People in my org are excited about an org MCP we built that lets you publish html docs to either a public or internal bucket But if you don’t want to build it yourself, Moda looks like a cool option

CEOs are uniquely prone to AI psychosis because they’re sufficiently distant from the last mile of work that still has to happen to generate most value with AI. So when they play with AI, they see the happy path results, often not considering the next 10 or 20 things that have to happen to get sustainable results from agents. “Look I made this awesome product prototype”. Yes but you didn’t have to review the code before it went into production and fix a bunch of issues. “Look I generated a contract”. Yes but you didn’t verify all the terms before it goes out to the counterparty and didn’t have to wire up all the past contracts to work with. The best thing you can do as a CEO is to use AI a *ton* to figure out the real implications of agents in the enterprise, and come out the other side with an appreciation for both the upside and the real work that goes into them.

I gave HTML over Markdown a try and now I see there are a lot of advantages. I’m prototyping a couple new tools in our org MCP to allow any user in our org to publish an HTML file to an Cloudflare R2 bucket to share with their teammates. It’s meant for one time artifacts, but you can deploy little single page apps with it too, haha

Killer article from @cathrynlavery on @ppressdev. A printed CLI is two things at once. The CLI is the deterministic core, same input same output. The skill on top teaches Claude Code, Codex, Hermes, and @openclaw to drive it. One install, every agent.And the real tell: we're printing them for many surfaces that never had an API.

I was an API + skill maximalist but there are some great points here for CLIs. Particularly if you use @ppressdev to make them.

Remote control was one of the biggest things I missed when I gave Codex a try from Claude Code…just released!

Not long ago I thought UI workflows like Zapier and Retool were convenient. Now after 1 min in the browser and I’m annoyed I can’t drive changes with Claude Code. I even tried having Claude slowly make changes through the browser and even it started complaining 😂

This is an incredibly approachable breakdown of how @openclaw works

Running /insights in Claude Code on a regular schedule is a great way to improve. Claude Cowork and Chat don’t have that command, but you can teach it how to pull past sessions in each tool and suggest skills, projects, or scheduled tasks you should be using.

I’ve been doing more work with documents, spreadsheets & slides If I can get a Microsoft office file on my laptop (sync’d with Box Drive), Cowork is great for big changes Claude in Word is nice for fine tuning With Google Docs, you’re stuck with the API and it’s painful

This is a great course! The examples are just complex enough to make it interesting. I have several modules bookmarked and refer back to them often

Wow Cloudflare is really pushing things forward

Built a “Code Mode” HubSpot MCP for my org, modeled on cloudflare/mcp, then ran a pipeline review with Claude and asked for feedback vs the default HubSpot MCP: >Roughly 3–5× fewer tokens and one round-trip instead of three. The bigger win isn't the token count — it's that filtering, pagination, and aggregation move out of the model's context and into deterministic JS. Two tools: search the API spec, and execute JS in a sandbox against HubSpot. The trick is borrowed straight from Cloudflare’s Code Mode post — the token stays server-side and the agent only ever calls a hubspot.request(...) helper, so I can lock down endpoints and it can’t see the secret or hit anything outside HubSpot.

Quick reset: this account used to be tied to my Jack Gray pseudonym/ StakNine project. I’m using it as myself now.

The official Elixir Language Server Protocol implementation GitHub repo is public!

Phoenix LiveView v1.1 is out: colocated JS, portals, improved change tracking, faster tests and more! Read the announcement: phoenixframework.org/blog/ph…