GitStart
@GitStart
GitStart accelerates engineering teams by taking tickets and turning them into ready-to-merge PRs created by people + AI agents ⚡
Very similar ethos here @GitStart ... we just shipped a new feature this afternoon for a customer that signed up today morning. We only merge code (hence the delay). GitStart had the PR ready in minutes. If your stack can't do that, ngmi! P.S every user get's slack support.
This post has been pinned to every virtual wall we have.
Or you can experience the future today, by looping in real engineers to build an end to end service (aka @GitStart) The real leverage goes up to 3-5x per team member scoping work and reviewing PRs!
Congrats on all the progress @SudoHamza and @GitStart team! Awesome to see how ya'll are building fast, staying focussed, and delivering practical, truly helpful ways of scaling engineering teams. 👏🙌
You are staying up all night fixing AI slop, while @GitStart PRs are getting merged to production! Here is how🧵
Coming soon to other chat clients near you 🤫
How I use @GitStart to collab in @SlackHQ and ship features
You are staying up all night fixing AI slop, while @GitStart PRs are getting merged to production! Here is how🧵
@GitStart got all of it done. 95% LLMs with 5% human assist (< 1 hr) LLMs are nearly there. Just need a little push. github.com/calcom/cal.com…
no it can't
@peer_rich we opened a PR also: github.com/calcom/cal.com… Generated by @GitStart AI and QA + touch ups by Rubens Rafael and @toledo99_
🔥🔥🔥
Feedback on @GitStart from eng team building a large mobile app (500m downloads, top 5 App Store): "GitStart got the first PRs shipped faster than new engineers get our app running on their laptop!" Solid Engineering with Code Reviews & QA, at the speed of AI Agents!
Feedback on @GitStart from eng team building a large mobile app (500m downloads, top 5 App Store): "GitStart got the first PRs shipped faster than new engineers get our app running on their laptop!" Solid Engineering with Code Reviews & QA, at the speed of AI Agents!
This has been the most significant announcement made by VS Code since GitHub Copilot's initial launch. VS Code is going to build a killer AI native IDE, which is going to open up the market for everyone else and reduce the need for forks to exist Unexpectedly exciting! 🚀
Today, we're announcing plans to make VS Code an open source AI editor. We believe AI development should stay true to VS Code's core principles: open, collaborative, and community-driven. Let's build the future of software development together. aka.ms/open-source-ai…
With @GitStart on top of full search you get: - Deep Search (rolling window across all files instead of plain RAG) - Tools to loop in engineers on our platform - Visual QA (human / sub-agents) - Stages (Scope, Spec, Code, Review QA) They have a long way to go!
No companies has any particular moat in this space. IDE switching cost is minimal if it’s all VS Code (extensions come over) All work on GitHub so you can switch different agent for every PR. Only moat is context and tools that others can’t add easily!!
today we are introducing codex. it is a software engineering agent that runs in the cloud and does tasks for you, like writing a new feature of fixing a bug. you can run many tasks in parallel.
We shipped the highest # of one touch PRs (merged directly or with a nit) AND we had the highest number of AI PRs merged where humans only reviewed and verified the code 🔥 Looking forward to what @GitStart ships next week 🚀
So much fun building stuff I actually love to use
That’s why we evaluate ourselves by duplicating recent PRs shipped by our customers instead of benchmarks at @GitStart Current benchmarks are practically useless in modelling real software engineering!!
Wrote about this before: data scientists at a few large tech/fintech companies I talked with all shared how they have internal evaluations on how AI models can solve *real problems* they give to new joiners. They keep evaluating this to stay grounded. Last I heard it was low