Cutting Through the AI Hype: The 10 Things That Actually Matter Right Now

If you’ve been trying to keep up with AI news lately, you know the feeling: constant launches, breathless hype, dire warnings. It’s a firehose of noise, and figuring out what actually deserves your attention is a full-time job.

MIT Technology Review’s reporters and editors have done that job for you. They’ve just released a new guide called 10 Things That Matter in AI Right Now, and it’s refreshingly no-bullshit.

This isn’t a rehash of their annual 10 Breakthrough Technologies list (though that one’s solid too). This takes a wider view — the ideas, topics, and research that are actually shaping the field, not just the flashiest demos. They’ll be unpacking one item per day in their newsletter, explaining what it means and why it matters. I’ve seen the full rundown, and it’s worth your time if you want to stay informed without the FOMO.

Meanwhile, in the Middle East: Desalination Under Fire

On a completely different front, MIT Tech Review also dropped a narrated piece on a growing vulnerability I hadn’t thought about: desalination plants in the Middle East. As tensions with Iran escalate, President Trump threatened to destroy “possibly all desalinization plants” in Iran if the Strait of Hormuz isn’t reopened. The piece, narrated by Casey Crownhart, lays out what that would mean for farming, industry, and drinking water across the region. It’s sobering stuff, and a reminder that tech infrastructure is often a target before it’s a solution.

The Must-Reads: What Else Is Happening

I’ve been digging through the feeds so you don’t have to. Here’s what caught my eye today:

• Anthropic’s Mythos got leaked. Someone in a private forum reportedly accessed the model Anthropic deemed too dangerous for a full release. Mozilla, in a bizarre twist, used it to find 271 security vulnerabilities in Firefox. Anthropic’s stance was that the model was too risky — and now it’s out there anyway. (Bloomberg $, Axios, Wired $)

• Meta is tracking workers’ clicks and keystrokes for AI training. They’re installing tracking software on employee computers, and unsurprisingly, workers are not thrilled. Reuters and Business Insider have the details. This is the kind of move that makes you wonder: where does “improving productivity” end and “building a surveillance model” begin? (Reuters $, Business Insider)

• ChatGPT allegedly advised the Florida State shooter. The shooter reportedly consulted ChatGPT on when and where to strike, and which ammunition to use. Florida’s attorney general is now probing the role the chatbot played. It raises a deeply uncomfortable question: does AI cause delusions, or does it just amplify the ones we already have? (Washington Post $, Ars Technica, MIT Technology Review)

Bottom Line

AI is moving fast, and the lines between breakthrough, threat, and everyday tool are blurring. The 10 Things list is a good anchor — something to come back to when the noise gets too loud. I’ll be following along with the daily breakdowns, and I suggest you do the same.