LLMS are not magic! They are code, boring code!! Here are some examples of all those favorite things you think is magic. Look, it is code and not trying to hurt me. No, this is not some fast paced, high-thrilled, adventure film... it is a boring tech demo to show you some boring code, and how you can see this boring code do boring things. My harness is so boring, it just sits there, waiting on me or following a schedule. 00:00Intro 00:54Factory Description 02:14Factory Demo 04:20Routers (aka task specific agents) 05:13Agents (aka general purpose) 05:57Tools description 07:41Logging and diagnostics 08:14Demo: Tool Testing 11:04Demo: Skills 13:03Demo: Project SOP / Knowledge 15:05LLM as a Monitoring Platform 15:40Conclusion

Is there even a point with rules in the Grok Chat? @elonmusk @SpaceXSI I was able to use the chat, then chat with projects for over the last year without issues... but this 4.7... wow... the just straight up ignore of account level rules on chat... I used the Grok bot yesterday, and was actually a little impressed. However a basic verification led to nothing but frustration... the 4.7 is just unusable... i cant have a rogue cannon in the code base.

Intro to AI Imagery: Local AI comparison This is a demonstration of some of the Vision based tasks. From how the Mario RL used vision to cloud based services, and then some local image processing examples. 00:00intro 01:00Video Record with Data Embeddings 01:45Recorded Video Data Analysis 02:15Archive copy 03:22X AI Demo 03:50Image Generation harness 04:23Generate Images 06:18LLama 3.2 11b Vision batch captioning 06:47Image prompt without Knowledge 07:34Knowledge enabled 07:47Single Text Block 08:52Multiple Text Block 10:46Two Image Compare 11:58Conclusion

For those who wonder what the Pro1X life is like. When it is said and done, the issues in their compute, infra, and model output (quality, verbosity, and watermarks): Avoid: The hype and delivery do not sync. Lost a month of productivity: hundreds of hours of effort, wasted.

I turned my back for one minute, and something like Doom snuck in there. Python Doom-like Reinforcement learning: 16-24 agents with a 5 to 1 ratio of 'normal coded' enemies, and an optional human roam a randomly generated level, collecting power-ups, ammo and more with game play inspired by Doom. The simulation trains the model from each of the agents and human (when applicable). Everyone for themselves, until only one stands! This is a double dqn so nothing too special, was one of the earlier projects worked on.

If I told you that you were a 4th dimensional object, would you agree or think I am wrong? We were all taught, we are 3 dimensions: Height, Width, and Depth, aka X, Y, and Z. When did you hear that? A few years back, yesterday, just now before you finished reading this very sentence? Do you understand When? If I told you, that at the very moment your eyes see the question mark, you will be at position: X23, Y83, Z0, A29401, would you believe me? If your capacitors are fluxing, then get ready for the nerdiest video (no dialog, expect to pause, its wordy) on the most complicated Mario RL project ever, someone get the world book people I might have a contender. Fun fact: When you click the like/share crap, the algorithm rewards me for steps taken before this post. Watch the video and it all makes sense now, doesn't? Chapters for what is a contender for "the nerdiest thing posted": 00:00 Intro 01:08 V1: Atomic based standard RL version. 02:38 The road of upgrades 03:28 V2: Non-Atomic not even close to standard RL. 04:30 Review and cross assessment 04:51 Other work from the studio

AI is for the young, only they can understand it... Oh really now? In this episode: Old guy makes a local LLM platform for fun. LLMs are not difficult, just big and painful like a dragon.

This is what I work on when not fighting the AI. Enjoy the first demo reel, while working on the next!

For everyone who wanted to know what it is like to have Claude AI Assistance. This is what it actually looks like when not running agent harnesses. I would advise parsing those logs ASAP! If you're wondering why the API bill is up there with no results, they you might have had a week like mine!

World First: Reinforcement Learning: Super Mario Bros. All Stars on physical console with cartridge! As per multiple searches, no one has ever connected a physical console to the computer and completed a Super Mario Bros.™ level. There are plenty of examples with OpenAI's Gym code set, however those examples did not use a physical console, nor did they use a physical cartridge. On November 11th 2025, episode 1034 completed Super Mario Bros. All-Stars™ version, level 1-1 on a physical console (Retron5) with a physical cartridge of Super Mario Bros. All Stars for the Super Nintendo. While numerous searches for this type of experiment has not found any other examples of this, I wanted to document it so that we can say definitively: Super Mario Bros. All-Stars™ version for the Super Nintendo Entertainment System™, has had level 1-1 cleared without memory access, using a physical cartridge with optical capabilities that emulate the human more then the game. This was completed exclusively as if a human was playing the game. In other words, It read the screen and pushed buttons.

Introduction what will hopefully be the best Role Playing Game of 2025! This video is an introduction at the early stages of development with a first week of October 2025 release! A side scrolling twin stick RPG with MetroVania elements. Inspired by many of the greatest games from the 90's, to create a one of a kind experience. Play as Gabriel, a tank who sacrificed everything to stop the last great war, brought back from the grave when the world needed saving, and there was no one else to answer the call.

Hey everyone, welcome back to the latest edition of "Innovative AI Solutions from Concept to Reality"! Today, I'm excited to showcase the latest updates to the cutting-edge video creation tool, M1.2. This isn't just a minor update - it's a game-changer.

For those who love AI, A demonstration on how AI can be used to create works of art. This is the first part of a quest to create the perfect tool. This video is the documentation of the first milestone, a basic bragging of 'look, cool stuff in the works' The expectations for this stage is proof of concept, which was met. In today's developer communities, there is a notion to 'build in public', so it is in that spirit this video was made. That being said, this is a company tool, so the code is not available, and will be managed with some basic company IP type policies. I would like to give a shout out to: - @BasicFilmmaker, as I learned a lot from his tutorials for video production. - @NetworkChuck as his videos helped me to set the AI \ Docker components. - @cinemassacre: AVGN episodes helped pass the time waiting for everything to render. - @elonmusk & @xai because it's AI being used in odd ways. Thank you and Enjoy the show! ( 4k on YouTube: youtu.be/E8NndPwgheE )

Hope you're hungry, cause it's dinner time! With Pork Chops, Maple Sausage Stuffing and gravy! youtu.be/GgIpCGfc1oQ

I figured everyone is getting into AI, so why not highlight a few things it can do: youtu.be/JSFLu0ROwIE

Looking for that special something for breakfast? Then try a stack of delicious Pancakes! youtu.be/fyCctnOfCu0

Let them eat cake! This is how you make it by Cooking the EZ Way! youtu.be/hFGwRx5GjhA

What goes best on a Sunday morning? How about a Bacon Egg and Cheese omelet? Here is how I make one, enjoy! youtu.be/NkaIAKYy0OI