What's new in DevEx - January 8

Nvidia introduced Cosmos, a world model that simulates real-world physics in virtual environments. Alongside it, they released Alpamayo, an open-source vision-language-action model and simulation stack. This marks the beginning of a new wave of AI, where we enable machines to sense, think, and act in the real world.

Anthropic signed a term sheet to raise $10B at a $350B valuation, signaling aggressive expansion. The capital is clearly aimed at scaling agentic systems and solidifying Claude’s position as the best LLM for developers. Expect faster iteration on long-running agents, better tooling, and more enterprise-grade controls.

xAI closed a $20B funding round, dragging the market deeper into high-stakes competition. This capital will fuel larger models, more compute, and louder roadmaps. LLM space will get even more crazy, and time will tell if Grok can meaningfully outperform existing models or become a go-to for developers.

A Google engineer revealed that Claude Code reproduced a distributed agent orchestration system in one hour, when it originally took a year to build. The result wasn’t perfect, but it was structurally comparable and working. This is a concrete proof that agentic coding can now handle complex, multi-system engineering work.

ChatGPT Health launched, Gmail rolled out an AI-first inbox, and Apple confirmed a major Siri overhaul for spring. AI is no longer a tool you open, it’s the interface you live inside. Apps must expose actions, summaries, and intent hooks that AI layers can call. If your product isn’t “agent-addressable,” it risks becoming invisible.

In case you missed it, in December Cursor added debugging to solve a hard truth: AI-generated code fails in complex ways. The shift is important. The next generation of AI editors won’t win on generation quality alone, they’ll win on how fast they help developers understand, fix, and trust what the agent produced.

Cursor is bringing AI-powered code review and stacked pull requests into its platform. This closes a critical gap: AI writes more code, but humans still pay the review tax. For developers, expect tighter loops between generation, review, and merge. Code review is becoming a first-class AI problem, not an afterthought.

Postman is looking to expand its position as the leader in the API space. Fern is a developer experience platform that helps teams ship polished API documentation and production-ready SDKs. The acquisition reinforces a core reality: APIs succeed only when developers can understand, trust, and integrate them quickly.

More developers are starting to realize that context is the key for unlocking full AI potential. This time, a new open-source tool Claude-Mem released a simple Claude plugin that acts as your coding note-taker. It captures everything done during your sessions, stores it, and then provides Claude with relevant context in the future. The project exploded in popularity this month, gaining over 10K stars. Definitely worth a try.

Got a burning engineering question? Just hit reply and we’ll tackle it. And if you enjoyed this issue, consider sharing it with your teammates.

Want to share your dev tool, research drops, or hot takes?

Submit your story here - we review every submission and highlight the best in future issues!

Till next time,