For an easy question—how many states are there in the United States?—mysterious groups on the internet seem intent on asking it 27,000 times.

Possible attackers are pairing scan-the-internet scripts with innocuous prompts like the aforementioned USA query to find large language models (LLMs) accessible without the need to authenticate, according to recent reports from cybersecurity training group SANS Institute and cybersecurity company GreyNoise.

Researchers said these LLM-ventories could support a range of adversarial tasks, including compute theft and extraction of restricted information.

After investigating their honeypot infrastructure, GreyNoise researchers detected:

• Two IPs launching “a methodical probe of 73+ LLM model endpoints,” access points that use APIs to interact with the system.
• Beginning on December 28, 2025, the scanners generated over 80,000 sessions in 11 days. According to the post, they aimed to find “misconfigured proxy servers” leaking access.

When life gives you exposed LLMs, make absolute havoc.—BH

There’s an assumption in cybersecurity that the youth—Gen Z—are more savvy than their older counterparts. That’s not necessarily true.

In a December panel discussion hosted by software security company Veracode, “How Gen Z Is Reshaping the Cyber Threat Landscape,” participants covered how different generations approach cybersecurity. The conclusions were somewhat surprising, particularly for Gen Z.

Young town. Digitally native Gen Z might be expected to exercise more caution online. But, according to panelist Caitlin Sarian, a cybersecurity influencer who goes by Cybersecurity Girl, that’s not the case. In fact, growing up immersed in tech may make younger people even more susceptible than older generations.

But don’t get it twisted: No one is safe, says one expert.—EH

Much like the Lumière brothers’ train, quantum computing is rushing toward the public—and users aren’t likely anticipating the changes in store.

At this year’s CES, Murray Thom, VP of quantum technology evangelism at D-Wave Quantum, told IT Brew that he feels people overcomplicate their thinking about quantum computing: “It’s energy-efficient compute for very hard problems.”

For D-Wave’s customers, he added, the technology’s applications are already streamlining processes and boosting efficiency, saving manpower hours for Pattison Food Group and helping Japanese phone operator NTT Docomo reduce paging signals.

“Those are very real-world, tangible applications; they’re not research projects, we’re not doing abstract mathematical problems,” Thom said. “And it’s incorporated in their business.”

What a quantum x AI looks like.—EH

Despite heavy investment in AI, only 12% of companies are considered true AI leaders. A new report breaks down what’s holding enterprises back and what separates AI maturity from experimentation.