[Dev Catch Up # 89] - Google's Gemini 3 Pro, Antigravity-Agentic IDE, OpenAI's GPT-5.1-Codex-Max, Meta's SAM 3D, Grok 4.1 Fast, reader3, Cloudflare acquires Replicate, git-worktree-runner & much more!
Bringing devs up to speed on the latest dev news from the trends including, a bunch of exciting developments and articles
Welcome to the 89th edition of DevShorts, Dev Catch Up!
For those who joined recently or are reading Dev Catch Up for the first time, I write about developer stories and open source, partly based on my work and experience interacting with people all over the globe.
Thanks for reading Dev Shorts! Subscribe for free to receive new posts and support my work.
Some recent issues from Dev Catch up:
Join 8400+ developers to hear stories from Open source and technology.
Must Read
Google has released Gemini 3 Pro, Gemini 3 Deep Think, and Antigravity. Gemini 3 was trained entirely on Google TPUs when others relied on NVidia. Gemini 3 Deep Think scores high on many reasoning tests. Antigravity is their new agentic IDE that brings agent-based coding with browser control features. Read Google’s blog about Gemini 3 and Antigravity Agentic IDE.
OpenAI has released GPT-5.1-Codex-Max. It is their new agentic coding model trained on software engineering, math, research, and more. It is built for long running tasks. It is faster, more intelligent, and more token efficient. It is now available in Codex. Check OpenAI’s post on GPT-5.1-Codex-Max for more details.
Cloudflare went down for six hours and took a big part of the internet with it. ChatGPT, Claude, Canva, Uber, Zoom and many others were offline. Cloudflare later released a detailed postmortem saying a small database permission change triggered the issue across their network. It is worth reading if you want to know more on recent Cloudflare’s outage.
Last week we covered Marble, the 3D image and video generation model. Here is another one to add to the list. Meta has released SAM 3D. It can turn a 2D image into a full 3D model of an object or a person. Check Meta’s SAM 3D for more details.
OSS Highlight of the Week
This week we are featuring agentset. It is an open-source RAG platform. It is used to build, evaluate, and ship production-ready RAG applications. It has built-in citations, deep research, 22+ file formats, partitions, MCP server, and more. It provides end-to-end tooling for ingestion, vector indexing, evaluations and more. Check the agentset Github repo to know more details.
Good to know
Cloudflare acquires Replicate to build AI cloud for developers. Replicate says all existing models and APIs will keep working as they do now. Check Cloudflare’s post and Replicate’s post for more details on the acquisition.
I have built CI CD pipelines, but not the runner. The pipeline is the flow of steps, and the runner is the engine that executes them. Check this post to learn how to build a full CI CD pipeline runner from scratch in Python. It covers both single and parallel execution.
The rise of agents also brings the complexity of managing them all in one place. Microsoft solves this with Agent 365. It is a control plane for AI agents built with different frameworks. Agent 365 helps you deploy, organize, and govern agents securely. It includes registry, access control, visualization and security features. Check Microsoft’s blog on Agent 365 for more details.
I came across SuperClaude Framework. It transforms Claude Code into a structured development platform. It includes about 30 slash commands, 16 specialized agents and MCP servers that cover the development lifecycle. Check the GitHub repo to set up your Claude Code workflow.
I found a simple visual guide on how to design good APIs. The diagram explains safe retries, API versioning, and clear resource naming. It is a quick reference if you want to build predictable and clean APIs.
Notable FYIs
xAI has released Grok 4.1 Fast & Agent Tools API. Grok 4.1 Fast is their best tool calling model. It supports 2M token context window. The Agent Tools API gives access to real time X data, remote code execution and more. Check xAI’s post on Grok 4.1 Fast.
I found a helpful post from Reddit about how they chose a vector database for similarity search. The team compared options like Milvus, Qdrant and Weaviate based on real engineering needs. Check the Reddit post on choosing a vector DB for ANN search to know more.
Coderabbit has open-sourced git worktree runner. It is a Git worktree manager that removes the pain of using worktrees. Git worktree is powerful but the commands are long and easy to mess up. This tool makes it simple with short commands and automatic setup. Check the git-worktree-runner repo for details.
Allen AI has open sourced the Olmo 3 models. The release includes 7B and 32B parameter models with full training code, datasets, and checkpoints. It comes in Base, Instruct, and Think variants and supports long context lengths. Check the Allen AI’s Olmo3 blog post for more details.
I came across reader3 by Andrej Karpathy. It is a simple self-hosted EPUB reader that gives you clean chapters ready for LLM reading. You can copy a chapter without any clutter and use it for summaries or notes. Check the GitHub repo to run it on your local machine.
That’s it from us with this edition. We hope you are going away with a ton of new information. Lastly, share this newsletter with your colleagues and pals if you find it valuable. A subscription to the newsletter will be awesome if you are reading it for the first time.
The git-worktree-runner project is a great exampl of how good tooling can simplify complex Git workflows. Worktrees are powerfull but the commands do get messy fast. Would love to see how this integrates with existing CI/CD piplines.