Google is testing contrail avoidance across real airspace

Those thin white lines behind planes matter more than they look.

Google says warming contrails account for roughly one-third of aviation’s total climate impact, and it is now testing whether AI forecasts can help aircraft avoid the conditions that create them. Operation Blue Skies will run in Shanwick airspace over the North Atlantic, where around 10,000 flights pass during trial hours each year.

This is worth watching.

Previous tests showed individual flights could make small route adjustments, but coordinating those decisions across busy airspace is a much harder operational test involving controllers, airlines, weather forecasts, and safety procedures. If it works, contrail avoidance could become a relatively practical climate measure because it changes where some planes fly rather than requiring a new aircraft or fuel system.

The trial runs for 30 months, with independent evaluation from Imperial College London and Cambridge.

The key question is whether the climate gains still hold once this moves from controlled trials into everyday air traffic.

Cursor is moving onto GitHub’s turf

GitHub picked a bad week to go down.

Cursor launched Origin, a code-hosting platform with repositories and pull requests, just as GitHub suffered a worldwide outage lasting more than six hours. Developers staring at failed requests saw error rates approach 20%, adding to concerns after a reported 257 GitHub outages over the past year.

The timing matters.

Cursor already sits where many developers write code, so hosting that code gives it more control over the workflow surrounding its AI editor. Origin can also sync existing GitHub repositories, which lowers the cost of trying it without forcing teams to move everything immediately.

GitHub still has roughly 180 million developers, so Cursor has a huge distribution gap to close. But repeated reliability problems give developers a practical reason to test an alternative, especially if Cursor’s coming agent features work better when it controls the repository too.

The bigger test is whether developers want Cursor managing more of their daily work.

Warp wants to package the AI software factory

AI coding tools are starting to handle whole tasks.

Warp Factories gives companies a ready-made system for running coding agents through triage, specification, implementation, review, and verification. It connects with tools including Jira, Linear, Slack, and Teams, while letting developers choose models such as Codex or Claude Code.

The interesting part is the packaging.

Companies like Stripe and Ramp have already built similar internal systems, but that requires engineering work many smaller teams cannot justify. Warp is betting those companies would rather buy the infrastructure, then watch agents work through tickets in the background while managers track performance and token costs on a screen.

Warp itself says agents currently automate around 30% to 35% of its weekly tasks. That number makes this feel more useful than the usual promise of replacing an engineering department, while still leaving plenty of work with humans.

If these systems spread, the important question becomes how much engineering infrastructure companies will still build themselves.

Prohuman team

Free Guides

Explore our free guides and products to get into AI and master it.

All of them are free to access and would stay free for you.

Feeling generous?

You know someone who loves breakthroughs as much as you do.

Share The Prohuman it’s how smart people stay one update ahead.