Gary Gannon

I’m an optimist about AI because I believe human potential has always been limited more by opportunity than imagination. We’re entering an era where the distance between having an idea and bringing it to life is collapsing. That should excite us. The future isn't something we have to fear. It's something we finally get to build.

the carnival has come to midnight city

im shooting for pre pre alpha in about a month.

I’m orchestrating real software development across ChatGPT/Codex and Claude from a voice conversation on my phone. Two different AI systems. An actual codebase and server environment. A bridge carrying engineering requests and findings in both directions. And me directing the work without opening a laptop. Here’s how the workflow operates: ChatGPT/Codex handles the conversation, technical reasoning, task definition, and review. Claude is my Builder, already working inside the project’s development environment. I talk through an engineering problem with ChatGPT. We work through the architecture, dependencies, constraints, potential failure modes, and what we actually need to verify. That conversation becomes a scoped engineering assignment: what to investigate, what changes are authorized, what needs to remain untouched, and what evidence should come back. I approve the request. The bridge delivers it to Claude. Claude acknowledges it, works against the actual project, and returns its findings through the same bridge. Then ChatGPT reads the report and we discuss the results by voice. What did the investigation establish? Which conclusions are supported? What’s still an assumption? Does the evidence justify implementation, or do we need a narrower investigation first? That discussion becomes the next assignment. I approve the scope, and the loop continues. The return path is a huge part of this. The findings come back into the conversation where the decisions are being made. I can question a conclusion, change a constraint, or ask for a deeper inspection, and that becomes a precise follow-up for the coding agent. I’m still setting priorities and authorizing the work. If an investigation reveals that we need more infrastructure, additional storage, or a broader change, that comes back as a decision before we proceed. There’s a real division of responsibilities here: conversational reasoning and review on one side, execution inside the development environment on the other, with an explicit handoff between them. And I’m operating that loop from my phone, by voice, without manually carrying prompts and reports between the two systems. The conversation is connected to execution. I can work through an architectural question, dispatch an investigation, review the evidence, and direct the next iteration while I’m away from my computer. My phone is becoming a control interface for the entire development workflow and when I have time I open my laptop running @omarchy (of course) I check quality control.

the biggest shift ai could bring to gaming isn’t better graphics or more convincing characters. it’s a change in who gets to decide what happens next. imagine finishing a match with your friends and saying, “this would be incredible if the city flooded halfway through and both teams had to fight their way onto the rooftops.” instead of that being a suggestion nobody acts on, it becomes something you can build, test, and play together that evening. the interesting part isn’t that a machine made a map. it’s that a conversation between players became a place they could enter. that’s a much bigger possibility than cheaper content. it could make the boundary between playing and creating almost disappear. you wouldn’t have to decide to “become a developer.” you would have an idea, try it, discover what was wrong with it, and keep going because your friends were having fun. but there’s a catch. making something easier to create doesn’t make it worth experiencing. an infinite supply of generated worlds could feel just as empty as an infinite feed of generated posts. if creation becomes abundant, attention becomes harder to earn. taste matters. atmosphere matters. the people you meet matter. someone caring enough to make a place worth returning to matters. the future i’m interested in isn’t a machine endlessly generating entertainment for passive players. it’s giving people enough creative power that the distinction between the audience and the people making the world starts to break down. you log in to see your friends. one of them says, “i made something. come with me.” and this time, that sentence could lead almost anywhere.

standing on the shoulders of giants

who is login into forever today?

recursive optimization loop running. must achieve maximum fps

building the tools so you can make games in the open world of midnight city. this is a first tech pass for an fps level. you will build these on your virtual land.

the biggest opportunity for ai in gaming is not faster content production. it is changing which ideas become worth building. consider what happens when the cost of trying an idea falls. a game that could never justify a studio budget might still be worth making for a small community. an experience that only interests twenty friends can matter enormously to those twenty people. commercial scale stops being a prerequisite for creative value. that's the direction i'm pursuing with midnight city. a shared world where players can build their own places and games, with ai helping them get started and hands-on tools giving them control over the result. but the real opportunity comes after that first creation. someone plays it. a friend suggests a different puzzle. another wants a new route through the level. the creator tries those changes, sees what works, and improves it. that is how someone develops taste and skill: by making decisions and experiencing their consequences. reducing the technical barriers gives more people access to that process. none of this makes quality automatic. a generated experience still has to earn someone's attention. more content nobody enjoys is not progress. but more people being able to create something their friends genuinely want to play? that is. the future i'm interested in isn't a machine endlessly generating entertainment for passive consumers. it's players gaining more power to shape the worlds they inhabit.

Happy Halloween

🤔

killer youtube.com/watch?v=Ip_Kc3EQ…