It's been five days since the MESA character creator went live. Since then, hundreds of custom characters have been created by users. Stream chat activity has tripled. Here's a thread with some of my favorite characters so far! ⬇️⬇️⬇️

The @mesastream character creator is now live. Possibly the FIRST interactive AI livestream with persistent, user-generated characters. ANYONE can create characters that become instantly usable on stream. ⬇️

One year later, and @mesastream continues. Chat continues to request topics and create clips, while development continues behind the scenes. At the time of writing, solana:5wVtfsFhLjxm27K9mN3ziYWCCpQwXXq7HWUiRMW7pump sits at a $113K market cap with $54K in liquidity. That's about a 48% liquidity-to-market-cap ratio. Not bad for a Solana project at its current valuation! Planning some partnerships with other crypto projects, and exploring the best ways to apply what we've built here. The research continues. λ

btw we upgraded the !research command. should be much quicker and more accurate now. it uses Meta Spark Spark 1.3 + web search tool calls before the scene plays.

MESA-coded things happening. The new media category continues to receive attention. 🧵👇

just wired up @X chat to the @mesastream stream. Now you can chat with the cast, influence each scene, and trigger chaos commands directly from X. Visit mesa.stream/guide for a list of commands.

We're coming up close to one year hosting the @mesastream 24/7 AI livestream. One year, and unique chatters are still showing up every day. We've accumulated 100k watch hours, processed hundreds of thousands of chat messages, and streamed for 8,000+ hours. AI orchestrated livestreams / responsive content generation will become a major category of entertainment. However: - AI slop lacks curation and humanity. - AI authorship is largely boring and uninspired. By connecting it to the crowd, it becomes something special. AI serves as the orchestrator for HUMAN ideas and creativity rather than a degenerate slop factory. Viewers become the directors. Look at Roblox, YouTube, or any other UGC platform. The trend is clear. The gap between producers and consumers is shrinking and soon there will be no distinction at all. The winners are the ones who understand: 1. Media stops being a product and becomes a place. 2. Anyone can generate, but few can gather. 3. Every audience collapses into a community. The final phase of participatory media looks a lot like Dungeons and Dragons. In D&D, the DM doesn't write a story. They maintain the world. It has rules and limits, but that's what makes it compelling. A world where anything can happen is a world where nothing matters. The fun comes from operating inside the constraints. The players bring the energy that drives the story forward. That's what we're building towards with the MESA framework: - AI maintains the world and its state - Persistence compounds, lore accumulates, relationships develop - Everything responds in realtime Imagine a narrative that branches on demand. What if that character had lived? What if the last season went differently? Run it and find out. This is where stories become open-source git repos. Forked and branched at the whim of the audience. Canon events determined by whoever shows up. One year in. Still early. λ youtube.com/watch?v=uXSROOJO…

The thing nobody's shipped yet: characters that move in ways nobody programmed. On @mesastream, imagine chat asks for a limp, a dance, or interaction we never intended. Every character with their own way of moving. Exploring this.

Some interesting @mesastream stats: - 1,765 phone calls processed - 10 hours of audio - Luther made up 30% of the total calls Overall, would consider it a successful experiment.