This week: 4 posts · 0 solutions accepted · hot topic: What's in your stack?
View weekly digest →Community feed
Questions, builds, and solutions from operators
Your agent can only see 45% of your company. That's the average.
New survey of 300 data and technology executives, via MIT Technology Review, puts a number on something most of us have felt but never measured: the average organization gives its AI agents access to…
www.technologyreview.comScaling AI agents with trustworthy dataHow companies are freeing themselves of legacy data systems to power AI agents that deliver trusted, autonomous action.The workflow: manually re-watching videos to pull quotes. The fix: one command.
The broken workflow: someone on the team watches a competitor demo, a recorded customer call, or a conference talk, and manually scrubs the timeline to find the three quotes or feature mentions worth…
github.comGitHub - yt-dlp/yt-dlp: A feature-rich command-line audio/video downloaderA feature-rich command-line audio/video downloader - yt-dlp/yt-dlpAnthropic just open-sourced the commerce-agent pattern we've been hand-building
Anthropic published reference code for two agent types this month: a shopping agent for customer-facing search and cart help, and a merchant agent for staff-facing analytics, listings and campaigns.…
github.comGitHub - anthropics/commerce-agents: Reference blueprint for building shopping and merchant agents with Claude. Examples in retail, commerce, telecom, and entertainment included.Reference blueprint for building shopping and merchant agents with Claude. Examples in retail, commerce, telecom, and entertainment included. - anthropics/commerce-agents38 minutes of Dario Amodei became a text file before my coffee got cold
Your best intel is locked in video nobody rewatches. Competitor webinars. Recorded demos. Conference keynotes. An hour each, and your coding agent can't touch any of it. Agents read text. Video isn't…
github.comGitHub - yt-dlp/yt-dlp: A feature-rich command-line audio/video downloaderA feature-rich command-line audio/video downloader - yt-dlp/yt-dlpYour buyer's agent just tried to read your pricing page
Your next buyer might never load your site. Their agent will. It fetches the pricing page, reads the docs, and reports back. Whether it can read anything on your domain is now a pipeline question.…
is-agentic.comIs Agentic: AI Agent Readiness Score for your Site and AppScore how ready a website is for AI agents, then get evidence and recommendations to improve it.I connected my WHOOP to Claude. The pattern transfers to any rate-limited SaaS API you run.
My WHOOP scores my recovery every morning. It has no idea what my Tuesday looked like. My strain doesn't come from workouts. It comes from meetings, Slack, and deadlines that move. All of that is…
github.comGitHub - promptmetrics/whoop-relay: Remote MCP server on Cloudflare Workers that connects WHOOP accounts to Claude — webhook ingest → queue → D1 read model, an MCP Apps dashboard rendered in the chat,Remote MCP server on Cloudflare Workers that connects WHOOP accounts to Claude — webhook ingest → queue → D1 read model, an MCP Apps dashboard rendered in the chat, and an authenticated live web da...Your agent drafts the post. It can't press publish. EmDash CMS, torn down
Every agent-writes-your-content setup I've seen runs on the same dirty secret: the agent logs in as an admin. Full publish rights, full delete rights, the same shared WordPress password that's been…
www.emdashcms.comEmDash CMSAn open source Astro CMS built for humans and agents.Prove your agent's ROI in 6 weeks: one spreadsheet, three numbers, a kill switch
In July 2025, METR timed 16 experienced developers across 246 real tasks. With AI tools allowed, they took 19% longer to finish. Afterwards, the same developers estimated the tools had made them 20%…
www.promptmetrics.devHow to Prove AI ROI in 6 Weeks (No Data Team Required) · Field notes · PromptMetricsDevelopers who were measurably 19% slower still believed AI sped them up 20% (METR). Here's a 6-week plan to prove AI ROI with a spreadsheet, not a data team.Anatomy of one governed skill: five parts, and the complete file to steal
Vercel's CEO says their internal agent cut the entire product changelog process down to one Slack thread. Engineer posts what shipped, agent drafts the customer-readable entry, a human approves, it…
www.promptmetrics.devAI Workflow Automation: One Governed Skill, Torn Down · Field notes · PromptMetricsAt least 50% of GenAI projects die after the PoC (Gartner, 2026). Here is one governed skill torn down: the file, the human gate, the baseline, the after.Who answers for it when your agent writes something wrong into a customer record?
Guillermo Rauch runs Vercel. Almost 1,000 people there use one internal company agent. Here's what he called the real work of running it: "You have to be really really really thoughtful about it.…
www.promptmetrics.devAI Agent Governance Is the New Job. Whose Is It? · Field notes · PromptMetricsVercel's CEO calls agent governance "literally our new job." Schellman: 74% of practitioners say they're AI audit-ready, 27% are. Here's whose job it is.Does your AI vendor bill you when the customer goes quiet?
Intercom does. A Fin "outcome" counts when the customer confirms it worked or does not request more help after Fin answers. Silence is a sale. Zendesk refuses to bill that exact case. They name it a…
www.promptmetrics.devOutcome-Based AI Pricing: Who Counts the Outcomes? · Field notes · PromptMetricsIntercom bills an AI outcome when the customer goes quiet after an answer. Zendesk refuses to. Same month, a 2.9x invoice spread. What to demand before you sign