The leading observability platform wants $83,700/yr just to monitor your app and catch issues.
So Replit Missions built DataCat: github.com/replit/datacat
100% free.
This isn't a prototype. Agent swarms explored thousands of workflows in the leading observability platform and rebuilt its features 1:1.
Andres retweeted
Your Android app no longer needs to ship its own ML runtime. 📦
LiteRT in Google Play services got a massive upgrade: CompiledModel API & ML Drift GPU acceleration → goo.gle/4zsB2xV
That means:
🔹 ~8 MB smaller APK size
🔹 ~1.4x faster GPU inference
🔹 Auto runtime updates
Andres retweeted
Today we’re launching the Gemini Decisions API, powered by our new Gemini Decision 1 model in @GoogleAIStudio.
Instead of generating text token by token, it’s built for closed-set choices scoring up to 1,000 questions across text, images, audio, video + PDFs in a single pass. A few details:
- 3 decision types: yes/no, multiple choice (up to 255 options) + rubric scoring
- $0.10 / 1M input tokens with $0 billed for output tokens (~5x cheaper than Flash-Lite)
- leads on several benchmark evals with lowest calibration error
Google teams are already building with it, below is a look at Stitch Rapid designing a full product flow in seconds (1x speed)
Andres retweeted
Designing for Android but can't find nice device frames in @figma?
Introducing Pixel Device Frames, a repackaging of every Pixel skin from 1-10 extracted from the public @AndroidStudio source. Once we have 11 there, I'll update it, may also get the watches in there too. Enjoy!
figma.com/community/file/170…
Andres retweeted
Introducing the Android Device Hub, a streamlined app to view and interact with all your local and remote devices and emulators, without needing to open an IDE.
We launched an early version of it last week at the @AndroidDev Summit, and would love your feedback on shaping its future, as we add familiar features from Android Studio, and also explore some new ones.
Try it out with `android device hub` via Android CLI!
Andres retweeted
A good engineer should rarely need to be assigned work.....they should know enough about project/priorities etc to anticipate what needs to be done and put themselves forward for particular tasks. In that way they effectively also gain more independence/autonomy.
Andres retweeted
I had tried out #A2UI in the ClimateTrace CMP sample (github.com/joreilly/ClimateT…) but had only been able to use on Android.
Now, thanks to @xx2bab's brand new a2ui-kmp library (github.com/BinaryTape/a2ui-k…), it's running on iOS, Android and Desktop! 🚀 #KMP #AI
Andres retweeted
Big news out of #AndroidDevSummit! 🚀
We announced the Connected Apps Platform that gives developers a unified entry point to surface their apps directly across Gemini and Search.
Check out the deep dive and live demos: youtu.be/-4G-4w9FKDw?is=cPvw…
Andres retweeted
When we launched Jules, it was an experiment to learn how asynchronous coding agents can help developers ship faster. Seeing what you built taught us so much about where agentic development is headed, directly shaping several products across @Google.
We’re working on bringing our learnings from Jules into @antigravity so you get the best of both worlds: autonomous remote background task execution and an agent-first development environment.
As we focus on this unified experience, Jules will be winding down.
Whether you used Jules to crush bugs, write tests, or launch features— thank you for shaping the future of agentic coding! 💙🦑
Andres retweeted
Android Dev Summit 2026 is happening now! 🤖✨
We built an ADS Connected App powered by MCP so you can explore sessions with an interactive UI right inside Gemini.
Ask Gemini about @AndroidDevSummit to try it live! 🚀
Andres retweeted
Turn agent replies into native app UI with Agent-to-UI (A2UI) for Jetpack Compose, now in alpha!
A2UI lets agents respond with a real interactive UI built from your app's own components instead of a wall of text.
goo.gle/4iMsRX2
#AndroidDevSummit
Andres retweeted
A2UI (developer.android.com/develo…) is such a cool technology. Very interesting live coding talk at #AndroidDevSummit showing how to use it (including adding custom components).
Andres retweeted
The coding interview is dead, at least for my team at Anthropic. I've seen enough to know that it no longer provides signal. I removed it from loops this month.
I have a spicy, optimistic take - I think there's an opportunity for junior engineers: felixrieseberg.com/hunger-be…
Andres retweeted
Native Markdown support is available across @GoogleDrive & @GoogleDocs. 🚀 Preview .md files directly in Drive, open and edit them in Docs, and collaborate in real time—all without converting your files. 👉 goo.gle/3W2dMbe
Andres retweeted
Gradle Skills 1.1.0 is out 🐘
• Best Practices skill now works better on weaker agents
• Wrapper Upgrade skill fixes rollback copy issues
Give your AI agent real Gradle know-how 👇
github.com/gradle/gradle-ski…
Andres retweeted
Andrej Karpathy:
"Prompting is going away. Delete everything, keep Graph."
In just 69 minutes, he shows how to build graphs and why they're the only layer that survives in the end
Prompts → Agents → Loops → Graphs
a loop keeps one agent running until the job is done
a graph decides which agents exist and what each one passes to the next
anyone can build an agent, almost nobody builds the graph that runs them
watch it today, then save the full guide on graphs below before everyone starts catching up ↓
Andres retweeted
We've just released a new official Android skill that can be used to statically analyse your #JetpackCompose code for performance improvements. ⚡️
Install it, and ask your agent to run the skill on your code. 🧑💻
We've seen some good results running it internally - give it a try and let us know any feedback!
github.com/android/skills/tr…
🤖 Made with AI
Andres retweeted
We now have a changelog for the Antigravity IDE extensions on the docs!
antigravity.google/docs/chan…
Andres retweeted
To make your agent produce high-quality output, equip it with a way to check its work. Tests are part of that.
Here are some of the ones I think are worth investing in.
- Write end-to-end tests that simulate real user flows and act as the ground truth
- Use property-based testing to specify what should never happen and generate thousands of test cases that attempt to break those constraints
- If replacing a system, compare the old and new versions using randomly selected inputs
- Require tests to run quickly and deterministically, which reinforces the agent to retry when the loop is slow or unreliable (the agent may learn to retry, not fix the problem)
Example tests say what should happen; properties say what must not.
























