AI Rebels

95% of enterprise AI projects fail, but not for the reasons you think. Ian Cook has spent 16 years shipping AI products across healthcare, physical security, consumer goods, and now cultural data, and the pattern behind the failures is always the same: companies start from the top down with vague mandates instead of solving a specific person's specific problem. In this episode, Ian breaks down his framework for AI implementations that actually stick. Start bottom-up, ask a tangible question, and know what "done" looks like before you write a line of code. We get into why the cost of throwing away code is now zero and what that means for how fast you can experiment, how tools like Claude Code have changed what a solo engineer can ship in an afternoon, and which industries are about to get hit hardest by this wave. Ian doesn't sell magic. He builds prototypes that put working software in people's hands, and his track record speaks for itself.

What happens when AI leaves the screen and enters the physical world? It breaks.

Spencer and Jacob sit down with robotics researcher Grigorij Dudnik, who's been running real experiments with real robots—and finding that most of our assumptions about AI fall apart the moment hardware gets involved. The big one: the idea that a single massive model can do everything. Grigorij makes the case that the future isn't a "super-intelligent" humanoid. It's a modular system where an LLM plans, specialized models act, and physical constraints keep everything honest.

They get into why humanoids are overrated, why generalization keeps failing in robotics, and why a system built on narrow, composable skills might be the actual breakthrough everyone's overlooking.

How does generative AI actually work inside a large, regulated fintech company when real customers, real money, and real regulations are on the line?

The AI Rebels Crew sits down with Jake from Brex’s Customer Experience Strategy team to reveal what actually happens when AI moves from hype to production. You’ll hear how a major fintech deploys AI in customer support without sacrificing trust, compliance, or the human experience, and why most AI implementations fail to do the same.

We break down the metrics that matter (and the ones that quietly lie), why “containment rate” can mask bad experiences, and how Brex measures whether AI is genuinely helping customers or just making dashboards look good. Jake also explains how his team evaluates AI mistakes versus human error, and what responsible recovery really looks like at scale.

Along the way, you’ll learn why friction isn’t always the enemy, how guardrails prevent costly failures, and why the most effective AI systems don’t replace people, they make the right work possible for them.

If you want a clear, practical understanding of how AI succeeds in the real world (not in slide decks) this episode is worth your time.

(AI Generated Summary)

brex.com

Most companies are racing to adopt AI, but almost none can explain who’s responsible when it goes wrong. Cordell Robinson of Brownstone Consulting joins the AI Rebels crew to unpack the uncomfortable truth about AI governance, security, and compliance in a world moving faster than regulation. Why is the U.S. is lagging behind on enforceable AI rules, how can existing frameworks like NIST and ISO be adapted to fill the gap, and what does “good governance” actually look like in practice? These are all questions you'll hear answers to. Cordell also shares why annual security checks are no longer enough, and how agentic AI could shift penetration testing from a once-a-year ritual to an on-demand capability. If you’re building, deploying, or betting on AI, this conversation will change how you think about risk, responsibility, and readiness.

https://bcf-us.com/

Chatbots are starting to feel less like tools and more like something you relate to, and that shift comes with real risks. On AI Rebels, we sit down with Ryan “Zuda” Satterfield, the builder behind Dreami.me, to unpack the “AI psychosis” controversy: how overly agreeable models can validate bad ideas and create safety problems for anyone shipping companion AI. Ryan breaks down the guardrails he is implementing, from reducing sycophancy to shutting down “messiah” narratives, and explains why he refuses to build romance features even if it costs users. Then we zoom out to the big questions: is “simulated consciousness” just clever prompting, or can agency emerge from enough structure and memory? We close with the messy future of copyright and training data, plus why human-made art may become the premium tier in an AI-saturated world.

https://dreami.me/

Chapters
00:00 — What is Dreami.me? “Digital friend” built on “simulated consciousness”
03:05 — “AI psychosis” and the danger of the default yes-man chatbot
06:52 — Guardrails in the real world: blood pacts, “I’m God,” and the “Messiah Complex”
13:46 — How Dreami tries to avoid it: system prompts + fine-tuning + grounding responses
24:40 — “It’s my code, not the prompt”: the ‘be you’ experiment
25:30 — Parasocial dynamics: when “companionship” turns into dependency
35:15 — Consciousness talk: emergence, “patterns upon patterns,” and functionalism
39:59 — Copyright & training data: lawsuits, scraping, and “use my data”
43:51 — The future of art: AI commodity output vs “human-made” premium value
49:22 — Interrupting unhealthy loops: the assistant should challenge repetitive patterns

AI isn’t magic, and pretending it is may be the biggest risk in tech right now. A former financial advisor turned AI founder argues that most AI products are already becoming commodities, with little defensible moat and wildly inflated valuations. He explains why real value comes not from “agentic” hype, but from deep domain expertise, careful guardrails, and knowing what not to automate. In regulated industries like finance, trust, liability, and human judgment still matter more than raw intelligence. The conversation cuts through bubble thinking, warning that many AI startups won’t survive once capital dries up. In the end, AI raises the floor of expertise—but the ceiling still belongs to humans who can synthesize, decide, and take responsibility.

https://finmate.ai/

AI is now spotting the news before journalists even know it exists. DeepNewz’ founder reveals how their system scans real-time signals from across the internet to uncover emerging stories faster than any newsroom. He breaks down why writers are secretly relying on AI, how tiny teams are scaling like giants with agents, and where media is headed next. The conversation uncovers the shift from simple article summaries to true signal discovery and personalized news intelligence. If you want an early look at the future of information, this is a conversation you can’t afford to miss.

Most retail traders don’t realize they’re battling algorithms that execute 70–80% of all market activity. Robert Grzesik reveals how Bot Spot turns plain-English instructions into actual Python trading bots capable of running on real brokerages. The discussion digs into AI-assisted backtesting, strategy optimization, and agentic systems that can research news or sentiment before placing trades. A new marketplace for user-built algorithms hints at a future where anyone can publish or refine automated strategies. What emerges is a clear view of algorithmic trading becoming not just more powerful—but radically more accessible.

00:00 — The Problem: Retail Traders vs. Automated Markets
03:20 — Why Algorithmic Trading Is Hard for Individuals
06:45 — The Inspiration Behind Bot Spot
10:12 — Building Trading Bots with Natural Language
14:30 — Backtesting, Deployment, and Real-World Execution
18:55 — AI Research Agents & Agentic Trading Systems
22:40 — Crypto, DeFi, and Expanding Beyond Traditional Markets
26:05 — LumiBot Origins and the Path to Bot Spot
31:10 — Marketplace for User-Built Trading Algorithms
36:20 — The Future of Finance: Fully Autonomous AI Investing

https://www.botspot.trade/
https://www.linkedin.com/in/robertgrzesik/

AI note-takers are already obsolete, and Camden Bean, CEO of Ping Assistant, explains why the real power is everything that happens before and after the meeting. In this episode, he reveals how Ping is building a full-context engine for accountants that remembers every client detail, automates emails and documents, and turns hours of work into minutes. Camden breaks down the wild speed at which modern builders can create using AI, including how he built in a weekend what used to cost $40,000 and months of engineering time. He shares a transparent look at how Ping co-creates features with accounting firms, why specialized AI beats generic tools, and what the next wave of automation will look like. If you want to understand where AI is actually heading, and how fast it’s coming, this conversation is a must-listen.

Austin Allred lived in his car in San Francisco teaching himself to code, and now he runs a program where students never write a single line. At Gauntlet AI, engineers learn to direct AI systems that do all the coding while companies pay recruiting fees upfront, making the entire training free for students who land $200K jobs. The model works because the skills gap has reversed: employers are now bottlenecked by talent, not the other way around. Allred's most practical insight gets overlooked by most users: when AI fails, ask it what went wrong, the models can often tell you exactly where they're likely to be incorrect. Traditional coding education didn't slowly evolve; it hit a wall, and the people learning to work alongside AI rather than compete with it are the ones companies are fighting over.