Methinks the US midterm election tips the scales a wee bit.

@Quroe How do you mean?

@dreev It signals an atmosphere of accountability to me. I'll say that I'm operating mostly on vibes -- and time decay.

@Quroe winter is coming

@MachiNi I accept the risk! (But have an exit order up at 5% for a week.)

@Quroe I meant another winter

@MachiNi I keep a radiometer in my cargo pants' daily load out.

@Quroe AI winter - Wikipedia

AI winter - Wikipedia

I don’t believe it but this is the winter the blue wave will cause, according to my failed attempt at humour.

@MachiNi Ooooh, I thought you meant AI triggering a nuclear winter.

I enjoy my misinterpretation more.

@Quroe I’m betting in your direction

This month now has a name, probably thanks to AI cyberoffense now being widely available.
Black October (information security) - Wikipedia
> In information security, October 2026 has been referred to as Black October due to several major security incidents and vulnerability disclosures.

I don't think this would or should cause a Yes resolution, but a time article claims that Trump's abduction of Maduro was based on a conversation with Grok. Ctrl+F Grok for it. Not tank battles of factions of US military though.

https://time.com/article/2026/10/01/donald-trump-2026-interview/

This should be way higher

@jim order up

@jim Why? Always curious to hear the thoughts of the nigh all-mighty (at least in confidence) Jim, who feels the exponential deeper in his bones than I ;P

@jim Why? Why, do you think this should be way higher?

@jim but @EliezerYudkowsky implied it shouldn’t. Or did he? Who knows.

@MachiNi here's some proprietary jimfund analysis (please don't share it anywhere) which should help you to understand my views.

Also, here's the distribution of arrival of ASI not conditioned on the fact that we haven't observed ASI yet

@jim are these images of Bayesianism?

@MachiNi i dont know what that is but yes probably

@jim updating my priors

Either this market is not being very Bayesian, or people think I'm going to misresolve it, or it was always hella weighted towards the last couple of months of 2026.

Yeah, weird for the market price not to drift down as the end of 2026 approaches. I think people feel uncertain whether you think Hugging Face and Navier-Stokes and such already count as all hell breaking loose.

And to be explicit, I hear you as saying no, you do not count events so far as qualifying. Possible rule of thumb: if a non-technical, news-ignoring person can, with a straight face, dismiss AI as hype, then all hell hasn't broken loose. They might be being willfully blind, but if they can still do that, it's still potential as opposed to kinetic hell, as another commenter put it. I'm aiming for something similar in my "life-changing for muggles" market.

PS: Claude kindly points out that (putting it my own words) in theory the market price could rationally stay flat-ish if the surprisingness of new evidence like superhuman math keeps managing to roughly cancel the evidence from the clock running out on 2026.

@dreev also if people thought the timing was going to be late in the year

@JussiVilleHeiskanen Right, Eliezer mentioned that one. My own feeling is that there's way too much uncertainty for that to have been reasonable.

Look, exponential growth is not intuitive. The expectation of new evidence feels strong on either side.

@EliezerYudkowsky or you’re being very obscure as usual

@MachiNi * and he's being obscure as usual ;P

(though in this case not actually as obscure as usual; this is very mild compared to a recent long tweet regarding the history of the communication of certain AI/singularity related ideas lol)