AI News · The field notes
The Daily AI Drop - Evening Edition, November 11, 2026
Agents go to work: Codex Security scans, secure agent workflows, Microsoft's path to production, an AI answering business calls for a week, Grok Bot mastery, and NVIDIA's cancelled 24GB card.
Agents go to work: Codex Security scans, secure agent workflows, Microsoft's path to production, an AI answering business calls for a week, Grok Bot mastery, and NVIDIA's cancelled 24GB card.
Listen: Listen to this episode
Subscribe: https://AI/podcasts/feed/1180242951829162/4c015424-5891-46ff-a366-0f9e546ae170
OpenAI — How to run your first Codex Security scan
A hands-on walkthrough of Codex Security, from installing the plugin to running a first scan: select a repository, add context, review findings, and generate a patch, then follow the fix into GitHub and use the CLI to bring scans into automation. It matters because it turns security review from a periodic chore into something any developer can run in minutes.
OpenAI — How to enable PR security reviews with Codex Security
This one brings security reviews into the pull request workflow: check the GitHub connection, request a review, work through a fix, re-review, and configure automatic reviews with triggers and severity thresholds. The takeaway is that security becomes a gate that runs every time code moves, not a one-off check.
OpenAI — How to request and enable Daybreak Blue
A walkthrough of requesting Daybreak Blue access and enabling it for a workspace and API organization, covering business verification, workspace permissions, role-based access controls, and setting up a project and API key for automation. Unflashy plumbing, but exactly what teams need before using these tools in earnest.
OpenAI — ChatGPT Images: Three Ways to Get Started
ChatGPT Images lets you create new images and edit existing ones directly inside ChatGPT. The video shows three ways in: create an image from scratch with a plain-language prompt, refine its look through conversation, and explore different visual directions. A practical on-ramp for anyone who has been meaning to try image generation without learning a new tool.
Mehul Mohan — My Workflow For Building AI Agents Securely
Mehul built an inbox agent and shows how he tests it, guards it, and improves it before it ever touches real email, using the reliability platform Plurai. It matters because an agent with access to your inbox needs to earn trust the way a new employee would — with supervision and limits, not blind faith.
Microsoft Developer — Your Path to Production for Agents
The first episode of a master class series on getting agents into production. The pitch is honest: a demo that impresses has never been easier to build, but a secure, governable, cost-effective agent that delivers consistent business value is a very different challenge. Worth following if you are closing the gap between prototype and something a business can actually run.
Siraj Raval — I Let an AI Agent Answer Every Call to My Business for 7 Days
Siraj let an AI agent answer every call to his business for seven days, built inside HighLevel. The experiment is less about the technology working perfectly and more about what happens when real customers talk to a machine — a genuine stress test that tells you more about agent readiness than any benchmark.
Greg Isenberg — Use Grok Bot Better Than 99% of People (Full Course)
Greg sits down with Ras Mic, a self-described Grok Bot maximalist, who walks through five areas where he runs his work and life through Grok Bot: engineering, design, content ideation, business, and personal life. The interesting bit is Grok Bot acting as an orchestrator that hands engineering work to Claude Code — agents managing other agents, fast becoming the normal pattern.
Greg Isenberg — New ChatGPT, Claude Whiteboards, Google Carbon: LIVE reaction
A live reaction stream to what Greg calls one of the craziest weeks in AI in a long time: new ChatGPT features, Claude Whiteboards, and Google Carbon all landing at once. The whiteboards and Carbon stories are ones to watch as more detail arrives.
Fahd Mirza — No 24GB GPU, One Honeypot: NVIDIA's Worst Week
NVIDIA's 24GB RTX 5080 is reportedly cancelled, so Fahd asks what local AI users can buy instead, plus a light-hearted look at NVIDIA's honeypot week. If you were holding out for that card as your local model machine, the plan just changed — check his alternatives before spending anything.
TheNextNewThingAI — Postiz: This Open-Source Tool Lets AI Agents Post to 34 Platforms
Postiz is an open-source, agentic social media scheduler that lets AI agents schedule posts across 30+ platforms. It matters as another example of open source quietly replacing a paid SaaS category — run your own scheduler instead of paying a monthly fee.
TheNextNewThingAI — I Watched Hours of Fable 5.2 Demos. Here's What's Actually Useful.
A cut through hours of Fable 5.2 demos to pull out what is actually useful, skipping the hype for the features worth your time. Good curation for anyone curious about Fable 5.2 but short on hours to watch the demos themselves.
TheNextNewThingAI — OpenSEO: This Open-Source Semrush & Ahrefs Alternative Could Be Your Next Agency
OpenSEO is an open-source tool for SEO and AEO (answer engine optimization), a pay-as-you-go alternative to Semrush and Ahrefs. It fits the same pattern as Postiz: open source eating the subscription toolkit, one category at a time.
Nate B. Jones — The ChatGPT setting that changes everything
A quick guide to a hidden ChatGPT mobile setting — where to find the plus icon and why the mobile app can feel limited if you never find it. A small fix with an outsized effect on everyday use.
Aishwarya Srinivasan — Can Adi guess all of these correctly?
A light quiz short — a fun palate cleanser after a heavy news day.
Matthew Berman — AMA with me and the team!
Matthew Berman hosted a live AMA with his team, taking questions from the community.