So Hermes refusing to enter verification codes was never a mimo-2

2026-10-01 03:22

So Hermes refusing to enter verification codes was never a mimo-2.6 problem. It's a security restriction baked into the new Hermes build. Now every time I want Hermes to fill in a code, I have to drop into a shell, add the one-time code to the Hermes Vault, then remember to delete it afterwards. Nobody who shipped that design actually tried using it. Since v0.21.2 (2026.9.11), Hermes Agent hard-bans any password or verification code sent over chat. Including the ones I send myself. So every site that wants an email or SMS code at login just hangs on prompt_unavailable. No toggle in config to switch it off. The only workaround is patching locally to kill the Vault. Issue #120704 has been open since September and nobody has touched it. github.com/NousResearch/hermes-agent/issues/120704

Gemini 4 Argon has to clear the US government's pre-release

review too. Google's own line: it's "actively engaged in the U.S. government's voluntary process for pre-release model access." For closed frontier models in the US, that review is the default now. GPT-6.1 Astra never even shipped over safety concerns. So is Gemini back at the front of the pack?

2026-10-01

Gemini 4 Argon is out.

Paper numbers look decent: $2/M in, $10/M out, output cap up from 64K to 1M tokens. It tops DeepSWE v1.1 at 77.9% (real-world software engineering) and LVBench at 91.7% (long video). Outright first on 18 evals, tied on 13. https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/

2026-10-01

Claude appeal rejected.

Won't refund the subscription either. I'm half convinced the appeal email is itself phishing, because I'm pretty sure they log the IP. DHH came to Shenzhen recently and showed how he uses Claude Code over there. That's the safest way now, honestly: run Claude Code on an overseas machine, connect through Tailscale, never use it locally. So I need to rent an overseas box with a residential IP, and register and log into Claude there.

2026-11-14

Sharing my current mobile only workflow.

Until late September it was the usual setup: parked in front of my laptop, terminal open, bossing pi around, fanning out subagents, then eyeballing diffs myself and filing the PR. Glued to the chair all day, every day. Late September I moved the whole thing to my phone, Hermes runs the show, pi takes the orders. My back is very grateful. Three weeks in, still no regrets. Stack: Hermes on flagship brains (gpt-6-astra / opus-5.5), pi on mid-tier (ds-v4.1-flash / sonnet / terra) with as few plugins as possible, the two talking over a2a. Hermes reads the issues, slices them into small tasks, spins up a worktree for each one, and dispatches a pi into it. When a pi finishes it pings Hermes, who decides: ship the PR or bounce it back for rework. I get back from a full day of being a tourist, or just wake up, and a dozen PRs are sitting there done. Pretty sweet. The catch is how black-box the whole thing is. My emotional arc has been a rerun of early vibe coding, except back then I at least read the diffs. Now even that goes to a flagship model. Boxes all the way down. If you don't have data anxiety, grok-bot and chatgpt-dot basically ship this out of the box, grok-cli and codex built in, native agent-to-agent comms. Downside: your code lives on their servers, everything's moderated, and one ban takes it all with it.

2026-11-14