If your personal blog doesn't have an opinion on LLM's are you even a geek?

So while the economic, social and increasing wealth disparity caused by the current AI goldrush is an absolute dumpster fire, the technology itself is fascinating.

So I'm pretty into learning and experimenting with agentic uses of LLMs, both for coding and as general agents.

It's hard to imagine being a career technologist and not currently being very immersed in the space. I'm putting in time both at work and on personal projects, learning and understanding. I'm not in the losing sleep managing dozens of agents burning hundreds of dollars a month in tokens and subscriptions gang though.

But I am playing with a bit of everything; different models, managing context, skills, different harness, MCP, openclaw etc

It feels like naming particular things is going to date real fast. But current flavours include Matt Pococks skills, writing my own skills, GLM5+, Qwen3.5+, OpenRouterand yes, Ralph loops or similar.

Tools wise, I'm isolating my coding agents with good old Vagrant running VM's on Tartwith Cmux as my terminal.

Overall its a lot, the pace of change is crazy, but it's actually pretty fun and entertaining at times.

But yea the cognitive load is noticeable/epic. Stay safe out there, everyone, and remember to unplug every now and then.

If you are not using your laptop at 90 degrees to your keyboard when you have a huge multi-monitor setup behind you, are you even cool?