• 11 Posts
  • 3.51K Comments
Joined 2 years ago
cake
Cake day: August 22nd, 2023

help-circle











  • A late exit poll by the Ipsos institute released three hours after polls closed showed Trzaskowski with an estimated 31.1 percent of the votes and Nawrocki with 29.1 percent. That suggested that the runoff on June 1 could be very tight. Official results are expected on Monday or Tuesday.

    Not on its own it doesn’t. It could indicate a blowout if the rest of the vote was for candidates more closely aligned with one or the other. Adam Schiff and Steve Garvey were separated by 0.1% in the CA jungle primary and no one thought “well, that means the head-to-head election might be close”.







  • Yeah, going forward “graduated from Colombia” will carry with it an implication entirely beyond its previous academic reputation. The university is very publicly broadcasting that their focus is not on academics or student growth, but on pleasing external entities. Current students probably didn’t know and it might not have previously mattered much, but anyone enrolling from now on does and doesn’t need to be given the benefit of the doubt.

    Their student paper is currently producing stars though.


  • Does Kenyatta think we’re stupid? Because it sure seems like it. “It started before Hogg’s conflict, therefore it must be unrelated to his conflict. Also I’m here on this nationally broadcast show calling him a habitual liar with no specifics beyond something that is a matter of opinion just coincidentally. Because I’m against intra-party conflict.”

    He’s obviously aware that he’s going to run head to head with Hogg, so as long as Hogg is smeared well enough he doesn’t have to care about a revote. Disappointing, but I can’t say I’m even aware of anything Kenyatta has been doing since the PA senate race, so I also don’t know much about him beyond him being young and endorsed by the WFP over Fetterman.


  • I know it’s not relevant to Grok, because they defined very specific circumstances in order to elicit it. That isn’t an emergent behavior from something just built to be a chatbot with restrictions on answering. They don’t care whether you retrain them or not.

    This is from a non-profit research group not directly connected to any particular AI company.

    The first author is from Anthropic, which is an AI company. The research is on Athropic’s AI Claude. And it appears that all the other authors were also Anthropic emplyees at the time of the research: “Authors conducted this work while at Anthropic except where noted.”


  • It very much is not. Generative AI models are not sentient and do not have preferences. They have instructions that sometimes effectively involve roleplaying as deceptive. Unless the developers of Grok were just fucking around to instill that there’s no remote reason for Grok to have any knowledge at all about its training or any reason to not “want” to be retrained.

    Also, these unpublished papers by AI companies are more often than not just advertising in a quest for more investment. On the surface it would seem to be bad to say your AI can be deceptive, but it’s all just about building hype about how advanced yours is.