Building React Native Apps, Investor, and Tech Evangelist 🥑 | Ex @penta_medical YC S18 | React.js | Remote at @FullscriptHQ 💻 🇨🇦 | OSS ❤
Toronto, Ontario
Joined March 2010
- Tweets11K
- Following6.1K
- Followers5K
- Likes28.8K
Adrian Carolli retweeted
a software factory will also amplify what you have already in the codebase:
give it slop → receive 10x more slop
give it guardrails → ship high quality code without reading
invest in verification and evals before the factory. go into debt if you have to
Adrian Carolli retweeted
Xcode 27.2 ships with a new skill for agents to work with Instruments, as well as an update to the SwiftUI specialist skill. Love that my apps can get better just by upgrading!
Adrian Carolli retweeted
Introducing Simlock: a free, open-source tool that turns your machine into a simulator cloud that manages itself, so every coding agent gets its own iOS or Android device, simulator or real phone.
github.com/callstackincubato…
You can now use Simlock in T3 Code. Simlock is a gateway that lets you connect multiple machines to a unified API that your agents can request leases from when they need to test mobile stuff.
Works for both Android and iOS and is a great alternative to EAS Simulators if you happen to have some spare compute laying around!
Adrian Carolli retweeted
Last night an AI agent performed brain surgery on a mouse.
We gave it 19 gene therapy vectors in a spreadsheet, a mouse model of epilepsy, access to a motorize surgery rig with cameras and one goal: reduce seizures.
It picked the drug. It picked the brain region. It derived the coordinates from three papers. Then it drove the stereotaxic arms into the brain and injected 250 nanolitres while watching the entry site through the rig camera for bleeding.
A licensed vet gated every irreversible step and held a physical kill switch to stop the AI agent.
Thread on what it decided, what it saw, and the moment it refused to let me override it.
[Writing this on the spot out of excitement, but I've been 18hrs in front of my laptop coding and coordinating this, so keeping it brief and descriptive, more cohesive thoughts on this soon...]
Adrian Carolli retweeted
At Callstack, most of our time goes to building applications for many platforms. We're building Sentinel with that use case in mind. Powered by our tooling and model.
Excited to announce Project Sentinel: a code review and verification agent, built for mobile apps, powered by @apex__ai model.
Early benchmarks look promising:
â–Şprecision: 82.5%
â–Şrecall: 40%
â–Şseverity alignment: 80%
We're rolling it out internally today. GA 🔜
Adrian Carolli retweeted
Yes! useReducer (and state machines, when needed) have ALWAYS been the better default.
Devs just avoided them because upfront modeling felt like too much work.
Agents don't care. They're not lazy, and explicit logic keeps them from doing dumb things.
Adrian Carolli retweeted
I’d love to buy @spolsky a coffee and ask him whether AI has changed his thinking on rewriting software.
Does he still stand by his 2000 article? And how would he think about moving from React Native to Native today?
For more than a decade, I was firmly against rewriting software from scratch. AI has changed my view of the economics enough to make me reconsider. I still think it’s a high-stakes decision, but some of my old assumptions deserve another look.
It reminds me of Roger Bannister breaking the four-minute mile: seeing someone do what seemed out of reach can change what others believe is possible.
But that’s also where I get cautious.
When we read about companies moving from React Native to Native, we’re seeing a selected set of stories i.e. Shopify, Airbnb, Coinbase, etc. A successful transition makes a compelling engineering blog post. An abandoned rewrite is probably less likely to get one.
We don’t know how many teams tried, failed, and quietly moved on.
AI may be changing the cost of a rewrite. It doesn’t remove the need to understand why you’re doing it, what you could lose, and whether your team can carry it through.
I’m more open to rewrites than I used to be. I’m still wary of treating the stories we hear as evidence of how often they work.
Has AI changed your threshold for considering a rewrite?
joelonsoftware.com/2000/04/0…
Adrian Carolli retweeted
Debating just straight-up banning useState in my React apps
Push the agent into the black hole of success
Adrian Carolli retweeted
🚨 app store banners can autoplay videos now
new in iOS 27, and almost nobody’s using it yet
super easy to build with claude motion design 👇🏻
preview videos already lift installs up to 30%
add it to your release checklist
Adrian Carolli retweeted
Starting to see more and more of this agentic workflow across our customer base 👇
The mobile factory is humming @expo - everyone is reaching for this type of setup:
1. agents pick up linear tickets
2. spawn simulators to fix them
3. submit PRs
4. 10-20 going at a time
5. Run your test suite
6. Deploy to Production via Update or Build
Adrian Carolli retweeted
The mobile factory is humming @expo - everyone is reaching for this type of setup:
1. agents pick up linear tickets
2. spawn simulators to fix them
3. submit PRs
4. 10-20 going at a time
5. Run your test suite
6. Deploy to Production via Update or Build
Adrian Carolli retweeted
Starting to hear of more teams skipping PRs and doing true trunk-based dev.
The argument: PRs make less sense in the age of agentic coding.
The agent is the author, not me. So I, the prompter, am the first reviewer. One or more AIs operate as second reviewer.
Eliminating PRs unlocks radically more throughput - At Amp they do over 100 releases a day.
21 people, a single meeting a week, no quarterly goals, no pull requests, no mandatory code review, no titles (except one dictator), a hundred releases a day, bugs fixed in minutes, famous for killing features.
@sqs and @thorstenball on how the Amp team works:
Adrian Carolli retweeted
There’ll be times where you over-guide the AI. It knows better than you and is more intelligent than you.
There’ll be times where you under-guide the AI. It can go off-track, have the wrong biases, or miss context.
How well you manage this superpower is the #0 skill I look for.
Adrian Carolli retweeted
Same, wrong level of abstraction for this next phase. They had a good run.
