Microsoft makes TypeScript faster; Crusoe adds AI services

Welcome to Runtime! Today on Product Saturday: A popular programming language just got a big boost from Microsoft, Crusoe adds new services to its AI cloud, and the quote of the week.

(Was this email forwarded to you? Sign up here to get Runtime each week.)

Ship it

All systems Go: The TypeScript programming language has become quite popular with software developers over the last decade who want to build JavaScript applications but like how well TypeScript works with a lot of popular code editors. But developers working on very large code bases tend to spend a lot of time waiting for the codebase to load, and Microsoft announced plans on Tuesday to port the TypeScript compiler to Go.

"New experiences powered by AI benefit from large windows of semantic information that need to be available with tighter latency constraints," Microsoft said in a blog post. The move "addresses long-standing performance pain points with a pragmatic solution, uplevels TypeScript’s development experience, and ensures that as apps and codebases grow, the tooling keeps up," according to Kate Holterhoff at Redmonk.

DevSecAgentOps (sigh): ServiceNow released three new pre-built agents for customers this week, adding to the first batch it released last year. The new agents include a security tool that manages incident reporting, a change-management agent that lets companies wargame possible strategy shifts, and a networking agent that promises to find and fix network issues.

The company also made its AI Agent Orchestrator and AI Agent Studio generally available to customers that want to use ServiceNow to manage existing agents and build new ones. According to VentureBeat, "administrators will be able to clearly see how many tasks the agent is closing and determine if it is working according to plan."

Manage up: As Nvidia puts the finishing touches on next week's GTC conference, Crusoe got out ahead of the rush with the announcement of two new managed services for customers of its GPU cloud infrastructure. The new services — Crusoe Managed Inference and Crusoe AutoClusters — are the latest part of the company's push to add software assets to its budding cloud infrastructure service, as Crusoe's Nadav Eiron told Runtime in January.

"Crusoe Managed Inference allows enterprise developers to quickly and easily run and automatically scale the deployment of machine learning models without the need to set up or maintain complex AI infrastructure," the company said in a blog post. AutoClusters was designed to help customers manage the uptime of their AI apps, and works with Kubernetes.

Kid A: Cohere is trying to compete with some deep-pocketed AI model developers by focusing directly on the needs of enterprise customers, rather than college students that need to generate a paper out of thin air by tomorrow. This week it introduced Command-A, a new model that it said outperforms OpenAI's GPT-4o and DeepSeek's V3 "across agentic enterprise tasks, with significantly greater efficiency."

"Command A builds on Cohere’s focus on retrieval-augmented generation (RAG), external tool use and enterprise AI efficiency — especially with regards to compute and the speed at which it serves up answers," VentureBeat reported. Cohere said Command-A can achieve those results while running on just two GPUs, which is a fraction of what other models need to run.

Total Eclipse: Coding is by far the most promising application for large-language models in the enterprise, and coding editors and IDEs are probably never going to be the same. The Eclipse Foundation released two new open-source projects this week that were designed to make it easier to use LLMs in popular development tools.

Theia AI "gives tool builders full control over how AI is integrated into their tool products" and is generally available. The foundation also introduced an alpha release of Theia IDE, which "demonstrates the potential of AI-enhanced development workflows, while maintaining user control and transparency," it said in a press release.

Stat of the week

Container-based workloads are more ephemeral than ever before, according to new research released this week by Sysdig. "For the first time, 60% of containers now live for 60 seconds or less," the company said, which compares to an average lifespan of several days not all that long ago.

Quote of the week

"These services are going to become a basic expectation. The user interface is evolving in a very visceral way for what used to be very structured applications." — Airwallex product chief Shannon Scott, describing at the HumanX conference how he believes generative AI services will impact almost all enterprise apps over the next several years.

The Runtime roundup

Here we go again: The Pentagon is drawing up requirements for a new cloud infrastructure contract, seven years after the JEDI debacle eventually led to the $9 billion JWCC contract awarded in 2022.

Cisco released a patch for a vulnerability that could allow an attacker to crash its IOS XR routers by sending a carefully crafted BGP message, according to Bleeping Computer.

Thanks for reading — see you Tuesday!