The Download: Apple’s headset challenges, and what AI can be taught from nuclear security

The ‘yet another factor’ introduced by Apple at its Worldwide Developers Conference (WWDC) this 12 months was the business’s worst-kept secret. The Apple Vision Pro, the tech big’s gamble on making blended actuality headsets a factor, has obtained a blended reception. Most of the priority has centered on the eye-watering $3,499 value.

But there’s an even bigger drawback: Whether there’ll be sufficient apps obtainable to make the price of the gadget value it. It’s an actual problem to revamp apps for a wholly new interface—and builders are involved. Read the full story.

—Chris Stokel-Walker

To keep away from AI doom, be taught from nuclear security

For the previous few weeks, the AI discourse has been dominated by those that suppose we may develop an artificial-intelligence system that can at some point develop into so highly effective it is going to wipe out humanity.

So how do firms themselves suggest we keep away from AI damage? One proposed answer comes from a brand new paper by DeepMind et al that implies that AI builders ought to consider a mannequin’s potential to trigger “excessive” dangers earlier than even beginning any coaching.

The course of may assist builders resolve whether or not it’s too dangerous to proceed. But probably it’d be extra useful for the AI sector to attract classes from a area that is aware of a factor or two about very actual existential threats—security analysis and threat mitigation round nuclear weapons.

—Melissa Heikkilä

Melissa’s story is from The Algorithm, her weekly publication providing you with the within observe on all issues AI. Sign up to obtain it in your inbox each Monday.