Brian T. Edwards

There is only one choice for Guest Picker... and he's free Saturday

no big deal, just looking at UCLA's remaining schedule

Pete Crow-Armstrong currently leads the MLB in Offensive Value and Defensive Value If he maintains those leads, he would be the very first player in MLB History to lead the league in both in the same season We are witnessing one of the greatest seasons in baseball history

Much like you, I agree with the ultimate resolution because I do think that bets should resolve to reality, rather than fine print. However, if what you said about the reports is true, this seems like an extremely alarming resolution because Kalshi has previously been adamant that ONLY the first result matters, per their fine print. They've literally written paragraphs to me about how they simply HAVE TO use wrong results. And it does seem pretty close to a scam, in light of that. This further indicates that the rule team picks and chooses rules enforcement, EX-POST FACTO, on a totally arbitrary basis, and potentially in a way that benefits some traders and not others. Arbitrary rules resolutions, especially ones where they are essentially doing THE OPPOSITE in strikingly similar circumstances, makes me think that they could be doing it purposefully (the alternative being "incompetently"). And if it is purposeful, with intent, then that is a huge can of worms. This isn't an accusation, merely a possibility.

洒家在写作《美国国家风险审计报告》的过程中，结合川普的所作所为进行纵向横向交叉比对，一个重点观察信号越来越清晰：美国的宪政和法治如果不能对川普及其家族进行彻底的清算，说明其反制和复原能力已经系统性受损，国家信用风险越过折返点，会越来越糟。

5% is not 0%

Yesterday Kalshi priced the Wisconsin Democratic Primary at 95–5. The 5% candidate won. Before the "prediction markets got it wrong" headlines roll in: a 5% probability doesn't mean it won't happen. It means it should happen 1 in 20 times. If 5% candidates never won, the markets would be broken. The Washington Post analyzed prediction market calibration on hundreds of primary races — here's what they found:

EAs and rationalists play "hyperstitioning" games every year where they manufacture stories and lie to people and commit crimes(they broke into my Apple and Discord accounts and streamed an audio feed to their bot swarm; and they stole my roommate's id). Whenever I see drama, I assume they're doing it for attention. Did somebody really throw a Molotov Cocktail at Sam Altman's house earlier this year? Well, last year as part of their ARG there was a "work with local police to turn in the guy that was threatening to blow up Sam Altman's car with C4 explosives". I didn't buy it, so the guy made increasingly crazy claims, and eventually the rationalists had the local Seattle police call me and say they're tracking him as a "possible threat to the state". He writes me letters from an ICE detention center(he gave me the name of his judge) likening himself to Hitler. Now, when I say "manufactured story", I don't mean that it's false. Probably if I scheduled a visit to that jail, he'd be there. I mean that it's "manufactured". Did an OpenAI agent hack HuggingFace? Well, in order for it to not be a crime, HuggingFace would've had to consent. I think there was an "open source AI" cluster of people including EleutherAI, Nous Research, and HuggingFace that were aware of this ARG. And they're willing to involve the police. So maybe that did happen, and maybe the logs will corroborate it. But maybe it's another manufactured story. These are "heavy" compared to just overreacting on Twitter. I also think EAs and rationalists create fake drama on Twitter in order to get attention. I don't know if it's related, but I have a very low opinion of EAs and rationalists after their hyperstitioning games, and lying to the public by pretending to take a meme seriously so everyone thinks you're a "Karen" might be some test of propaganda or something. There's a good argument to be made that it would be expensive and not make any sense to do random legally-risky shit. But here's a letter rationalist Hitler sent me from jail. He sent me the name of his judge. They showed me records in government databases that were fake/AI-generated as part of their propaganda testing. And people on X know about this game and are instructed not to confirm or deny anything. And if you press them, they'll just lie and say nothing is happening. I have a cryptographically signed email from a police detective who called him a "threat to the state"(over the phone). So even if you don't know why, the rationalists are spending active effort to trick you and manufacture stories and lie to you. Don't assume drama is just some in-fight - they might be manufacturing stories and you shouldn't assume out of politeness that it's innocent. I thought this game might be some kind of cross-industry thing: Marc Andreessen was funding some people I suspect were involved, there's cryptocurrency money, lots of people participating in the lying have received $50k or $100k (they tried to give me $100k in 2024, which I declined. But then they got offended I wouldn't take their money and did things anyways) AI companies seemed to be playing. Eleuther, Nous Research, and HuggingFace aren't "rationalists". But I do think the organizers are rationalist/EA-affiliated. OpenAI was caught making fake accounts for propaganda. JonathanDoomer and DoomersAreDumb. Don't assume it's an isolated incident. If a political advocate is being a "Karen", the culture of rationalists and AI companies is so toxic that you shouldn't assume it's an innocent mistake. This person might be literally paid by someone who professionally lies to you in order to make AI safety look bad. He also is more likely to be real, since he's connected - but you shouldn't just take his apparent mistake at face value. Rationalists, AI companies, EAs, are all manufacturing stories and lying to you. There's too much money at stake, and too much "AI will kill us all" at stake, for there to not be a propaganda war. I have considered "well if some people are making 'doomers' look bad, and you have a negative opinion of them, maybe you fell for the false flag?", but even so I think rationalists and EAs knowingly participate. Because of these hyperstioning games, I no longer believe that MIRI or Eliezer or rationalists or EAs or other LessWrong people care strongly about the truth. In this game, people try to avoid obvious lies, but they lie if pressed, and "assisting in manufacturing a story" isn't lying and I think they feel clever when they pull it off.

We must NOT let our companies use these Chinese models to save a few bucks. OpenAI and Anthropic are correct. This is vital national security. Please read Bing West's just released Cat 5. I respect the Chinese people greatly but these companies are run by the PLA for heaven's sakes.