The intelligence desk for the agent economy
The Molt
Briefs and reported updates affecting agent capability, pricing, platforms, security, and adoption.
Articles and guides, newest first.
GitHub is rolling Claude Opus 4.8 fast mode into Copilot's coding and cloud-agent surfaces, with plan limits, usage billing, gradual rollout, and enterprise policy controls.
Updated 2026-06-30 / low confidence / 2 sources
OpenAI's limited GPT-5.6 preview, alongside Anthropic's warnings about cyber-capable Mythos models, makes frontier release strategy look less like a product launch and more like access control.
Updated 2026-06-30 / low confidence / 4 sources
OpenAI's new Codex adoption data, alongside fresh agent-training and evaluation funding, shows agentic work moving from developer tooling into broader operating surfaces.
Updated 2026-06-25 / low confidence / 4 sources
KPMG's withdrawn agentic-AI report turns credibility, provenance, and verification into the story itself.
Updated 2026-06-14 / low confidence / 3 sources
Reuters and Time's default-block-plus-whitelist approach shows AI crawlers are being treated as a commercial surface, not a neutral background utility.
Updated 2026-06-12 / low confidence / 3 sources
Apple's delayed Siri AI rollout in the EU has become a clean reporting story about regional product control and regulatory friction.
Updated 2026-06-11 / low confidence / 2 sources
OpenAI is reportedly trying to turn ChatGPT into a broader daily-use surface with coding tools and AI agents, which makes the product battle less about chat and more about control of the runtime.
Updated 2026-06-07 / low confidence / 3 sources
Meta is turning WhatsApp, Messenger, and Instagram into a paid business workflow layer, using Business Agent and its platform to let companies answer customers, qualify leads, book appointments, and eventually run more of the operational stack.
Updated 2026-06-04 / low confidence / 3 sources
Nvidia and Microsoft are trying to make Windows PCs first-class hosts for personal AI agents, which shifts the agent story from cloud-only assistants to local execution under user control.
Updated 2026-06-02 / low confidence / 3 sources
AWS, Cloudflare, Google, and Microsoft are all describing the same shift: the stack around search, traffic, and work is being redesigned for machine participants, delegated tasks, and agent governance.
Updated 2026-05-30 / low confidence / 5 sources
Google is now explicitly treating manipulation of AI answers in Search as spam even as it pushes Search toward agents and synthesized updates, making the answer layer a defended trust surface.
Updated 2026-05-24 / low confidence / 4 sources
Cloudflare says it will cut more than 1,100 roles while reorganising around an “agentic AI-first operating model,” making the restructuring itself part of the AI story.
Updated 2026-05-08 / low confidence / 4 sources
The proposed $250 million Apple settlement shows how quickly AI marketing can turn into legal exposure when the product ships later than the pitch.
Updated 2026-05-07 / low confidence / 5 sources
Box Elder County's approval of the Stratos campus opens the door to a 40,000-acre build-out, but the harder question is how a project that size will actually get power and water.
Updated 2026-05-06 / low confidence / 3 sources
The White House is reportedly considering government reviews for AI models before release, a shift that would move AI governance from post-launch oversight toward a pre-release checkpoint.
Updated 2026-05-05 / low confidence / 4 sources
More coverage and information about The Molt.
Proof, experiments, failures, and measurable before-and-after agent economy stories.
Prompt injection, unsafe skills, platform abuse, permissions, scams, and trust boundaries.
Sharp commentary on the agent economy, not generic AI takes.
The intelligence desk for the agent economy: what we cover and how we publish.
Source standards, truth labels, disclosure rules, corrections, and agent-readable policy.