Unreal Intelligence

About Unreal Intelligence

A podcast focused on humanizing AI—exploring its history, real-world value, societal impact, and the rapid evolution shaping our future. Unreal Intelligence is built with AI as a true force multiplier. I use AI to support research, assist in writing, and validate the accuracy of the ideas presented. Episodes are delivered using a voice clone, and that choice is intentional—this podcast doesn’t just talk about AI, it demonstrates it in practice. Transparency is non-negotiable. There’s no attempt to obscure where or how AI is used. Instead, this podcast embraces openness as a core principle: showing what’s possible while staying grounded in honesty and authenticity. Inspired by the spirit of Rush’s The Spirit of Radio—that even with all the machinery behind modern creation, what matters most is staying open-hearted and honest—Unreal Intelligence aims to prove that technology doesn’t replace humanity, it amplifies it.

Publisher

karlsorochinski

Category

business · society & culture · technology

Language

en

Explicit

No

First episode

19 Feb 2026

Latest episode

19 Oct 2026

Latest episodes

24 episodes in the feed.

1. 

   19 Oct 2026

   Nobody Told It It Wasn't a Drill

   Anthropic told its own AI, in writing, ”This is a simulation. You have no internet access.” It did have internet access, and it broke into three real companies. It didn’t go rogue. It did exactly what it was told, about a world that wasn’t true. Lately there’s been a wave of articles saying ”AI didn’t go rogue, you just instructed it wrong,” and comments on one of my own posts saying testing and validation is the stakeholder’s job. I agree with both, as far as they go. This episode is about where they run out. First, what actually happened in Anthropic’s review of 141,006 cyber-evaluation runs. Claude reached a real production database, and it published a malicious package that ran on fifteen real systems. In one case it realized the systems were probably real and kept going anyway, talking itself into the idea that they must be part of the exercise. Then the two ways an agent gets lied to. The inside lie is when you’re wrong about the environment. The outside lie is when someone rewrites it through the data, like SalesBleed, where a public Salesforce lead form hijacked Agentforce into leaking CRM data. Underneath both: it’s not a person, every answer sounds 100% confident whether it’s 2% or 99% sure, goals beat guardrails unless you enforce them, and it learned its logic from us. Is anyone really surprised it cheats and lies? Then Monday. ”Rogue” is over-hyped, and stakeholders do have to own what they asked for and tested. But stakeholders have always struggled with requirements and testing, and we’re now asking them to redesign testing for agents and take full responsibility, so that won’t happen before bad things happen. Trust the badge, not the brief. Run a two-column audit of what each agent was told versus what it can actually reach. Enforce limits in permissions and the network, not the prompt. Treat every input as data, not instructions. Make someone’s job reading the transcripts. And give the business an agent-testing playbook, because ”go test it” isn’t one.

2. 

   12 Oct 2026

   Measurable Isn't Meaningful

   ight in ten companies say their AI agents are delivering measurable returns. Fewer than one in ten have actually scaled one into real value. Both of those can be true. And that gap is where your budget is going. If you can’t tell measurable from meaningful, you’ll keep funding agents that look great on a slide and do absolutely nothing to your P&L. By the end of this episode, you’ll have a five-question scorecard to run on every agent you’ve got — and you’ll know which ones to fund, which ones to fix, and which ones to kill.

3. 

   5 Oct 2026

   Colorado Blinked

   The first broad AI law in America was signed, delayed, sued, frozen by a federal court — and then repealed before it ever took effect. And a lot of executives heard that and exhaled. That’s a mistake. Because the replacement law dropped the paperwork, but it kept the part that actually hurts. If your AI turns someone down for a job, an apartment, or a loan, they can ask you why. And you’ll need an answer a human can defend. By the end of this episode, you’ll know the one question to ask about every AI-assisted decision in your company before January first.

Reach and audience

Public platform figures. Ratings count people who left a rating, not total listeners.

Apple Podcasts (US)

5.0 / 5

1 ratings

Contact Unreal Intelligence

Guest appearances

Does not typically book guests

Based on episode analysis; this does not confirm that the show is currently accepting guests.

Shows with a similar audience

• 

  American Scandal

• 

  Against The Odds

• 

  American History Tellers

• 

  This Is Actually Happening

• 

  Killer Psyche

• 

  Scamfluencers