Newsletter
Positive Intelligence pt.III and other things
May 30, 2025
Hello,
Here’s everything since my last little missive to you:
Excerpt: A lot of people were upset with me for teasing the ‘neuroscience-based’ coaching programme ‘Positive Intelligence’, so I thought I’d do a little autopsy. This is part three, on the brain science… Such as it is.
Main idea: This might be the most comprehensive example of the neuroscience confidence game I’ve ever written about. That and a heavy dose of self-indulgence. Neuroscientific self-help, not so much.
The entangled brain. I liked this one, as an intro into the frontiers of how we think about cognitive neuroscience (i.e. how does the brain do thinking):
When thousands of starlings swoop and swirl in the evening sky, creating patterns called murmurations, no single bird is choreographing this aerial ballet. Each bird follows simple rules of interaction with its closest neighbours, yet out of these local interactions emerges a complex, coordinated dance that can respond swiftly to predators and environmental changes. This same principle of emergence – where sophisticated behaviours arise not from central control but from the interactions themselves – appears across nature and human society.
As in abstractions as gods, looking at things at the level of near-agentic systems often seems like the right level at which to consider the really complicated stuff, brain included. Pretty article.
–
Get ADHD Meds with ChatGPT. Research paper on the more unexpected utility of AI coaching:
This preregistered study aimed to assess whether AI-generated coaching helps students to successfully feign attention-deficit/hyperactivity disorder (ADHD) in adulthood. First, based on questions generated by 22 students, we conducted an extensive ChatGPT query to develop a concise AI-generated information sheet designed to coach students in feigning ADHD during a clinical assessment. Second, we evaluated the effect of this coaching in an experimental analogue study in which 110 university students were randomly assigned to one of three groups: (1) a control group (n = 42), (2) an ADHD symptom–coached simulation group (n = 35), and (3) an AI-coached simulation group (n = 33). All participants underwent a clinical neuropsychological assessment that included measures of ADHD symptoms, functional impairments, selective attention, and working memory. Our preregistered data analysis revealed that the AI-coached simulation group consistently moderated their symptom overreporting and cognitive underperformance compared to the symptom-coached group in small to medium size, resulting in lower detection sensitivity. We conclude that publicly accessible AI tools, such as current versions of chatbots, can provide clear and effective strategies for feigning ADHD during clinical neuropsychological assessments, posing a significant threat to the validity assessments.
–
Superstition In A Godless State. On the rise of superstition in the notoriously anti-superstitious Chinese State. It’s hard not to see parallels in our own countries:
Young Chinese are not naturally more superstitious. But they are trapped in an unstable system, and with no clear future, they are buying ready-made ones. These crystals and tarot cards aren’t ancient traditions—they’re quick-fix stories built from what’s left in the marketplace. Meanwhile, sellers and platforms continue testing how much people are willing to pay to ease their fears.
–
How To Rig A Clinical Trial. This one’s a great example of the problems with p-values:
Here’s a dirty little science secret: If you measure a large number of things about a small number of people, you are almost guaranteed to get a “statistically significant” result. Our study included 18 different measurements—weight, cholesterol, sodium, blood protein levels, sleep quality, well-being, etc.—from 15 people. (One subject was dropped.) That study design is a recipe for false positives.
Think of the measurements as lottery tickets. Each one has a small chance of paying off in the form of a “significant” result that we can spin a story around and sell to the media. The more tickets you buy, the more likely you are to win. We didn’t know exactly what would pan out—the headline could have been that chocolate improves sleep or lowers blood pressure—but we knew our chances of getting at least one “statistically significant” result were pretty good.
Whenever you hear that phrase, it means that some result has a small p value. The letter p seems to have totemic power, but it’s just a way to gauge the signal-to-noise ratio in the data. The conventional cutoff for being “significant” is 0.05, which means that there is just a 5 percent chance that your result is a random fluctuation. The more lottery tickets, the better your chances of getting a false positive. So how many tickets do you need to buy?
With our 18 measurements, we had a 60% chance of getting some“significant“ result with p < 0.05. (The measurements weren’t independent, so it could be even higher.) The game was stacked in our favor. It’s called p-hacking—fiddling with your experimental design and data to push p under 0.05—and it’s a big problem. Most scientists are honest and do it unconsciously. They get negative results, convince themselves they goofed, and repeat the experiment until it “works.” Or they drop “outlier” data points.
–
Does Testosterone Effect Economic Preferences? This paper says no:
Participants were randomly allocated to receive a single dose of either placebo or intranasal testosterone. They thereafter carried out a series of economic tasks capturing social preferences, competitiveness and risk preferences. We fail to find any evidence of a treatment effect for any of our nine primary outcome measures, thereby failing to conceptually replicate several previous studies reporting positive findings that used smaller sample sizes. In line with these results, we furthermore find no evidence of an association between basal testosterone and economic preferences, failing to also conceptually replicate previous correlational studies.
People really love the nature account, possibly because it absolves us of personal responsibility. But nature is nurture and vice versa, and genetics always seems to get pushed further into the margins.V
–
THE WAY OF CODE: The Timeless Art of Vibe Coding. It’s a Rick Rubin project, so you know it’ll be interesting. Ripping off Lao Tzu in a playful way (I mean, I assume it’s not serious, Rick Ruben can’t have that big an ego). See vibe coding as a meditative practice. It does capture a bit of what I get putting this site together.
–
Pitfalls of AI for information gathering. Nominally about openai’s Deep Research, but applicable to all:
- it excels at straightforward questions, struggles at creative ones; and
- it suffers the tyranny of the authority—drawing “on ideas that are frequently discussed or published, rather than the best stuff”.
The third point the author raises is that:
you find yourself taking intellectual shortcuts. Paul Graham, a Silicon Valley investor, has noted that AI models, by offering to do people’s writing for them, risk making them stupid. “Writing is thinking,” he has said. “In fact there’s a kind of thinking that can only be done by writing.” The same is true for research. For many jobs, researching is thinking: noticing contradictions and gaps in the conventional wisdom. The risk of outsourcing all your research to a supergenius assistant is that you reduce the number of opportunities to have your best ideas.
I think this is just unthoughtful use though, so it’ll only affect people who are already doing this in other areas.
It all still points, at least in the short term, to the need for human supervision. I used o3 to help with my teardown of Positive Intelligence, and while it sped up the process of clarifying my intuitions, it failed to get past superficial critiques, and made a bunch of truthy, but ultimately inaccurate claims.
All of which raises the question of how do we get non-experts (e.g. kids) to the place where they can supervise, when they’ll be using ai to get there.
–
I hope you found something interesting.
You can find links to all my previous missives here.
Warm regards,