Will James

@iamlukethedev OS level security still going to create a million hassles for you. No escape.

I’m blown away by how easy this was. I asked Hermes to install Computer Use, control my machine remotely, take a screenshot, and send it back to me. And it just worked. No setup nightmare. No fighting configs. No stealing my cursor. Just my agent using the computer while I’m away.

@BrianRoemmele Was thinking of you when I found this house. There was a liveliness that was expressed in everything mid-century! x.com/i/status/20677…

They had optimism about who were are and can be. Their buildings would declare it. Here is the testimony.

When the past still looks like the future…

@taziku_co @nexta_tv This is super cool

エアホッケーでほぼ負けないロボット。 注目点は「反応速度」より、現実の不確実さをどう学習させたか。 このプロジェクトでは仮想の試合を何千回も繰り返し、16カ月かけて、パックの摩擦、壁の跳ね返り、モーター応答、カメラの誤差までモデル化。 via：@nexta_tv

🧠Write more about whatever you think on X Note to myself and follower friends I've seen multiple cases of very big accounts launching some trend when others were already writing about it and doing it months earlier. At least post for a track record of being early.

@zarazhangrui nobody can out output codex 😜

A good rule of thumb for preventing AI slop (in writing, design, etc): Is your input (the context) longer than the output? I've found that for the AI to produce quality results, my input is often 3-5 times the length of the output. If your input is much shorter than the output, it's almost certainly going to produce slop

There are a lot of ways to automatically capture data passing through whatever agent system you are using. I wrote this about 3 I use, but in reality could use just one to do multiple parts, and gbrain is similar as I touch on. x.com/willjamesusa/s… There are hook mechanisms in Openclaw and Hermes where you can use their built in process or swap or add-to processes you want to run. It's actually not so complex once you grok the basic idea. All these solutions basically ask an LLM to organize information that is passing through the agents activities as a side job.

I regularly use openclaw for monitoring my business but i've thought about switching over to Hermes because i think the concept of passively saving data about the users preferences and style will be really valuable also to enhance your productivity while coding. I've tried the thing about "second ai brain" but i dont think that's have really worked so for right now i just edit the .md files myself

People here discussing what happened with OpenClaw. The hype died down. We improved quality and grew a team. We created a non-profit whereas competiors are VC funded and have other agendas. This is our strongest week so far.

@lkr Only if the potential penalty is 10x the real line by line review time cost.

Should I a) actually review my 99 page tax return before it gets submitted b) just kill myself right now

A lot of good information video is loaded with filler - just the nature of humans talking. My preferred lesson extraction is: Youtube Transcript -> Key Point Extraction -> #Hyperframes video with ai voice over

Seems to me a lot of teams and individuals are rebuilding this software development loop over and over. OpenAI took a stab at it with Symphony. @aoagents is this. @ryancarson doing this with his code factory framework and @devinai. Many lesser known engineers like @elvissun have demoed this since the spring. I've seen internal corporate initiatives building the loop from scratch. All generally matching on the NASA systems engineering framework I'm just amazed we don't have a dominant opensource solution to the development loop.

If you're curious how I managed to do over $20,000 in inference on the last 48 hours, here's a video all about it. Spoiler: loops are really powerful

This is an easy setup to make memory in your #OpenClaw much better. It includes lossless-claw, adding QMD and #ByteRover.

You can see the entire thing on redfin: redfin.com/CT/Cheshire/20…

What a basement! #midcentury

👀 Houses like this are about to be gone forever! 1/6 🧵

@hyperagentapp How could a video artifact make the work more memorable or playful? I send my boys a hyperframe when they dont do their homework, data pulled from the tracker.

We got early access to the Gemini Omni API Google is calling this model "Nano Banana for video" 3 things we built with it👇 1. Landscaping proposal from customer video. Customer submits a video of the current lot. Hyperagent designs and renders the transformation. Perfect realism, no surreal changes to the surroundings. 2. Animated professor who explains your dashboards. Hyperagent runs analysis on a business question, then generates an explainer video to walkthrough the findings. 3. 8-bit morning briefing. Hyperagent builds morning briefings based on calendar, email/chat, and market intel. Then generates a sidescrolling platformer video showing the goals you'll clear today. Our take: > Video has been a vastly underused input for agents. This is the first model to make video directly malleable to our agents. > Get more imaginative with your outputs. How could a video artifact make the work more memorable or playful? coming soon to Hyperagent

@ScottyBeamIO Do you put all skills into one agent or do you segment by specialist agents within one install?

MOST PEOPLE ARE STILL USING OPENCLAW WHEN HERMES AGENT IS QUIETLY DOING THE SAME JOB BETTER Tested both for a month on the same machine. Here's what nobody is talking about. Reason 1: Hermes gets smarter every time you use it. OpenClaw runs the same skill the same way forever. Hermes updates its own skill and saves the workflow based on what actually worked after every task. A research agent that improves with every single run. It even runs a weekly curator that removes unused skills automatically so it never gets bloated. OpenClaw slowed down the more skills got saved. Hermes doesn't. Reason 2: One command undoes everything. Before Hermes touches a single file, it snapshots your working directory. If anything breaks, /rollback puts everything back exactly how it was. You can technically do this with OpenClaw and Git. It's nowhere near as seamless. Reason 3: Fewer updates, fewer breakages. OpenClaw ships updates almost every week and several of them introduced bugs or slowdowns. Hermes updates less but breaks far less often. OpenClaw still has the bigger ecosystem and more integrations. That's real. But if you want an agent that actually learns and protects your work by default, the gap is starting to show. Bookmark this post. Full comparison in the video below.

@HeyGen @epidemicsound its just money 😆

@willjamesusa @epidemicsound 👀👀👀

The difference between a video that looks fine and one that feels expensive is one thing nobody talks about: Sound HyperFrames now searches music and SFX by describing the vibe, layered in as you build Help us fine-tune the catalogue, what SFX do you use on everything?