I got sick of writing the same weekly plan over and over, so I built an app that roasts me for the gap between what I said I'd do and what I actually did.
It's called GrindProof. It's live. It has no streaks, no badges, no mercy.
https://t.co/p2EGElVxde
We built a cyber-capable AI model that, based on benchmarking and real-world testing, is perhaps the most capable offensive-security LLM we have seen
It scores 95% on Cybench, solving 38/40 tasks
But benchmarks only go so far, so we pointed it at real systems
Through authorized programs, we tested it against a Fortune-500 technology company, a major global telecom, and leading Web3 protocols.
The result: a 100% success rate - it found a viable path into every authorized target in our test set.
It adapts, reasons, and finds routes that aren’t obvious, even in systems that already passed audits and security reviews.
If this holds at scale, most current security assumptions break.
Privacy note, since it reads your setup: read-only, local-only, no telemetry — the output is one offline HTML file. The optional classification call sends item names/descriptions only, never transcripts. Rough mode sends nothing at all.
I asked my AI coding setup what it's actually good at.
103 skills, agents & MCP servers installed. 90 never fired in the last month.
One command draws the whole thing as a map:
npx agent-atlas-cli
Works with Claude Code today. Open source (MIT), adapter interface ready for Codex CLI / Cursor — PRs welcome.
npx agent-atlas-cli
Repo: https://t.co/ilH5zZ1h8I
Deep dive: https://t.co/mRa9a3usNK
ORGN is evolving.
We're going from a confidential AI IDE to a full confidential agentic stack: infra, agents, apps
Confidentiality is by design, down to the silicon.
ORGN CDE is now just one piece of the system.
It's free right now. If you've rewritten the same plan three Sundays in a row, this is for you:
https://t.co/JgaW3m8mnH
And if you try it — tell me if the roast lands. I read everything.
I got sick of writing the same weekly plan over and over, so I built an app that roasts me for the gap between what I said I'd do and what I actually did.
It's called GrindProof. It's live. It has no streaks, no badges, no mercy.
https://t.co/p2EGElVxde
I built it because the only thing that ever worked for me wasn't another app being nice to me. It was being looked at — someone who knew the plan and asked whether I did it.
GrindProof is that, automated.