Ben Cera's Polsia reportedly reached an $18.7M annualized revenue run rate with zero employees. The intriguing part isn't the headline number. It's the possibility that AI agents could fundamentally change the relationship between workforce size and business scale. But revenue isn't profit, and AI agents aren't free. The real breakthrough will be building sustainable, profitable companies with dramatically smaller human teams.

Viggle's JST-2 is worth watching, not simply because it generates 3D content, but because of the direction it represents. We're moving from AI that generates pixels toward AI that understands spatial structure and motion. Gaming, animation, and eventually physical AI could benefit enormously. The next frontier isn't just generating what the world looks like. It's modeling how it behaves.

The next enterprise AI bottleneck may not be model intelligence. It may be organizational memory. Scuba's $34M launch is an interesting bet on this. AI agents can execute tasks, but without understanding a company's decisions, priorities, and history, they can still get the work wrong. Context could become a major competitive advantage.

Something interesting is happening in Nigerian gaming right now. Lagos Life exploded in days, Danfo Run showed how local culture could travel, and more browser-based games are appearing around everyday Nigerian experiences. The bigger story isn’t just virality. It’s the combination of: • AI-assisted development • Instant browser distribution • Hyperlocal cultural ideas • Social sharing • Much lower experimentation costs That changes what kinds of games can get built. Instead of asking whether a highly local idea can justify a traditional studio budget, developers can now prototype fast, publish behind a link, and let the audience decide. That’s why danfo rides, NEPA, Lagos rent, traffic, local food, and city life are suddenly becoming playable systems. We also built DanfoRush - @DanfoRush , a Lagos bus-driving simulator, and it’s been fascinating to watch different developers independently turn the same cultural material into very different game loops. The takeaway: AI is not just making games easier to build. It may be making culturally specific games economically possible to test. And Nigeria could be one of the clearest places to watch that shift happen.

Nigeria’s AI gaming moment feels bigger than one viral hit. Lagos Life, DanfoRush and the new wave of browser games are showing how AI + local culture + instant distribution can unlock ideas that traditional game economics might never fund. This could get very interesting.

Lagos danfo game 🟡 DanfoRush — Oshodi to Ojuelegba, play free from Friday.

I see game development taking off in Nigeria like the fintechs did.

Home made games are picking up in Nigeria - made possible with the use of AI

The real disruption isn't AI making games. It's game creation becoming a form of content creation.

Founders, give @attnwatch a trial. List your products free. Launch your product updates. Let me know what you think. If you want your products added by our team, provide the url here 👇

AttentionWatch @attnwatch

Hello everyone. I’m a founder in Washington DC. What are you building? What solutions are you looking for? Let’s connect. 👇

If you are using Elevenlabs, their v4 is free on the web till October 12.

Who are the founders of all these AI product launches? Please consider launching on AttentionWatch @attnwatch . It’s free. It’s a directory. It has MCP and plugins for all popular Agents - Claude, ChatGPT, Muse… Support a fellow founder 🙂

OpenAI yesterday: agents, Dots, plugins and AI that keeps working. Google today: Gemini 4 Argon for complex engineering, cyber defense and enterprise work. Different strategies. Same direction. The AI race is moving from who can answer best to who can actually do the work.

OpenAI DevDay 2026: Will AI Agents choose your solution?

There was a lot to digest from OpenAI DevDay 2026. Dots. GPT-6.1 Sol. Computer-using agents. ChatGPT Space. Pages. Collaborative slides. Codex in the cloud. A $500 Pro tier. But one announcement keeps coming back to me: OpenAI is turning ChatGPT into a distribution platform for software. The new plugin system is much more than the old idea of connecting an app to ChatGPT. Developers can now build richer experiences that live inside ChatGPT, including sidebar presence and interactive interfaces. OpenAI is also improving plugin recommendations so relevant tools can be surfaced contextually while you're having a conversation. And ChatGPT now has 1.2 billion weekly users. That creates an interesting question. For the last 25 years, companies have spent enormous amounts of time thinking about: Can people find us on Google? Then mobile created another question: Can people find us in the App Store? Social media added: Can we win attention in the feed? AI agents may introduce another: Can AI find us? Imagine asking an agent to accomplish something and, instead of you searching for the right software, the agent discovers the appropriate tool, connects to it with your permission and uses it to complete the task. At that point, software discovery starts changing. Your customer may not always discover your product first. Their AI might. That means a new form of optimization could eventually become important: making products understandable, accessible and trustworthy not only to humans, but to AI agents deciding which tools can accomplish a job. OpenAI is already putting pieces of that system together: Plugins give agents tools. MCP gives them connectivity. Events tell them when something has happened. Dots give them persistence. And ChatGPT provides the distribution surface. This is why I think DevDay 2026 was about much more than another set of AI features. We may be watching the beginnings of an agent economy—where AI increasingly sits between intent and software. If that happens, the battle for attention doesn't disappear. It gets a new audience.