Samuel Stroschein

@ganstlr you could write a plugin for lix github.com/opral/lix

why does git for latex not exist? feels very useful in a large research team capacity

matches my observations as well. filesystems for agents solve infrastructure issues but are not a (consumer-facing) product on their own

@weird_ceo the company as repo makes click. you might want to try deepline.com and model your crm locally in the repo for better context any issues you are running into with your stack?

I turned my company into a GitHub repository. Every recurring process — daily planning, sales ops, customer responses, sales answers, RFPs, security assessments, client onboarding, meeting follow-ups, product update emails, competitive battlecards, weekly investor updates, candidate screening — is a markdown file in a repo. AI agents read them and run them across our tools. At SegmentStream, every team member commits every day. Doesn't matter if it's sales, marketing, or customer success. If you produced any work for a client that can be repeated, you commit it — as a skill, an example, an artefact, anything that feeds the agents. The repo IS the company. The minimum viable stack to copy this: → @linear MCP (tasks) → @attio MCP (CRM) → @SlackHQ MCP (communications, messaging artifacts) → Gmail MCP (email - private inbox, email artifacts) → @resend MCP (email - broadcasts for agentic workflows) → Google Calendar MCP (scheduling meetings) → Google Drive MCP (signed contracts, etc)@ → @firefliesai MCP (meeting artifacts) → @github MCP (primary storage for all company data) → @Xero MCP (accounting data) → @segmentstream MCP (marketing measurement) → @claudeai Code or @OpenAI Codex to orchestrate it all Write your processes in plain English. One file per process. Review them in PRs like code. A month in: my morning queue scan, my customer drafts, my daily plan, my meeting bookings, my RFP responses, client onboarding, performance reviews — all run by agents, reviewed by me. The ops team of the future isn't people. It's a repo with good PRs from every department. The next step — train fully functional AI Operational Director. Any useful agentic-first apps I'm missing in my stack?

company as a repo 🤌

graphite is building a github competitor

from their blog post github.blog/news-insights/…

The number of pull requests, commits and new repositories on GitHub doubled over the past 4 months due to AI agents Clear argument for "AI agents need new infrastructure"?

@bwarrn congrats on getting the beta out!!

I'm so excited to officially share what we've been cooking up at Mesa: the most powerful filesystem ever built for AI agents. The dirty secret of every "production" AI agent today: the filesystem is held together with duct tape. Teams are stitching together S3, GitHub, sandbox-local disks, and homegrown diff logic to give their agents something resembling persistent, versioned storage. None of it works. S3 isn't designed for parallel agents - concurrent agent writes silently overwrite each other. GitHub has the semantics but rate-limits you into the ground at agent scale and doesn't give you filesystem ergonomics. Sandbox disks vanish the moment the container dies. And your agents don't want to git clone and git push anyway. 𝗧𝗵𝗲𝘆 𝘄𝗮𝗻𝘁 𝘁𝗼 𝗿𝗲𝗮𝗱 𝗮𝗻𝗱 𝘄𝗿𝗶𝘁𝗲 𝗳𝗶𝗹𝗲𝘀. 𝗟𝗶𝗸𝗲 𝗲𝘃𝗲𝗿𝘆 𝗽𝗿𝗼𝗴𝗿𝗮𝗺 𝗲𝘃𝗲𝗿 𝘄𝗿𝗶𝘁𝘁𝗲𝗻. So we built the missing layer. Mesa is a durable, POSIX-compatible filesystem with version control built in. Branches, diffs, history, rollback, access control — every primitive a codebase has, for any file type, at agent scale. You mount it. Your agent uses it like a normal filesystem. We handle the rest. Private beta is live. Link in comments.

@_CallMeMacy yeah overlaps with us churning from attio. interesting pattern the remaining benefit is the data model (which claude can copy in one go) and integrations like last contact via email etc. the ui and their ai features became completely irrelevant

@samuelstros I guess the question is what value am I paying for at this point?

Lately, I've been advising founders to vibecode their own CRM that reads their emails, calendars, and keeps things up to date in real time. It took me 2 hours to create my own using Claude Code (and I've never coded in my life). Am I crazy for suggesting this or do others agree?

@francescagnuva @grok who do you side with?

Damn. GPT siding with Elon in the Musk/Altman feud LMAO

@danielwise92 @ZohranKMamdani @CitiBikeNYC is already operating on a 2% margin. Not much “rent seeking” from their side. they should publish where their costs go. Maybe the city could wave fees or make operating citi bike cheaper other wise

@ZohranKMamdani As someone who likes to use citibike as much as you do is your office coming up with a plan to make citibike a viable affordable means of transportation for NYers?

perfect weather to bike home

@lizwandersworld congrats on launching!!

Today we're launching Era and announcing $11M in funding. We're building the intelligence layer for a new ecosystem of AI devices — the platform that lets any device manufacturer, brand, designer, or creator make objects that think, respond, and act in their own style. We're entering a Cambrian explosion — new form factors, new creators, new objects worth desiring. Made by people who've never had the tools to make them, before today. Welcome to the new Era. Backed by @AbstractVC, @BoxGroup, @topology_vc, @betaworks, @CollaborativeFund, @MozillaVentures, and @AIResidency.

@rincidium congrats!!

Bloom is joining YC for the Spring batch. We’re building the future of branding and marketing. The vision is for Bloom to be the “Brand Operating System” - a system that deeply understands your brand and powers everything it touches. Since launch, over 40,000 people have used Bloom to create on-brand assets. Our revenue has tripled in the past month. Thank you @koomen and @ycombinator for the support. We’re building our team in San Francisco. If you want to own massive scope and help build something huge, link below.

@enrique_goudet @lixCCS 1 weekly active user on @posthog 🚀

first user of @lixCCS by @samuelstros

@rabanspiegel 🚀

@jik_wtf @HSVSphere what is your use case? why wouldnt git work for non code text files?

@HSVSphere I'm about to wrap _something_ to get version control over some text files that aren't code. Could I wrap JJ for this? E.g. are JJ merge conflicts strictly a subset of git conflicts? or am I out to lunch and don't get it at all

Git is just bad. Use jj, and stop complaining about concepts when the implementation is bad.