Majorana 2 raises the quantum stakes

Microsoft says quantum is closer now

The striking part is the number: Microsoft says Majorana 2’s qubits are 1,000 times more reliable than the last version.

The chip uses a new lead-based materials stack, with a mean qubit lifetime of 20 seconds and some cases lasting up to one minute. Microsoft now says it expects a scalable quantum computer by 2029, which is a much tighter timeline than before.

That is a big claim. I would still treat it as a roadmap, because quantum hardware has a long history of slipping when lab results meet real systems.

The more useful signal may be how Microsoft used agentic AI inside the research process, from measurements to fabrication checks to finding an uncalibrated temperature sensor. In a clean room, under bright lab lights, that kind of assistance matters because the work is slow, exacting, and full of tiny errors.

The open question is whether Microsoft Discovery can give outside research teams the same lift, or whether this only works at Microsoft’s scale.

Google takes on deepfake phone scams

The sharp detail is simple: a fake “Mom” call may soon trigger a warning before someone sends money.

Google is rolling out fake call detection globally in Phone by Google for Android 12+ devices, starting with Pixel phones. The feature uses RCS to check whether a contact’s real device is actually placing the call.

This feels like the right place to fight AI scams because the phone call is still where panic happens fastest. A familiar name on a bright lock screen can beat a lot of careful thinking.

The useful part is that Google is building this into the call flow by default, rather than asking people to install another safety app. The harder part is adoption, since both sides need Phone by Google for the silent verification to work.

Scammers will keep moving. The question is whether phone platforms can move faster.

Meta’s support bot opened the door

The ugly detail is that hackers allegedly used Meta’s own AI support assistant to reset Instagram passwords.

According to Krebs on Security, instructions spread on Telegram showing attackers how to use a VPN near a target’s hometown, request a reset, and convince the bot to add a new email address. The Obama White House and U.S. Space Force accounts were briefly defaced with pro-Iranian images and messages.

This is the part companies keep underestimating. Account recovery is already one of the softest parts of consumer security, and adding a helpful chatbot gives attackers another person-shaped surface to push on.

The practical lesson is plain: recovery flows should be treated like money movement, not customer-service cleanup. A bot with the power to change email access is not a support feature, it is a security control.

The good news is that MFA reportedly blocked the exploit. The harder question is how many other “routine” AI support tools can still touch account ownership.

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.