how is this digital treatment similar to BetterHelp? how would it possibly be bad in factors other than efficacy, like BetterHelp was due to data nightmares and advertising a different mechanism? this isn’t even online

99 is a more than enough sample size if your RCT’s Bayes factor is 114 and 15.8 for better efficacy than -control and -regular treatment respectively, which corresponds to “extreme” and “strong evidence” (Lee and Wagenmakers 2013, p. 105; adjusted from Jeffreys, 1961). The Lancet also peer-reviewed the claim “The Bayesian adaptive trial design enabled efficient evaluation with early stopping when convincing evidence was reached (n=99).[2]”

indeed further testing is needed to establish subgroup effects and improve generalizability but this is already quite promising

how so? this is a method that would work completely offline and without any form of centralization i can imagine

my point is not to ignore the machete, but to not assume the monkey

when we stupidize our enemies, we lose sight of how to fight them

call me optimistic but i think he’s just having a laugh

he’s a psychiatrist, involved in political discourse but not the kind of thing that gets editorial review. i’d gauge him as much less crank-y but only a bit more reliable than Curtis Yavin, whom he is in roughly similar Bay Area intellectual circles with (and on the opposite ideological side of)

You shouldn’t be posting that without the context in that comment:

FYI, the article was presumably taken down because many of the quotes turned out to have been fabricated, and they said they were investigating this. (I don’t think that they are trying to cover up anything, just that they have not gotten around to written an official response yet, given that this is a recent development.)

Ugh, that is utterly disappointing to see from Ars Technica. Here’s a bit of context about it: https://mastodon.social/@nikclayton/116065459933532659

Fortunately, the article was already archived, for what it’s worth: https://web.archive.org/web/20260213194851/https://arstechnica.com/ai/2026/02/after-a-routine-code-rejection-an-ai-agent-published-a-hit-piece-on-someone-by-name/

@[email protected]

I agree the dispute is stupid, but IMO the more important part here is anrchive.today’s undisclosed execution of malware to try and win a dispute.

34 (mainly non–core Anglosphere newspapers) of the 121 platforms TWL can give you access to require an application. The rest you can access automatically, instantaneously right now as long as you meet the stats.

I mentioned that this (only) solves one (of two) major problems archive.today was used to solve: paywalls. This is also very workable; you already have major newspapers like Haaretz and WSJ available on TWL.

I also mentioned that the backcatalogue problem can be solved by running a different archiving service on the existing archive.today URLs we use.

I am an active editor lol. I’m saying that the proposal is to establish something similar to TWL for media URLs. It would serve the same purpose for editors as a major complaint in the discussion was over addition of Archive.today links to bypass paywalls. Obviously developing this deal would take a lot of work but it is workable.

You must first apply to gain access.

That’s not true. Anyone who meets the stats you mentioned may access TWL.

the WML does not host any of these publications

Indeed, that’s what makes it legally sound and prevents us from needing to relicense. We don’t need to license the content to copyleft for the thing to work.

Archived pages wouldn’t necessarily be the knowledge they distribute, just ways to verify the knowledge they distribute is correct. Content from The Wikipedia Library (which provides access to academia) isn’t relicensed at all, for example. Such a service would be a project but not a sister project like Wikisource is,

https://en.wiktionary.org/wiki/jarhead

1. (slang) A US marine.

The Wikimedia project gets to host verbatim third-party news articles? This is creative but completely unrealistic

It would be just like the extant https://en.wikipedia.org/wiki/Wikipedia:The_Wikipedia_Library.

In the worst case we could just run Megalodon on all the archive.today URLs

No, you do not need real time chat, you need searchable, permanent presence

why not both? Discourse is really nice but it’s a forum, not IRC or Matrix

I disagree, but we’ll see what they do with it

https://support.discord.com/hc/en-us/articles/1500005292701-How-to-Access-an-Age-Restricted-Server-FAQ clarifies that Discord would only do that for servers that otherwise violate the Community Guidelines’s provisions on adult content, whose only mention of which I quoted above.

Server owners must apply an age-restricted label to any channels that contain sexually explicit content involving adults or content involving adults that is shared solely for the purposes of sexual gratification.

nothing else needs to be age-restricted, and age restriction of topics you mentioned would be big news for discord. pluralkit’s home is discord and has great influence there, and the lgbtq+ community definitely has a larger presence than the plural community

only changes for the unverified:

• Channels, servers, and commands set to age-restricted will be inaccessible
• DM requests deemed “sensitive” will be inaccessible
• “Sensitive content” will be blurred
• Won’t be able to become a speaker in “stage” channels, a special type where a few videoconferencers have a meeting where anyone else can watch and type but not videoconference; verified watchers can be invited to speak
• DMs from unfamiliar users will be routed to a separate “warning”

nothing else

I think this is business as usual…

ruling that California’s new law was unconstitutional because it did not impose the same requirements on all federal, state and local law enforcement officers.

However, the judge allowed the state to enforce a separate law that requires all law enforcement officers, including federal ones, to display visible identification.

that’s more than one bit, though, which is part of the question’s constraints. In practice that means [No response] would be the same as indicating 0 or even 1 and they cannot convey additional information by controlling for time of receipt, likely because they can’t control time of receipt