Just made a video on How to Make AI Videos That Don't Feel Like AI Slop Also finally nailed down my voice model to sound really good. The whole video shows its own making, from the voice to the storyboard, the animation and the render, all out of one folder on my computer. Even shows how quickly you can create different styles as well! I'm also adding the templates to the vault tonight, and updating the Implementation Playbooks to focus on this, with a whole new set of lectures. More soon! If you are loving my work I really would love some support, if you subscribe to premium you can support me to make more free resources for everyone else! Upgrade Here!

Rung two built the window. Rung three is what survives it, open from today. There is something you have told the model three times this month. Nothing big. The client signs as Sam, never Samuel. We stopped saying "the portal" in June. It got it right every time and it forgot every time, and somewhere around the third time you stopped noticing you were the one doing the remembering. That is not the model being slow. That is a window closing, with nothing on the other side. 📚 What just opened Memory Engineering, seven posts, the whole rung at once. The Overview draws the line the rung stands on, this turn against what survives between turns, and the three classes carry one claim each: (1) Most memory fails by keeping too much (2) A memory you cannot find is worse than none (3) A memory is a belief, not a record Then the payoff, the whole series on one card, my perspective on rung three, and the closing post on where memory sits inside ICM. By the end you own a memory: one small file the room reads every time you open it, a drawer behind it for the rest, and a date in your calendar for the day you read it back. No system in it, and you will notice you never built one. 🪜 Where you are standing (Rung 1 & Rung 2 should be completed) If not, you can find them here they are in order and should be taken from series 1 to series 2 and so on: PROMPT ENGINEERING 📚 Series 1 Overview CONTEXT ENGINEERING 📚 Series 2 Overview Climbed rung two? Bring the room's file, the seven lines, because this rung decides what gets written into it, what stays out, and what gets crossed out the day it stops being true. The Overview asks one more thing before class one: the three things you type to the model in every fresh window. Three, not thirty. Not there yet? Rungs one and two are still open and still the order, the Start Here post is pinned. Honest word on this one: it is the rung people stop on, because nobody ever asked most of us to decide what should be kept. So the steps are smaller and every wrong version is shown in full. If you have never built anything that had to remember, that is who I wrote it for.

Full disclosure before anything else: I don't use Claude Code. I'm a proponent of open source, both the products and the models underneath them. That's not ideology, it's economics. I experiment constantly and I pick on value per project, not on brand. This post is one of those experiments, and it earned its writeup. Most people rent five APIs, lose the keys, and wonder why the bill reads like a ransom note. Stop that. Nous Research is the same lab that ships Hermes Agent. They also run Nous Portal. One account. One key. One bill. Their own Hermes lineup plus a catalog of 300+ models: Claude, GPT, Gemini, DeepSeek, Qwen, Kimi, GLM, MiniMax, Grok, NVIDIA, and the long tail of the agentic frontier. Don't take our word for the count. The full list is public at portal.nousresearch.com/models. If you already run Hermes Agent, this is the move. The people who built the harness also sell the inference surface. Less friction. Less vendor-hopping. Memory stays yours. The stack stays inspectable. And the pricing isn't passive. There are 25 active promotions on the catalog right now, spanning Kimi K3, Tencent HY4 Preview, OpenAI GPT-5.6 Sol Pro, Mistral Nemo, and more. That's the point of one bill with 300+ models: the promo finds you, you don't go hunting across five dashboards for it. WHAT YOU ACTUALLY GET - 300+ models through Nous Portal. One subscription. Switch mid-session with /model. - Hermes models (4 / 4.3 family) on their own backend. Cheap relative to the closed labs, open weights when you want local. - Tool Gateway on the same bill. Search, extract, image gen, TTS, browser. Not just chat endpoints. - Same OAuth Hermes Agent already speaks. Run hermes setup --portal and you're grinding. - The agent stays MIT. Portal is optional. BYOK still works. No lock-in sermon required. HONEST CAVEAT Nous say it themselves: Hermes 4 is a chat and reasoning model, not the best pick for the agent's rapid tool-call loop. So flip the script. Use Portal for the 300+ agentic catalog. Use Hermes 4/4.3 for writing, research, local runs, and cheap bulk. Right model for the job, not brand loyalty.

A new animation style on top of the 16-bit one that I created, as well as an amazing interview I had today with Jackson Bennett Keep an eye out for another YouTube video coming soon where I give you all a tutorial on how to make animations and videos like this

Hi guys, has anyone had a chance to try using the ICM approach within Copilot Studio following the most recent update? I had a really useful conversation with Microsoft yesterday, and they shared the attached PowerPoint for me to review. The idea I am exploring is how I can bring the ICM approach to life directly within the new Copilot Studio. I was just wondering whether anyone has already tried this and been able to use it successfully?