Building AI-powered software that ships. Full-stack + AI partner for startups & enterprises—from idea to production with ops automation & strategy consulting.
How do you get Claude Code to check its own work before handing it back?
Watch how you can encode your manual checks so Claude closes its own feedback loop:
Raw updates migrating hermes from openclaw
--> used hermes migrate: move off openclaw
--> ported discord gateway
--> switched model to codex: will use gemma/minimax
--> imported all claude code skills to hermes
--> setup initial llm-wiki (video below)
--> porting existing obsidian over
"Yeah but where's your MRR bro?!"
X and open source are for audience building & experimenting - the learnings get baked into community and product.
Purpose built Hermes agents for auto research is my favorite so far
start moggin github issues with your claude code max subscription and docker sandboxes
bun install -g @bobbyg603/mog
mog init
mog workingdevshero/automate-it 123
from issue to PR with the power of claude code and the safety of docker sandboxes
learn more 👇
https://t.co/dAzDuScldX
Use Claude Code and MCP to make Revit drawings!
Had a lot of fun working on this and can't believe more people aren't using MCP for CAD. It's time to build!
@SHGI_SparxFire@workingdevshero
Your AI team is assembling. Behind the scenes, a new era of automation is taking shape—agents working seamlessly across every platform you use, executing complex tasks with precision. The future of work is almost here, are you ready? 🤖
Davos debate: Tech leaders clash on AI's path forward. Elon Musk predicts AGI this year, while Yann LeCun argues language models aren't enough. The divide shows how uncertain AI's future really is.
https://t.co/UEHFRpoTyj
Tech's AI spending is reaching staggering levels. Big tech companies are projected to spend over $110 billion on AI infrastructure in 2026, with Goldman Sachs forecasting as high as $125 billion—a 57% increase from 2025.
Read more: https://t.co/MrgJNWUBKu
AI hallucinations slip into top research: At least 100 fake citations found in 51 papers accepted at NeurIPS. Fictional author names and non-existent DOIs bypassed peer review, exposing publishing vulnerabilities.
https://t.co/zyl3Z1UZht