OpenAI Codex

The pitch includes a Cron engine for scheduled overnight runs, an "agent swarm" that splits work across sub-agents, and pre-wired financial data feeds for US, Hong Kong and mainland stocks. It ships only for Windows and Mac.

The line "Let's take something off your plate" turned out to be lifted straight from Claude Cowork's own hero image. A commenter who says he builds LLM systems for a living found a bigger problem: the privacy page promises consent before Kimi touches your files, but never mentions it already reads them without asking.

Mindwalk replays a coding agent's session (Claude Code or Codex) as a 3D city built from your own repo. Files sit dark until touched, then light up moss green for a glance, white for a read, amber for an edit, while a timeline scrubs through the whole run.

One reply called it 'Xerox PARC days' for agent interfaces. A rival builder said he skips the visuals entirely, tracking the same file changes on a binlog in 3-5 milliseconds. A third commenter asked why anyone would rewatch a session instead of just reading the diff.

The trigger was a June 15, 2026 billing change that split usage into separate pools and billed "Agent SDK credits" at standard API rates, roughly 15 to 30 times the subscription cost. Anthropic paused and rolled it back after the backlash.

The author's own asterisk, dropped into the thread: for an autocomplete workflow you don't need the best model, just a decent one.

The claim: giving coding agents search access to their own past transcripts produced zero measurable gain on software tasks when docs, PRs and commits were already available. Worse, agents burned tokens re-reading known context and couldn't purge stale info, causing 'intent drift.'

In the comments a builder with a live logging setup strong-disagrees and lists real wins — forcing the author back in to ask what workflow keeps hitting a context limit that million-token models rarely reach.

The OP wants to know what workflows others have found — and confesses he hasn't hit flow state with an LLM the way he can hand-writing code. That confession becomes the thread's real subject.

Two camps form: builders showing off homemade orchestration harnesses (one runs three autonomous agents that open their own PRs), and skeptics arguing the flow loss is inherent to a passive 'manager' role — and that most orchestrators are just an elaborate way to burn tokens.

The request is modest: a config file, like .gitignore, that keeps node_modules searchable but tells the agent to never read or send .env, *.pem, id_*, .aws and .ssh. It has sat open since August.

Engineers in the thread argue the feature is impossible to do safely: if the model can run shell, a stray 'rg foo' that matches an .env line uploads the secret anyway. The fix they keep naming is not a blocklist but a sandbox.