Don't Stop Us Now AI Edition

Ming delivers a genuine cluster of non-obvious, research-backed claims per minute - gamma band suppression linked to cognitive decline, the automator/validator/cyborg taxonomy drawn from original experiments, the colonoscopy degradation finding, and the Polymarket prediction result where cyborgs beat both solo humans and state-of-the-art AI. Some repetition and host-filler dilute the density slightly, but the ratio of real claims to padding is above average for the genre.

“What I see in my research is this measure of cognitive engagement that is very robust, called the gamma band. Activity goes way down. Like the difference between working on a Hard math problem and just watching tv. Enough of a drop that it actually worries me about things like early cognitive decline, dementia rates, Alzheimer's rates.”

“naive kids, uh, having to make 10 predictions in one hour were doing not just better than the state of the art model. At that time. I took the questions off of Polymarket and they were doing nearly as well as the professionals with millions of dollars on the line at polymarket.”

The 'AI passivity harms cognition' thesis is increasingly common discourse, but Ming earns originality points for her specific reframe of the Turing Test, the proposal to benchmark AI models on all-cause mortality and friendship-network size rather than autonomous task performance, and the concrete nemesis prompt technique - these are not recycled LinkedIn frameworks. The overall frame still sits in recognisable 'human-AI collaboration' territory.

“I wrote a blog post saying actually humans failed the Turing Test because a true Turing Test pass would be 50 50. You can't tell the difference. Humans thought GPT was the human 75% of the time.”

“which of these models has the lowest all cause mortality of any user base? Which of these models has the highest income and wage accumulation or the largest friendship networks?”

Ming is a genuine practitioner: 30 years building neural network models before LLMs existed, active empirical research (the Berkeley/Polymarket experiment, the 63,000-company gender-wage analysis), six founded companies, and current projects spanning Alzheimer's neurotechnology and a hybrid intelligence benchmark - she is decidedly not a career podcast guest or a pure thought leader. The main caliber caveat is that her current work is largely philanthropic/research rather than at-scale commercial operator.

“I built my first model nearly 30 years ago. So this is long before agentic models or the Nobel Prizes recently.”

“I analyzed the faces and read the quarterly reports of 63,000 companies and came up with a novel finding about gender wage gap”

Ming names specific studies (Anthropic Claude Code paper, Turing Test at UC San Diego, MIT/Harvard cyborg research, Polymarket-sourced questions), specific models (Gemini 3, Llama, Opus 4.8), specific metrics (gamma band activity, 75% Turing misidentification rate), and specific companies (63,000 quarterly reports). Docked for repeatedly citing studies without full attribution ('a group at MIT,' 'a great study I think it needs to get replicated') and some claims that hang without empirical anchors.

“Anthropic's own research on clog code shows a similar messy story, which is for a large percentage of clog code users, but not all of them. Their conceptual understanding of coding gets worse over time”

“Some colleagues of mine at UC San Diego ran the classic Turing Test, and I think they really genuinely did it. Not one of these thought leader statements. They ran a robust example of the Turing Test.”

The hosts keep a coherent arc and occasionally contribute useful paraphrases that sharpen Ming's arguments, and the question about what she would tell AI lab founders is well-placed. However, they never push back on a single empirical claim - including bold ones about dementia risk and cognitive decline - and default to affirmation ('I love that vision,' 'Yeah, fascinating'), leaving multiple threads underexplored and several strong assertions unchallenged.

“So if I was to paraphrase what you've kind of just been talking about, the superficial, potentially harmful way of engaging AI is to delegate in a kind of almost like a one shot.”

“I love that vision. I think it's uh, it's so compelling.”