TechSpective

Podcast: Play in new window | Download Subscribe: Apple Podcasts | RSS For most of the last year, the AI security conversation had a clear villain. Attackers were using AI to write better phishing lures, adapt malware mid-attack, and move faster than defenders could keep up with. That story was easy to tell because it fit the shape we already understood. Bad guys get a new tool, they use it against us, we build a defense.

National Geographic’s “Lion” premiered August 19 and moved to Disney+ and Hulu the next day, closing out a four-year production that followed a cub named Kio from an eight-week-old newcomer to a shot at leading his own pride in Kenya’s Maasai Mara. Jon Favreau, who directed the 2019 photorealistic Lion King, executive produced the series with BBC Studios veteran Mike Gunton. Hans Zimmer, who scored the original 1994 animated film, wrote new music for it.

Meta recently agreed to a seventeen-billion-dollar settlement resolving ongoing youth addiction lawsuits. The deal forces Facebook to implement strict daily time limits for teenage users. Executives are paying billions to actively lock kids out of their own platforms. This won’t end well for Facebook or the kids they are trying to protect. Meta Attempts to Drag Rivals Down Meta faced a massive trial regarding teen addiction and platform surveillance.

Podcast: Play in new window | Download Subscribe: Apple Podcasts | RSS Think about every account you’ve ever created and then forgotten about. A login for a gym you canceled two years ago. A free trial you signed up for and never used again. A vendor portal from a job you left. Nobody deletes those things, and nobody circles back to check whether they still work.

Malware hits your network — and by that point, attackers have already done the hard work. Weeks of it, sometimes months. Cybersecurity conversations tend to fixate on blocking malicious code, but that framing misses something critical. The bulk of successful attacks lean on reconnaissance, social engineering, and stolen credentials long before a single line of malware gets deployed.

I track these industry shifts from my home office here in Bend, Oregon. I serve as the Trustee for a shopping center-owning trust (or did). Recently, I finalized the official closing date for the shopping center we held in Tustin, California. Selling that physical asset was a highly calculated move. You look at market fundamentals to decide when to hold an asset and when to liquidate it entirely. My grandfather shared a specific historical anecdote with me about market momentum.

Every large company has some version of a vendor risk program by now. A vendor list. Risk tiers. Questionnaires and certifications. Insurance requirements. A workflow for approving exceptions when something falls short. Building and running all of that costs real time and real money. Unfortunately, none of it tells you what your actual exposure is. That’s the argument in “The Missing Measure in Third-Party Information Risk,” a paper from HITRUST founder and executive chairman Daniel Nutkis.

The Google Pixel launch event in New York was amazing! Trevor Noah opened the stage with his standard comedic delivery. A musical set from PinkPantheress followed him. The entertainment quickly gave way to a stark corporate message when Rick Osterloh grabbed the microphone. Osterloh runs the devices division at Google. He told the crowd that modern technology must get more helpful and simpler simultaneously. He mapped out a future driven entirely by Gemini Intelligence.

Podcast: Play in new window | Download Subscribe: Apple Podcasts | RSS Barak Sternberg spent a chunk of this year proving something that should worry anyone who has plugged an AI coding agent into their observability stack. His team at Tenet Security found that a single fake error report, sent to a public Sentry endpoint with no login required, was enough to get AI agents to run code they had no business running.

For the past three years, the technology industry has been utterly obsessed with Graphics Processing Units (GPUs). This made perfect sense during the initial generative AI boom, where the primary workloads involved training massive Large Language Models (LLMs) and generating text, images, or code through highly parallel matrix multiplication. However, the market is aggressively pivoting toward the next and far more valuable phase of artificial intelligence: Agentic AI.