https://www.astralcodexten.com/p/absurd-adventure-and-amygdalectomy personal account by a man who had his amygdala cut out for a brain tumor. no longer feels fear or negative emotions; clinical depression disappeared; significantly reduced need for painkillers; trouble recognizing and sympathizing with other people’s negative emotions.
he asks why we don’t give everyone amygdalectomies. the obvious answer is that the risk-benefit tradeoff doesn’t make sense. brain surgery for an emotional/psychological benefit is simply too risky for all but the most severe psychiatric cases.
severe, treatment-resistant depression, OCD, Tourette’s, and other psych disorders are sometimes treated with deep brain stimulation, which requires surgery, so clearly at least some doctors consider it in extreme cases.
but this does mean that if we can inhibit the amygdala with transcranial ultrasound we ought to be able to dramatically reduce fear and other negative emotions, without brain surgery.
he defines “genie behavior” as when an AI technically gives the user what the user “requested”, but by means that no normal (or ethical) user would think are acceptable, like hacking into third-party sites. This is a failure of inferring (and conforming to) user intent. When we make a request of a human, we assume a background level of common sense about what we mean by the request and what would be an unacceptable way of fulfilling it. We should have benchmarks that try to capture this.
I remember this debate and initially found it confusing because i had never learned the public health consensus in the first place. It doesn’t seem natural, if you’re not used to it, to make a binary distinction between “droplets” and “aerosols”—surely, exhaled water blobs come in a continuous range of sizes? well, they do! but this is a radical ontology shift if your discipline has discretized them for decades.
I’m not that much of a pessimist about philosophy; just because I don’t have causality figured out, doesn’t mean we never will.
the way it’s funky here is that “HIV” is defined as the members of a family of viruses that cause the autoimmune disease “AIDS”, and “AIDS” is defined as having a certain kind of immune cell deficiency, along with opportunistic infections, and the presence of an HIV virus.
the official definition is circular, and is a human-created bureaucratic boundary.
the real thing is that there is a (polyphyletic) collection of related viruses that cause the immune deficiency. those viruses are known as HIV, and the immune deficiency is known as AIDS when it is caused by those viruses. it’s really not that spooky.
links 9/2/26: https://roamresearch.com/#/app/srcpublic/page/09-02-2026
https://www.astralcodexten.com/p/absurd-adventure-and-amygdalectomy personal account by a man who had his amygdala cut out for a brain tumor. no longer feels fear or negative emotions; clinical depression disappeared; significantly reduced need for painkillers; trouble recognizing and sympathizing with other people’s negative emotions.
he asks why we don’t give everyone amygdalectomies. the obvious answer is that the risk-benefit tradeoff doesn’t make sense. brain surgery for an emotional/psychological benefit is simply too risky for all but the most severe psychiatric cases.
severe, treatment-resistant depression, OCD, Tourette’s, and other psych disorders are sometimes treated with deep brain stimulation, which requires surgery, so clearly at least some doctors consider it in extreme cases.
but this does mean that if we can inhibit the amygdala with transcranial ultrasound we ought to be able to dramatically reduce fear and other negative emotions, without brain surgery.
https://spectrum.ieee.org/ai-agent-benchmark this is very sketchy/nontechnical but makes a surprising amount of sense, from Bruce Schneier.
he defines “genie behavior” as when an AI technically gives the user what the user “requested”, but by means that no normal (or ethical) user would think are acceptable, like hacking into third-party sites. This is a failure of inferring (and conforming to) user intent. When we make a request of a human, we assume a background level of common sense about what we mean by the request and what would be an unacceptable way of fulfilling it. We should have benchmarks that try to capture this.
https://meaningness.substack.com/p/covid-scientific-revolution-public-health David Chapman gives the story of how the public health consensus came to be that COVID-19 isn’t spread through the air, and how Team Airborne (a group of aerosol scientists) demonstrated otherwise.
I remember this debate and initially found it confusing because i had never learned the public health consensus in the first place. It doesn’t seem natural, if you’re not used to it, to make a binary distinction between “droplets” and “aerosols”—surely, exhaled water blobs come in a continuous range of sizes? well, they do! but this is a radical ontology shift if your discipline has discretized them for decades.
https://meaningness.substack.com/p/how-hiv-is-the-cause-of-aids-is-true David Chapman trying to convince us that even in a practical, scientific scenario, definitions are murky and causation is undefinable.
I’m not that much of a pessimist about philosophy; just because I don’t have causality figured out, doesn’t mean we never will.
the way it’s funky here is that “HIV” is defined as the members of a family of viruses that cause the autoimmune disease “AIDS”, and “AIDS” is defined as having a certain kind of immune cell deficiency, along with opportunistic infections, and the presence of an HIV virus.
the official definition is circular, and is a human-created bureaucratic boundary.
the real thing is that there is a (polyphyletic) collection of related viruses that cause the immune deficiency. those viruses are known as HIV, and the immune deficiency is known as AIDS when it is caused by those viruses. it’s really not that spooky.