I spent 25 years as a Filmmaker. Showrunning, directing, and producing TV, films, and commercials for dozens of major networks and global brands including Apple+, HBO Max, Discovery, MTV, YouTube Originals, and many more. Won an Emmy. Led crews of 300+. Led projects with Mark Hamill, Joseph Gordon-Levitt, Kim Cattrall, Logic, the Kardashians, Nelly, Meek Mill, and a U.S. Presidential campaign. Then I went quiet.
I became a Creative Technologist and AI Systems Architect. Spent the last few years heads down, learning every tool, trying every framework, breaking things, rebuilding them, pushing past the limits of what these systems can do. And my own limits. I iterated over and over and over. When something broke, I figured out why. When a tool couldn't do what I needed, I built the tool. When I hit a wall, I went around it. I'm now building custom agentic AI ecosystems — multi-agent systems that run 24/7, autonomously managing operations, generating content, and scaling business workflows for elite individuals and major companies.
I haven't had this much fun in years. It's the same feeling I had as a kid picking up a camera for the first time — falling in love with something so deeply that you can't stop thinking about it. Except this time the creative possibilities are truly endless. We are only limited by our own fears — not the tools. I'm all in.
This is where I'll share what inspires me, what I'm experimenting with, what I'm building, what I'm breaking, and everything I'm learning along the way — at the intersection of Tech, AI, Agentic Systems, and 25 years of creative leadership. Follow along if you're curious about the exciting and unknown that lays ahead — from someone diving in head first every day, not just talking about it.
Real-world logic, now in your video generations.
Gemini Omni pulls from Gemini’s deep knowledge of history, biology, and narrative logic to generate videos that look and behave like the real world.
See how it works 🧵 ↓
@Tech2Wild thats what I've been wondering all day is will it be so slow that isn't worth it locally on 2 sparks. Seems like you felt its not practical to run it that way?
@Teknium Just fyi... been testing it out all afternoon but having lots of issues. It disconnects after one rough of Hi, then with some fixes it disconnects after 2nd round and havent yet been able to get it to do endless back/forth with answers.
Creative people don’t need AI to “be creative” for them. They need it to remove the drag around the creative call: research, versions, summaries, boards, logistics, notes. That’s where the leverage is.
What I meant was for example if you are using Hermes with Comfyui - what is the brain of Hermes at that time. Where I'm going is if its a frontier model the censoring may be happening there, if you are end to end only using local then Im unsure why you would get the issue, but I want to know what you find out.
I keep coming back to one rule for AI work: don’t automate the judgment until you’ve mapped the consequence. Drafting is cheap. Acting is expensive. The system should know the difference before it touches anything real.
Few questions - in the past getting twilio to allow you to use the number requires red tape and approvals. For this use case do you have an easy to follow instruction of how to fill out the correct answers for twilio for this use case? Also can you setup more than one agent with same number?
Introducing Agent Arena: real-world agentic evals at scale.
How do you evaluate agents doing actual work? We measure millions of live sessions where real users accomplish real tasks.
On Arena, models now get web search, filesystem, and terminal tools to complete complex workflows: writing code, creating slide deck, researching the web, building apps, and analyzing documents.
Every session produces rich signals. Users iterate with the agent turn-by-turn: approving, editing, correcting, praise or expressing frustration. The environment gives feedback too: shell errors, tool failures, recovery attempts, and more.
Our leaderboard measures each model's agentic performance using causal inference across five signals: task success, steerability, error recovery, user praise vs. complaint, and tool hallucination.
This leaderboard snapshot is built from 300K+ tasks, 2M+ tool calls, and 40M lines of code by agents.
Top labs in Agent Arena:
- #1 @OpenAI: GPT-5.5 (High)
- #2 @AnthropicAI: Claude-Opus-4.7 (Thinking)
- #3 @Zai_org: GLM-5.1
- #4 @GoogleDeepMind: Gemini-3.1-Pro
- #5 @Kimi_Moonshot: Kimi-K2.6
More analysis in the thread, with the full technical blog below.
@DJLougen Agreed, for me I have spun up a lot of specialized agents separate instances with specific job categories and then have them orchestrate and spawn subs. But hearing others have less instances and more subs.
@seanward Agree fully and I couldn't love cinema, the movie theatre, storytelling more. It's been the love of my life since I was a few years old. But there is not one way to do things, not 1 type of paintbrush or process. I look forward to new stories told by new artists in new ways.
I hear you but I disagree that it is that simple or narrow. I am an artist have been my whole life. Have been creating on all levels from indie to studio to network to AI. I get it. I also know you can hone your own models, your own loras as well. So I think you can and are not stuck with just the frontier closed source.
AI won’t replace creative taste. It will expose who has it. When everyone can generate a decent image or cut a rough video, the scarce skill becomes judgment: what to keep, what to kill, and why.
@artopiumcom@ComfyUI is open source/free. @ltx_model is open source/free. DavinciResolve is free. You can literally create end to end for free now - was impossible to do that 30 years ago.
Any artist anywhere can get access to incredible tools right now. I actually think there is more democratization of creative tools then there ever has been. When I was a kid you had to save up a fortune or get 35mm donated. It was near impossible to make a film. Not the case now.