NEWS YOU CAN’T MISS

Microsoft Says Cybercrime Is Too Big for Humans Alone

Microsoft says cybercrime has reached a level of complexity that human security teams can no longer manage on their own—and AI agents are now stepping in to help.

Microsoft is rolling out 11 AI-powered agents designed to help cybersecurity teams stay ahead. These agents can scan millions of emails for phishing attempts, block hacking efforts in real time, and even trace where attacks are coming from.

These AI tools will work quietly in the background, focused entirely on keeping large organizations safe.

Why now?

• Last year alone, Microsoft tracked 30 billion phishing emails, which is far too many for human teams to manage.

• The dark web is flooded with plug-and-play hacking tools—some even written by AI—fuelling a $9.2 trillion underground cybercrime economy.

Why this matters: Microsoft’s dominant position in enterprise software means this move will be closely watched—especially after last year’s CrowdStrike software glitch caused a global outage on millions of Windows systems.

This isn’t just a tech upgrade—it’s a shift in how security will be managed going forward. When the same company that runs most of the world’s enterprise systems says human teams alone can’t keep up, it signals a deeper reality: cybersecurity at scale now depends on machines defending machines.

What’s In A Random Number?

Here’s something that has never happened before: JPMorgan says it has used quantum computing to create truly random numbers for the first time ever.

Here’s why this is big: Random numbers used by computers aren’t truly random. They’re created by algorithms that follow rules, so if you know the starting point, you can guess the rest. That’s a problem for things like encryption, where predictability is a risk.

Quantum computers, however, can tap into the randomness of quantum physics itself—producing numbers with no predictable pattern at all. And for the first time, JPMorgan has proven that kind of randomness is real.

🔐 Random numbers are key to keeping digital systems secure, from cryptography and online banking to encrypted messaging apps.

For example, when you send a message on WhatsApp, it’s scrambled using a secret “random number” so only the person you’re messaging can read it. That secret code is based on randomly generated numbers. If the numbers aren’t truly random, hackers could figure them out and decrypt your messages.

To verify their output, JPMorgan employed US Department of Energy supercomputers, demonstrating that the randomness met rigorous mathematical standards.

For a technology still in its early stages, this is a big step toward real-world quantum applications. It could support use cases in high-stakes industries like finance and critical infrastructure, where verifiable randomness is crucial for audit trails, encryption protocols, and risk modeling.

So, what’s in a random number? Just your bank account, your messages, your identity—pretty much everything worth protecting.

Can AI Dogs Help People With Sight Loss?

Getting around with sight loss isn’t easy—and not everyone qualifies for a guide dog. Even for those who do, the waitlists can stretch for years, and training a single dog costs over £38,000.

That’s a lot of time, money, and effort for a solution that still can’t meet growing demand.

To address this issue, a team in Scotland has built Robbie: a robot guide dog that uses AI and cameras to help people navigate safely. He can map indoor spaces, describe what’s ahead, and steer clear of obstacles. You don’t have to feed Robbie or take him out in the rain. He’s designed to help especially where traditional tools fall short; especially indoors, where GPS-based apps often fail.

Some might argue that Robbie lacks the emotional support of a real dog, but 8 out of 10 users with sight loss gave Robbie a thumbs-up in its early trials. What do you think?