TypeScript 5.9 Brings Less Friction, More Features
Typescript 5.9 was released Aug. 1, with a slate of new features, including expandable hovers for preview, support for import defer and minimal and updated tsc --init.
Daniel Rosenwasser, the principal project manager for TypeScript, wrote that for the last few years, “running tsc --init created a very ‘full’ tsconfig.json, filled with commented-out settings and their descriptions.” The intent was to make options discoverable and easy to toggle, he explained.
“However, given external feedback (and our own experience), we found it’s common to immediately delete most of the contents of these new tsconfig.json files,” he wrote. “When users want to discover new options, we find they rely on auto-complete from their editor, or navigate to the tsconfig reference on our website (which the generated tsconfig.json links to!).”
It turned out generating the tsconfig.json was often overkill for what developers actually did. They also wanted the tsc --init to be initialized with a few more prescriptive settings than were enabled thus far, he wrote.
“We looked at some common pain points and papercuts users have when they create a new TypeScript project,” he said.
For example, he wrote, most users write in modules rather than global scripts. --moduleDetection can force TypeScript to treat every implementation file as a module.
“Developers also often want to use the latest ECMAScript features directly in their runtime, so --target can typically be set to esnext,” he wrote. “JSX users often find that going back to set --jsx is needless friction, and its options are slightly confusing.”
What’s more, projects were loading more declaration files from node_modules/@types than TypeScript actually needed. Specifying an empty types array can help limit this, he said.
In the full blog post, Rosenwasser showed the new tsconfig.json that a plain tsc --init with no other flags now generates.
Another change in this update: Many of the DOM APIs in TypeScript only linked to MDN documentation for the API, but that didn’t provide a “quick summary” of what the API actually does. Now, thanks to the work of Adam Naji, TypeScript’s MDN documentation includes that information.
This update also incorporates:
- Support for import defer
- Support for –module node20
- Expandable Hovers (Preview)
- Configurable Maximum Hover Length
Rosenwasser also includes a look at what is planned for TypeScript 6.0, which he writes will serve as a “transition point” for developers to adjust their codebases for TypeScript 7.0, which is focused on the native port of TypeScript.
OpenAI Releases Two Open Source Models
OpenAI released two open-weight language models — get-oss-120b and get-oss-20b — which it says will “deliver strong real-world performance at a low cost.”
Available under the Apache 2.0 license, the models outperform similarly sized open models on reasoning tasks, according to the OpenAI blog that provides evaluations for the models.
They also demonstrate strong tool use capabilities and are optimized for efficient deployment on consumer hardware, it added.
gpt-oss-120b matches OpenAI o4-mini on core reasoning benchmarks while running on a single 80 GB GPU, the company stated, adding that gpt-oss-20b delivers similar results to o3-mini and runs on edge devices with just 16 GB of memory — making it ideal for devices, local inference, or rapid iteration.
As part of the release, OpenAI also published a safety research paper and model card that detail its protocol for security in “worst-case scenarios.”
Weights are live on Hugging Face and GitHub.
Name.com API Supports AI-based Domain Registration
AI can design a web page and produce the code, and now…it can buy a domain name for you. Name.com, the ICANN-accredited domain registrar and web hosting company based in Denver, Colorado, now offers a name.com API.
The AI-native domain platform will change how companies build custom domain search, registration and management into their services and applications, the company claimed.
“With support for Model Context Protocol (MCP) and OpenAPI specification, name.com domains are modernized for the age of agentic AI, enabling AI agents to interact directly with the API,” the company added.
Canva Expands AI Footprint
For the DYI developer, Canva launched a deep research connector for ChatGPT on June 26. The company also revealed that Canva GPT — which generates designs — usage surged 375% year over year. It’s one of ChatGPT’s top productivity apps, the company stated.
Canva offers a one-click integration with Salesforce’s Agentforce and plans more AI integrations.
The company also officially launched its MCP server. The Canva MCP Server is an open platform that allows any AI assistant to tap directly into a user’s full Canva workspace to generate designs, draft or improve design copy, and resize assets, among other design tasks.
MCP integrations with Claude, ChatGPT, Salesforce and more will be available soon, the company noted.
Designs generated through MCP are “visually rich and context-aware,” because it has real-time access to both the user’s Canva account and the ongoing AI conversation.
The Canva MCP Server allows AI assistants and agents to:
- Generate any design type, from social posts to presentations, with context from chat;
- Autofill charts with labeled, formatted data from AI-generated insights;
- Resize and export branded templates; and
- Import PDFs or files directly from a link, no upload needed.
