UnHacked the podcast with hosts Justin Shelley, Bryan Lachapelle, and Mario Zaki.
Justin's AI subscription quote looked reasonable, until the bill came in at 10x the price. No warning, no explanation, just a massive jump. So what's really going on with AI pricing, and is the "AI will end humanity" panic actually about something else entirely?
In this episode of UnHacked, Justin Shelley, Mario Zaki, and Joshua Holloway go fully unscripted and unstructured to dig into why AI companies keep dropping doomsday headlines about their own products while simultaneously racing to build bigger, faster, more powerful models. They lay out three competing theories: it's a real security problem, it's a push for regulation with a hidden agenda, or it's a marketing stunt designed to keep businesses locked into expensive subscriptions instead of running AI in-house.
The conversation gets specific fast. Justin breaks down the real invoice that tripled his AI costs by 1,000%, not 10%. The team talks through why "slowing down AI development" is a meaningless promise when no single company or country can actually afford to fall behind. They also connect the dots to the exploding cost of server hardware (from roughly $15K to $50K with three-month lead times) and why that pricing surge benefits the same companies pushing the fear narrative. Along the way, Mario shares a real story about a client who got sold an eight-year-old server as brand new, a reminder that shady vendor practices aren't limited to AI.
What you'll learn:
- Why one AI vendor's pricing jumped 1,000% with zero warning, and what that says about subscription lock-in
- Why "pausing AI development" is likely impossible to enforce and may just be marketing
- How server and GPU hardware costs tripled in under a year, and who actually benefits from that
- Why AI companies might be more afraid of businesses self-hosting open-source models than they let on
- How to spot when a vendor sells you old hardware disguised as new, and why that's already fraud
