This is the most practical approach I've seen to actually building an AI native work operating system that people use Huge contribution that Ramp shared this!

Ramp is on 🔥 with agent pilling the org - so far ahead of most cos and much to learn... One of key 🔑 is not only making it easy to let engineers get max AI/agents from day one but everyone else and then making all work which are skill easily reusable, shareable and version

Show me another company/organization doing this. Putting compounding space between themselves and their competitors. And here is the kicker, these moves dramatically increase odds of entering businesses and new opening revenue lines that can’t be fathomed right now. What’s

This playbook by Ramp is 100x bigger than the Claude Code leak 🧠

Is anyone else losing their minds watching Ramp produce such interesting research as they implement their own harnesses and agentic systems internally? Very impressed by Ramp Labs

99% of Ramp uses ai daily. but we noticed most people were stuck — not because the models weren't good enough, but because the setup was too painful and unintuitive for most. terminal configs, mcp servers, everyone figuring it out alone. so we built Glass. every employee gets a

This is what an AI first company looks like.

don’t just build a differentiated product, build a differentiated company

Ramp built their own internal Claude Cowork and Openclaw competitor. And it seems way better??? Product over everything.

it’s not just marketing. it’s actually that good btw

Ramp must be the largest company to transition from a traditional software company to an AI-native software company thus far. This transition will absolutely be mandatory business school reading

This is brilliant. Being the go-to person to implement this across companies and organizations is an obvious million dollar opportunity. Reminds me of when Michael created a million dollar business overnight by going and around setting people up with OpenClaw🦞 on

One of the interesting things we've observed from building glass is that the way to build agents is pretty consistent regardless of team / problem type. 1. Curate some basic context for how to approach the problem, and make sure it has the right tools (mcp, cli, etc) to

For those who've asked, sharing a great write up from Shane Buchan on how we built glass! No big engineering team or massive investment, just a small crew having a lot of fun and pushing vibe coding to its limits. Vibe coding is a skill like any other and you can build some

Models smart enough to cure cancer, UX bad enough most people can't get a decent email out of them. Seamless UX is going to be a bigger moat than people think.

Spent a few days in New York catching up with folks: 1.) Few have heard of Mythos 2.) While people have heard of Claude Code, few have tried Cowork or knows what a skill is 3.) Many F500 non tech companies here aren’t getting any value out of AI because of politics 4.) Too many

We held a company-wide AI hackathon and gave everyone tokenmaxxing permission. The results were... insane.