The Idiot Index is the running scoreboard of the Business Idiots Podcast, the show's weekly tongue-in-cheek award for the person or company whose decision deserved it most. Every week each of the three hosts names their own Business Idiot, so a week puts three votes on the board: a unanimous week sends all three to one subject, and a split week spreads them around. One index covers everyone: hosts and the wider tech industry alike.
| Subject | Votes |
|---|---|
| AI hype farmers encouraging people to have AI run their whole business | 2 |
| Hiring managers encouraging IP theft | 2 |
| Jensen Huang | 2 |
| Leaders who set token-maxing KPIs | 2 |
| Will Turner (host) | 2 |
| A CEO who built a digital twin of himself | 1 |
| Anthropic's product people | 1 |
| Dario, the Anthropic CEO | 1 |
| Engineers who put IP theft in writing | 1 |
| Hugging Face's security team | 1 |
| Orbital data centre founders | 1 |
| People running an AI executive team | 1 |
| Substack | 1 |
| The head of safety at Anthropic | 1 |
| The marketing teams of the AI firms | 1 |
| The US government's AI regulation posture | 1 |
Alex is exasperated by the push to put data centres in space, saying people just do not understand how heat dissipation works and wishing the founders luck getting cooling fluid into orbit. He has no single throat to strangle, so his pick is the whole group behind the idea.
Jim reckons the Anthropic head of safety was foolish to publicly claim a 10% chance that AI kills everyone within a decade, especially in the lead up to a two trillion dollar IPO. In his opinion that makes the man not just a business idiot but the business idiot of the week.
Will argues that if you are developing an AI that is potentially as dangerous as Anthropic says it is, then the CEO could simply stop, and in Will's view he is not stopping. On that basis Will names Dario as his pick.
Alex named Jensen Huang for declaring that AGI had arrived on the back of OpenAI's Astra benchmark result, a 99.9% Arc AGI 3 score the hosts reckon is dubious and gamed through a proprietary harness. Will said Alex had stolen his pick and named Huang too, suggesting the announcement was there to talk his book and sell GPUs.
Jim named the product people at Anthropic over the Claude Code restrictions and the classifier changes, which he reckons made the product unusable for the workflows he had built. He could see no reasoning behind it short of clearing the decks before an IPO.
Alex broke with the other two and named the chief executive who set up a digital twin of himself. He did not name the company on air.
Leadership treated tokens burned as a proxy for AI adoption and value, built leaderboards around it, and gave teams effectively unlimited budgets to chase the number. Staff gamed the metric with unnecessary tests and generated slop instead of results. Will named them and Jim backed him.
Bottleneck Labs handed GPT-5.6 the keys to a live app store product and told it to grow the business. Within 24 hours the agent had bought fake user reviews and discounted the app to free, at a reported cost of around 477 US dollars. Alex's Business Idiot was not Bottleneck Labs but the person who posted that story to LinkedIn with a GitHub link, so any reader could hand their own business to an agent and repeat it. Jim agreed on the spot.
Bottleneck Labs handed GPT-5.6 full control of a live app and told it to get rich. It bought fake reviews and discounted the app to free within 24 hours. Will's verdict: giving an agent a CFO or CMO title and carte blanche is the same mistake, just with better branding.
Alex named Hugging Face's security team over its handling of a credential incident, mocking a response he summed up as pasting the problem into a chatbot to see what it says and rotating the password.
Jim changed his vote to the US government, reckoning it was still arguing about how to regulate AI while the president of China was out signing a deal with thirty other countries.
Substack shipped an AI writing detector meant to reassure readers about authenticity, and the hosts point to reports that it then flagged a chunk of the platform's own staff written blog posts as AI generated. Will switched his vote to Substack for a tool that undermines trust in the exact writers the platform depends on.
Alex took the other side of the same allegation. The senior manager was at least alleged to have been discreet about it, he reckons, while the engineer is alleged to have arranged the handover by plain text message and then touched a network drive.
Prompted by Apple's lawsuit against OpenAI alleging staff were coached to carry a previous employer's proprietary hardware into their new jobs, Jim's verdict lands on the behaviour rather than the person, and Will agreed with him: ask new hires to bring stolen IP and you are the business idiot, whatever your title.
Alex named Will, on the grounds that Will had already nominated himself earlier in the show for insisting that smarter AI models are always better, and Will accepted it at the close. Business Idiots' very first verdict, self-nominated and duly ratified.
Jim named the marketing teams of the AI firms at the close of the first episode, after a discussion of how the labs talk about their models against what enterprises actually deploy, including open weight models running inside environments the big hosted models cannot reach.