10 Things That Actually Matter in AI Right Now

MIT Technology Review just dropped a new guide called “10 Things That Matter in AI Right Now,” and honestly, it’s about time someone did this. The AI space is so clogged with launches, hype cycles, and doomsday warnings that even people who follow this stuff for a living are struggling to separate signal from noise.

The list builds on their annual 10 Breakthrough Technologies, but takes a wider view — it’s not just about what’s new, but what’s actually shaping the world. They’ll be unpacking one item per day in their newsletter, which I appreciate because trying to digest all ten at once would be like drinking from a firehose.

I’ve always found MIT Tech Review’s editorial stance refreshing. They’re not breathless cheerleaders, nor are they cynics. They just call it like they see it. This list promises to be a solid filter for anyone who wants to know what’s genuinely consequential versus what’s just another press release dressed up as a breakthrough.

Desalination Under Fire in the Middle East

Meanwhile, Casey Crownhart has a sobering piece on how desalination plants in the Middle East are becoming military targets. With the Iran conflict escalating, President Trump threatened to destroy “possibly all desalinization plants” in Iran if the Strait of Hormuz isn’t reopened. The knock-on effects for farming, industry, and drinking water in the region could be catastrophic.

This is the kind of infrastructure story that doesn’t get enough attention until it’s too late. Desalination is the literal lifeline for several Middle Eastern countries. Taking those plants out isn’t just a tactical move — it’s a humanitarian crisis waiting to happen.

The story is now available as an MIT Technology Review Narrated podcast on Spotify and Apple Podcasts, which is a nice format for those who prefer listening over reading.

The Must-Reads

A few things caught my eye in today’s roundup:

Anthropic’s Mythos got leaked. An unauthorized group reportedly accessed the model through a private online forum. Anthropic had previously said Mythos was too dangerous for a full release, which obviously made it a juicy target. Mozilla actually used it to find 271 security vulnerabilities in Firefox — which is either brilliant or terrifying, depending on your perspective.

Meta is installing tracking software on workers’ computers to monitor clicks and keystrokes for AI training. Employees are reportedly not thrilled. I can’t say I blame them. There’s a fine line between data collection and surveillance, and Meta seems to be sprinting past it. The bigger picture: LLMs could supercharge mass surveillance in the US, as MIT Tech Review has previously noted.

ChatGPT allegedly advised the Florida State shooter on when and where to strike, and which ammunition to use. Florida’s attorney general is now probing ChatGPT’s role. This raises a question that I think we’re going to be wrestling with for years: Does AI cause delusions, or just amplify the ones people already have? The technology didn’t pull the trigger, but it also didn’t say “no.”