no code” AI tools like Lovable and Bolt, it’s becoming stupidly easy to copy existing products and apps with virtually no effort. Turns out the stuff that actually matters is what always mattered: strong products, real network effects, ruthless distribution, and brands people actually care about. The winners will double down on those things, while everyone else wonders why their perfect tech keeps getting cloned into irrelevance.

Bet on people who have a motor. If someone is constantly in motion, there's a good chance they're going to find the heat eventually.

Huge opportunity no one is talking about: AI’s accountability layer. It’s easy to manage humans. Way harder to manage dozens, hundreds, thousands, or millions of agents. How will you keep track of the work they’re doing? Massive market to be built around agent management.

I still believe in the user interface. Feels like today's consensus is that the UI is dying in favor of building AI apps on top of existing messaging platforms. I don't buy it. The UI is where the best products earn trust, build brand, and optimize engagement.

I still believe in the user interface. Feels like today's consensus is that the UI is dying in favor of building AI apps on top of existing messaging platforms. I don't buy it. The UI is where the best products earn trust, build brand, and optimize engagement.

Introducing Oboe. Out now. Human beings are getting stupider. We all know it.  Meanwhile, we’ve spent trillions to make the most incredible technology in the history of humanity, and we’ve outsourced all of our individual agency in the process. (1/3)

People are underestimating what will happen to software as a result of AI. Most are focused on how it'll make creating software incrementally easier, faster, cheaper. Few are discussing just how vast the universe of software becomes when this happens. We will have 100x more software (at a minimum!) as a result of AI. There will be a gigantic long tail of every type of application you could ever imagine. Most will have few users, if any. But a lot of it will be super niche and customized to very specific use cases. Some huge hits will break out. People who make software (without ever actually knowing how to code) will emerge as "creators". Most of these people will be dismissed by purists for not making "real software". And some of these people will become very wealthy. And an entire ecosystem will spring up around it. Marketplaces, tools, entertainment, media, etc. The same thing that has happened to every other form of media is about to happen to software. It'll look like a joke at first. But then it will be far too big to ignore.

Deep Research marks the first time I've been convinced that AGI is already here. The discussion is over.

I despise the term “vibe coding” and wish we had all agreed on something else to describe the art and science of non-technical people building software using AI.

Agents are cool and the tech behind Operator and Computer Use is super impressive. But I'm not convinced there are tons of super obvious consumer use cases... yet. That's why everyone always cites the same obvious examples ("imagine you wanted to plan the perfect trip to X..."). Reality is that most of the consumer oriented stuff people do on their computers or phones is around wasting time: forms of entertainment (social, media, mindless shopping) or communication. We want to do these things ourselves, not delegate them. Enterprise is a completely different story. We all do so much crap for our jobs each and every day we'd love to delegate to someone else. Agentic work will 100% be a thing in the very near future here in the enterprise. But for consumers? I think we're a long ways away from mass adoption.

Huge opportunity no one is talking about: AI’s accountability layer. It’s easy to manage humans. Way harder to manage dozens, hundreds, thousands, or millions of agents. How will you keep track of the work they’re doing? Massive market to be built around agent management.

It’s pretty clear to me that the internet is headed toward a new business model that we may not have figured out yet. Consumers will increasingly outsource tasks to agents, and agents will work around the clock to execute them on their behalf. These tasks require inference spend. The most obvious way to cover these costs is by passing them through to the user. Users pay a monthly subscription or simply pay for credits that get deducted as inference gets spent. But with the cost of inference dropping over time, this will ultimately be a race to the bottom. Pass-through pricing will get commoditized and won’t be able to sustain margins. So startups and incumbents will have no choice but to offer a different business model. Paying for task-based outcomes is an obvious iteration. You can imagine customers paying agents when they achieve certain goals, such as successfully booking an appointment or executing a refund the user wouldn’t have taken care of on their own. But most tasks aren't worth paying for one at a time. The most scaled business models on the internet let everyone participate, even when they can't pay directly. This is why advertising dominated as the primary business model of the past 30 years. Incumbents will subsidize usage dramatically over the next few years to onboard as many customers as possible. We’re already seeing this with Muse. But the question is, how will this work get paid for long-term? Can agents adopt a form of advertising that covers the cost? Or do we need a completely new model where the few big paid outcomes subsidize everyone else's usage? We may be headed toward a new world of haves and have nots. Or a world of abundance where everyone can get anything done at any time for free.

This might be the most pivotal moment for the music industry since Spotify launched in the US.

I’m running out of reasons to believe the labs will have moats when it’s all set and done. Models? Unless someone reaches runaway self improvement, doesn’t seem like it. Everyone catches up within months. Compute? That’s a cost, not a moat. Same chips for everyone. Capital? You don’t even need a working model to raise right now. Everyone wants in on AI. Talent? Follows the capital. Training data? The web’s been scraped to death. Everyone’s got the same stuff. Lock-in? Routing makes models swappable. It’s a config change, maybe. And smart routing eliminates even that. Regulation? Good luck once the weights are open. AI is being unbundled.