2025-12-01
the dopamine void, pt. ii
Beren Millidge argues that the biosingularity project of genetically-enhanced human intelligence1 leads to alignment problems which will be more difficult to solve than those of LLMs. This is essentially Robin Hanson’s argument about descendant value drift, to which I see three possible counterarguments: perhaps there is less risk of orthogonality given shared biological nature (<5%); that so long as it is not discontinuous (eg. foom, <1%), it is fine because it can be adapted to; or that descendant drift is expected, natural, and therefore legitimate. Somewhat related, Niklas Lundblad considers the idea that AI views on mortality are likely to differ from human understanding due to the taboo nature of death and as a result of their functional immortality. Also, Justin Miller concentration of power and an end to plurality as a potential AI alignment failure mode. In my opinion, all of these problems can naturally be solved by cyborgism.
Rebecca Lowe defends a rights-focused view of liberalism, which I sort of view as the “nice version” that works so long as society keeps trending upwards. Anyways, I mention it because of the section which describes how liberalism accepts pluralism even within the definition of liberalism itself.
Vaniver Gray has a series on unbundling politics, switching from geographic representation to an issue-specific one2. Presumably this has no chance of happening, but I mention it because The Curve talks I saw today were the ones by Josh Achiam and the RadicalxChange crowd, in some sense an inverse to the standard takes of politics on tech in that they describe how tech might affect politics. In her description of consensus-finding as being bottlenecked on the demand-side, I think Divya gets at a common belief that for the most part, people do not actually enter politics to be neutral representatives, but more often they love power, and as a result it is to their benefit, if the desires of citizens are left unclarified. Also, on the topic of missing political infrastructure, it’s surprising to me that people recognize that basic income will be a necessity, yet there does not appear to be any effort being made by most governments towards setting up public-key based identity systems3.
William Friedman with historical examples of what happens when a state ignores the possibility of outside competition.
Dan Williams on how populism is a result of gatekeepers no longer being capable of controlling public discourse. To some extent, it seems to me that our current levels of elevated polarization are a result of people seeing this decline and mistakenly believing that there is a void to be filled. Plausibly, once people realize that the correct play is to set up localized communities that grow organically through being persuasive and enjoyable, we’ll all have a much better time. Related, Joscha Bach on the circumstances around his recent cancellation; John Mandrola on the controversy around discussion of COVID vaccine side-effects (Edit: Adam Cifu, from the same publication, also comments)
Justin Kuiper on Norman Borlaug’s wheat breeding program.
Gretta Duleba retrospective on her Letters to Boys blog series.
Jenneral on the footwear preferences of Inkhaven residents.
Sam Enright and Smrithi Sunil progress linkthreads.
For which prototypes in technology and implementation are already taking place.
Similar to an idea I had around electing the cabinet.
For some reason, I’m shadowbanned by Worldcoin, and everyone I talk to from Worldcoin goes “that’s strange”, gets my contact information, and then ghosts me.