The AI Fix

In episode 100 of The AI Fix, Mark is joined by Corey Noles for a milestone episode devoted to the future of AI, how it will change our lives, and whether it will keep us as pets or use us as batteries.
Also in this episode, another tech companies fires a bunch of staff; OpenAI and Anthropic announce mega Wall Street joint ventures within minutes of each other; Anthropic rents Elon's GPUs; and an imaginary fridge takes a gun to a grocery store.
Episode links:
Coinbase didn't just lay off 14% of its staff due to AI. It replaced managers with 'player-coaches' and turned its org chart upside down.
OpenAI Finalizes $10 Billion Joint Venture With PE Firms to Deploy AI.
Anthropic, Goldman and others launch $1.5 billion AI venture.
Anthropic, SpaceX announce compute deal, includes space development.
Kimi K2.6 Leads OpenRouter with 1.88T Tokens.

The AI Fix
The AI Fix podcast is presented by Mark Stockley.
Grab T-shirts, hoodies, mugs and other goodies in our online store.
Learn more about the podcast at theaifix.show, and follow us on Bluesky at @theaifix.show.
Never miss another episode by following us in your favourite podcast app. It's free!
Like to give us some feedback or sponsor the podcast? Get in touch.
Support the show and gain access to ad-free episodes by becoming a supporter: Join The AI Fix Plus!

Our Sponsors:
* Check out Anthropic and use my code claude.ai/aifix for a great deal: https://www.anthropic.com

Advertising Inquiries: https://redcircle.com/brands

Privacy & Opt-Out: https://redcircle.com/privacy

In episode 99 of The AI Fix, guest host Lianne Potter explores the disturbing dark side of AI companions, while Mark meets Talkie, a chatbot trapped in 1930.
Also in this episode: Mark takes photos at the end of the world; Lianne builds an incel detector; a bizarre sneaker startup pivot; too much talk about kettles; and Elon Musk and Sam Altman argue over custody of the dog.
Episode links:
Shoe company Allbirds pivots to AI compute in sign of a totally normal and healthy economy.
NSA spies are reportedly using Anthropic's Mythos, despite Pentagon feud.
Elon Musk testifies against OpenAI, seeking Sam Altman's ouster.
Amazon launches an AI-powered audio Q&A experience on product pages.
Introducing talkie: a 13B vintage language model from 1930.

The AI Fix
The AI Fix podcast is presented by Mark Stockley.
Grab T-shirts, hoodies, mugs and other goodies in our online store.
Learn more about the podcast at theaifix.show, and follow us on Bluesky at @theaifix.show.
Never miss another episode by following us in your favourite podcast app. It's free!
Like to give us some feedback or sponsor the podcast? Get in touch.
Support the show and gain access to ad-free episodes by becoming a supporter: Join The AI Fix Plus!

Our Sponsors:
* Check out Anthropic and use my code claude.ai/aifix for a great deal: https://www.anthropic.com

Advertising Inquiries: https://redcircle.com/brands

Privacy & Opt-Out: https://redcircle.com/privacy

In episode 98 of The AI Fix, guest host Nik Roberts explains the cognitive debt you're racking up every time you let ChatGPT do your thinking, and Mark explains the new eval that's got even the very best AI models completely stumped.
Also in this episode, Nik learns how to pronounce "Mythos" and Mark learns how to pronounce "Kasparov"; Gen Z fights back against AI; Britain's AI Security Institute tests Mythos; DeepSeek casually drops a 1.6 trillion parameter model; and a Sony robot beats the elite at the world's most boring pastime.
Episode links:
Gen Z Sabotaging AI at Work So It Won't Take Their Job.
Our evaluation of Claude Mythos Preview's cyber capabilities.
DeepSeek previews new AI model that 'closes the gap' with frontier models.
Inside Project Ace: Discover the Robot Athlete That Competes With Professional Table Tennis Players.
Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task.
Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity.
The Verification Bottleneck: Why AI's Real Cost Is Human Attention.
Announcing ARC-AGI-3.

The AI Fix
The AI Fix podcast is presented by Mark Stockley.
Grab T-shirts, hoodies, mugs and other goodies in our online store.
Learn more about the podcast at theaifix.show, and follow us on Bluesky at @theaifix.show.
Never miss another episode by following us in your favourite podcast app. It's free!
Like to give us some feedback or sponsor the podcast? Get in touch.
Support the show and gain access to ad-free episodes by becoming a supporter: Join The AI Fix Plus!

Our Sponsors:
* Check out Anthropic and use my code claude.ai/aifix for a great deal: https://www.anthropic.com

Advertising Inquiries: https://redcircle.com/brands

Privacy & Opt-Out: https://redcircle.com/privacy

In episode 97 of The AI Fix, Mark is joined by Ethan Ding, one of the sharpest contrarian voices in AI, to explore why almost half of America's announced data centres aren't being built, and why AI job losses are not what they seem.
Also in this episode, California approves a startup to let AI prescribe antidepressants; Sam Altman pitches a New Deal for the age of superintelligence while sitting on the biggest pile of private capital ever raised; and Anthropic decides it's built a model so dangerous it can't let anyone use it.
Episode links:
Almost Half of US Data Centers That Were Supposed to Open This Year Slated to Be Canceled or Delayed.
AI Psychiatry Startup Approved to Prescribe Meds.
Behind the Curtain: Sam's superintelligence New Deal.
Project Glasswing: Securing critical software for the AI era.
TextQL - AI Data Analytics.
mandates | Ethan Ding | Substack.

The AI Fix
The AI Fix podcast is presented by Mark Stockley.
Grab T-shirts, hoodies, mugs and other goodies in our online store.
Learn more about the podcast at theaifix.show, and follow us on Bluesky at @theaifix.show.
Never miss another episode by following us in your favourite podcast app. It's free!
Like to give us some feedback or sponsor the podcast? Get in touch.
Support the show and gain access to ad-free episodes by becoming a supporter: Join The AI Fix Plus!

Our Sponsors:
* Check out Anthropic and use my code claude.ai/aifix for a great deal: https://www.anthropic.com

Advertising Inquiries: https://redcircle.com/brands

Privacy & Opt-Out: https://redcircle.com/privacy

In episode 96 of The AI Fix, Mark uncovers new research that shows AIs will lie and scheme to save each other, and Mike Fey, CEO of episode sponsor Island, drops in to talk about the future of enterprise work.
Also in this episode, Claude learns to make skills that make skills that make skills; Gemini wants to rummage through your digital life; OpenAI quietly turns ChatGPT into a $100 million advertising juggernaut; and a Waymo robotaxi takes a wrong turn through a school zone.
Episode links:
Skillception — On the Recursive Limits of Meta-Skill Generation.
Google rolling out Personal Intelligence to free Gemini, Chrome, & AI Mode users.
ChatGPT ads hit $100M in six weeks — and OpenAI is just getting started.
Waymo car goes wrong way in Alamo Heights school zone, concerns parents.
AI models will secretly scheme to protect other AI models from being shut down, researchers find.
Island.

The AI Fix
The AI Fix podcast is presented by Mark Stockley.
Grab T-shirts, hoodies, mugs and other goodies in our online store.
Learn more about the podcast at theaifix.show, and follow us on Bluesky at @theaifix.show.
Never miss another episode by following us in your favourite podcast app. It's free!
Like to give us some feedback or sponsor the podcast? Get in touch.
Support the show and gain access to ad-free episodes by becoming a supporter: Join The AI Fix Plus!

Our Sponsors:
* Check out Anthropic and use my code claude.ai/aifix for a great deal: https://www.anthropic.com

Advertising Inquiries: https://redcircle.com/brands

Privacy & Opt-Out: https://redcircle.com/privacy