Pierre retweeted
Agents generate code faster than any repo host was built for. Code Storage is Git infrastructure made for machine-scale writes.
by @fat @pierrecomputer
I've been redrawing the icons we use at @pierrecomputer. Along the way, I've been whipping up plugins to help me manage them—like this Layer Swap tool for "composite icons." Allows me to swap every instance of each layer with ease. Much faster, much easier.
Pierre retweeted
if you're at Git Merge, go find the @pierrecomputer folks and talk all things code storage!
Powered by Code[dot]Storage
Introducing @indent, one coworker for your whole company
Indent is a single agent that is built to work with every member of your team, simultaneously. Instead of every one setting up their own agents, use Indent to build one together.
Try it at indent.com
Pierre retweeted
We are huge fans of @pierrecomputer and code.storage.
They've literally had 0 downtime. It's unbelievable
Pierre@pierrecomputerAug 25United States
United StatesConnected via WebAccount-level information, not a live location or per-post device.
Me: "should we tweet this?"
Alex: "say something subtle about uptime being on people’s minds"
Pierre retweeted
We have created over 66.8 million git repositories.
We started in 2023 creating all repos on github in the hot path.
By 2024 we had millions of github repos in one github org which broke down. And resulted in a lot of attention around 'lovable bringing down github' - see lovable.dev/blog/incident-gi…
During the day of the incident we had a war room, and we moved to our own git-remote-s3 setup. It served us well.
We have since worked in tight collaboration with Pierre to optimise the system resulting in a 230% improvement in p50 fetch latency over our s3 setup.
code.storage/changelog/lovab…
xitter.cf/pierrecomputer/status/…
Pierre@pierrecomputerAug 24United States
United StatesConnected via WebAccount-level information, not a live location or per-post device.
At peak, Lovable creates 250 new repos per second on code.storage.
Simultaneously they improved fetch latency 230% from a more purely object storage solution.
If you’re a product builder and it’s scale and speed that you’re after, signups are now open.
code.storage/changelog/lovab…
Pierre retweeted

Theo - t3.gg@theoAug 25
United States
United StatesConnected via United States App StoreAccount-level information, not a live location or per-post device.
I think we’ll quickly hit some kind of balance here where software goes into a few buckets:
1. AI-powered human interfaces (Codex, T3 Code, Claude Cowork)
2. programmatic interfaces (anything with a good MCP, CLI, or agent ready APIs)
3. legacy human interfaces (apps like Excel, Gmail, Photoshop)
Investor world seems to think everything will become category 2. Anyone who’s tried to use one of these awful bridges knows that will never happen. I can’t even automate my GitHub usage without hitting endless rate limits 🙃
Category 1 is getting saturated fast with chat interfaces. Category 3 is fighting to pretend they still matter by baking awful agents into their products.
I think Category 2 is the best place to fight right now, but you have to do it without pretending you can just staple a slow, rate limited API onto some existing software.
I think @pierrecomputer is the best example right now. It seems like fixing GitHub to handle the unique needs and load from agents isn’t viable, and they built something tailor made for the needs of agents.
Powered by Code[dot]Storage <3
Pierre retweeted
Code Storage is an incredible piece of technology
They have seamlessly scaled to tens of millions of Git repos backing Bolt projects without a hiccup
Jacob and their team have been excellent partners for us at Bolt - can't recommend higher
Signups are now open on Code[dot]Storage.
Today, more new repos are created per second on Code Storage than any other platform.
Lovable, Bolt, Amp, Poke, and many more of the largest agentic workloads store their code and agent memory at scale using Code.Storage.
Pierre retweeted
fun fact: @warpdotdev factories start on code.storage
📣 Massive day for @pierrecomputer — Code Storage is now generally available!
Expressly built for machines and the massive agentic volume we’re seeing. More repos/second than any other code host and already home to some of the best platforms.
↪ code.storage
Signups are now open on Code[dot]Storage.
Today, more new repos are created per second on Code Storage than any other platform.
Lovable, Bolt, Amp, Poke, and many more of the largest agentic workloads store their code and agent memory at scale using Code.Storage.
Edit is now live on diffs[dot]com.
This isn't just another Monaco or ProseMirror wrapper.
This is an entirely new edit implementation, built on top of our file and diff rendering primitives.
- Over 189 commits
- Over 900 new tests
Try it out now – diffs.com/edit
Pierre retweeted

Mitchell Hashimoto@mitchellhJul 29
United States
United StatesConnected via United States App StoreAccount-level information, not a live location or per-post device.
I've started a new company: @superlogical! We're going to begin by building a terminal multiplexer. The entire vision is much larger, but the multiplexer is the foundation. Sign up for the newsletter to get beta access and devlogs (product updates only I promise).
superlogical.com/
personal update: i've joined the primitives team for a while to make diffs.
Pierre@pierrecomputerJul 28United States
United StatesConnected via WebAccount-level information, not a live location or per-post device.
Our primitives team is working very hard on a new and exciting update for diffs[dot]com
We're excited to share it with you very soon, we just have to make a few final edits
Pierre retweeted
your website should inspire joy
Pierre@pierrecomputerJul 24United States
United StatesConnected via WebAccount-level information, not a live location or per-post device.
If you haven't gone and played around on the code[dot]storage website you definitely should
For the new editable diffs[dot]com functionality, we pointed agents at existing MIT-licensed editor test suites (Codemirror, Monaco, Atom, etc.) and had them analyze the nearly ~6k tests for missing behavioral coverage.
After about 14 hours, we added 195 net new test cases, 30 of which exposed real oversights across selection management, whitespace collapsing, search, history, and more (we now total ~1,281 tests).


























