[Dev Catch Up #57] - GPT 4o Image Generation, Gemini 2.5 Pro, Bun v1.2.7, Next.js 15.2.3, Biome, ttyd, Cascii and More

Welcome to the 57th 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:

• How to build you own MCP Server?

• LLM Glossary

• How to do Local AI Code Generation and Completion using Open-source Llama.cpp?

• What are AI Agents and Must-Try Agents for Every Developer

Join 7000+ developers to hear stories from Open source and technology.

Must Read

• We all know people are going crazy over Ghibli style images lately. It is because of newly added image generation capability in GPT-4o which turns prompts into stunning visuals with impressive precision. Read OpenAI’s release post to know more details.

• We have been covering Google’s Model updates every week. And this time, its Gemini 2.5 Pro Experimental. It is a Reasoning Model with long context window and now available on Google AI studio. Explore more about Gemini 2.5 Pro on Google’s blog.

• For Bun developers, it’s been a busy week with two back-to-back releases v1.2.6 and v1.2.7. These updates bring performance improvements, better TypeScript support, and a new way to handle cookies more easily. Check out the changelogs for v1.2.6 and v1.2.7 for full details.

• For Next.js users, version 15.2.3 is out to patch a critical vulnerability impacting self-hosted apps using next start and output: 'standalone'. Backported fixes are available for older versions. This is a recommended update. Read the full advisory for more details.

OSS Highlight of the Week

Biome is a performant toolchain for modern web projects. It works as both a formatter and linter, supporting JavaScript, TypeScript, CSS, GraphQL, and JSON, all without extra setup or Node.js. It can be used via CLI or integrated into your IDE with LSP support, helping you catch issues as you code. It is still evolving and showing steady progress. Check out the GitHub repo for details.

---

---

Good to know

• Being a developer, I often work in the terminal and check other’s terminals too. ttyd makes that easy with a simple command line tool that shares your terminal over the web in real time. Check the GitHub repo to explore more.

• Text-based diagrams are great for docs, READMEs, or quick planning. Cascii is a simple web-based tool to create ASCII and Unicode diagrams with zero setup. Check the GitHub repo or use cascii.app to build and share instantly.

• Matching dynamic URLs is now easier in Node.js v23.8.0 with the new URLPattern API, contributed by Cloudflare. It brings cleaner routing and better consistency across runtimes. Read more in the Cloudflare blog.

• We’ve been highlighting many tools for data validation because it's a common need in apps and APIs. Valibot is a lightweight, dependency-free schema library that can run in any JavaScript environment and makes data validation easy and type-safe. Explore more in the GitHub repo.

• LLMs are getting better at reasoning, but we still don’t fully understand how they think. Anthropic’s latest research explores Claude’s internal behavior and shows it can plan ahead, reason across steps, and even make things up. Read Tracing the thoughts of a language model blog post to know more interesting details.

Notable FYIs

• Every frontend developer knows how difficult it’s been to style dropdown menus on the web. Browsers controlled most of the look, leaving only limited options. Read this Chrome Developers blog to know how Chrome version 135 changes that with full CSS customization.

• Generating high-quality images with diffusion models usually takes a lot of inference time. Researchers at UC Berkeley introduced “shortcut models” that reduce inference time and enable faster image generation. Check this DeepLearning AI article to learn more.

• Working on AWS resources regularly, I’ve seen how important it is to choose the right policy for access control. Fog Security blog explains how Service Control Policies (SCPs) and Resource Control Policies (RCPs) can work together to strengthen security. Read more in the Fog Security post.

• We all know how Chain-of-Thought (CoT) reasoning, improved performance in text generation models. This paper brings the idea of ImageGen-CoT, which makes models generate reasoning steps before image synthesis. Read the research paper to learn deeper.

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.