This

$GRASP PR to @openclaw. it needs some prompt turning but I'm comfortable with the v1 implementation. If you use openclaw and want to better understand the risks that you're exposed to (not so you necessarily stop but so you can patch what you want and accept the rest) please

What if risk and safety weren't gatekept by enterprise. I tasked the factory wth making a PR for openClaw that allowed users to run a GRASP analysis over their agent/s. Likely will take some time to merge and needs some fine tuning but a strong start. As with all GRASP

Just design to be afk, that should be your target mode

Secure slack integration for the ATLAS coding factory up and running (I lovingly call my factory the cookpit (Thanks @dexhorthy )). I can now trigger my very opinionated SDLC workflow from slack anywhere in the world.

Think of it like the shift from horses to cars. Those who get it aren’t asking whether cars are better than horses. They’ve already started building their companies around the new technology; supply chains, workflows, skills. The question is settled. Those who stay

NSW Govt has just released their AI Risk framework and my fear is this moves us in the wrong direction. I've reached out to the CSIRO and Digital NSW to discuss further but for me this is going to falling to the vicious cycle of: - Governance teams forces IT to answer too many

Morning all, Another week. Over the weekend I was thinking about the project and at the moment it's kind of conflating the construction of AI Factories and the governance of AI agents. I don't think you can build AI factories without governance but you do need governance even

Adding an SDK too so everyone can play with factories

After talking to a bunch of big businesses it's becoming scarily clear that there is a bifurcation occurring between those that get it and those that can't. On the innovative edge you have wild conceptual revolution, what is a computer anymore? what does an efficient programming

After doing a bunch of code reviewing yesterday I added another Claude max plan a started building an SDK for the ATLAS system too. The idea being that I want to provide interfaces that allow people to implement factories how they want to implement them with common standards

Full RYORA Interview On @MCGlive Today: Agent Oversight Platform With a background at Optiver, one of the world’s leading market makers, @GoblinRack onboarding Web3 with a strong from a place of deep operational rigor. As even Microsoft’s CEO has publicly warned, companies are

Thanks for having me on @MCGli @mynt_josh & @DineroDom0. Hope I made sense for the morning and I appreciate you letting me go on about the sexy world of AI governance. To be clear, governance isn't sexy, and I'm sure most people don't want to talk about it, it's why they get

This was the lab when I was building the house.

It 4:30am, I suppose that's a good as time as any to go on @MCGlive https://t.co/2FcOQyr5wp

https://t.co/pTeZsQgGtG Didn't get a lot of time to build today, thankfully the loop is doing most of the heavy lifting, tomorrow gitflow

https://t.co/TRoWLCfPlR

Yeah I use AI overdubs, but I hate my voice on recordings. This is content piece 1 being dropped tonight. It's focused on a more LI audience, meant to scare the shit out of them. AI is moving orders of magnitude faster than enterprise can keep up with (Without help).

https://t.co/Gj9C2M9chN AKA building multi agent systems is hard. You have to really consider what roles they play and how that factory pipeline works.