Gordon Mohr ꧁🧿🧿꧂

one way to pave the way for a generatively-faked POTUS is to make sure any preceding real vids already low-liveliness
bsky.app/profile/segy...

all the classic sci-fi tropes where supercomputer hyperliteralsm is funny, or frustrating, or responsible for its ultimate defeat are obsolete.

now – but harder to portray – supercomputers will struggle to communicate all subtleties & caveats that they richly understand to oversimplifying humans.

not saying I've achieved this, but I believe those who reliably lift have somehow managed to make its effects on their body/strength/mood just as 'tangible', mentally, as those externally-legible "landscaping, building, moving"-type goals you mention.

This reverse Turing Test video is one of the more unsettling pieces of AI content I've seen. The perspective of being an Other in which their mannerisms are normal in an embodied setting is quite spooky.
youtu.be/0MmIZLTMHUw

are you implying we're some other ancient civilization's x-disaster grey goo?

or what if it's not the *accounts* with popularity-regularization, but the *posts*. limited-edition skeets, only viewable by X thousand people at a time – as if they were being checked out & returned to a library.

www.msnbc.com/opinion/anal...

Center for the Alignment of AI Alignment Centers We align the aligners

Q. Who aligns the aligners?
A. alignmentalignment.ai

Today I’m humbled to announce an epoch-defining event: the launch of the 𝗖𝗲𝗻𝘁𝗲𝗿 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗔𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁 𝗼𝗳 𝗔𝗜 𝗔𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁 𝗖𝗲𝗻𝘁𝗲𝗿𝘀.

despite its problems, Grok is often better than the average Twitter (or for that matter Bluesky) user in weighing evidence. so if it can't navigate the feed frontier, what hope do modal human users have?

return to Xanadu

in a proper functioning democracy, every election would have a guy who wears the same red beret literally everywhere, to the point where nobody knows what he looks like without it. he should be referred to in the constitution as Hat Guy. if you ever poll below Hat Guy you must end your campaign

thanks! oddly NYT's native search doesn't find this on various combos of [yudkowsky book], [yudkowsky soares], [build it everyone dies yudkowsky], etc

perhaps, NYT's indexing agent has decided against it being found. :)

is an NYT review available somewhere? who's the reviewer?

Baudrillardian hyperreality. you've (quite reasonably) sensed Baudrillardian hyperreality.

Venn diagram titled “Shit-posting & Shit-posturing.” The left circle is labeled “Shit-posting,” the right circle is labeled “Shit-posturing,” and the overlapping center is labeled in bold “NO VACANCY.” The overlap is larger than the non-overlapping areas.

shit-posting xor shit-posturing

…reporting automated CSAM detection/blocking led to no credit for proactivity, but instead new grandstanding about "soaring" problem possibly just improved detection or even faked activity, by those seeking to probe oracles, chaff-clog detectors, frame innocents, or manufacture casus regulation. /

…wrt Backpage, even before FOSTA/SESTA, their cooperation with law enforcement twisted to attack them as more "knowing". their deployment of automated blocks against known-coded langauge was criticized as coaching bad actors: damned if they did, damned if they didn't.

wrt Facebook et al, …

& more generally, the law's treatment of how platforms handle suspect-but-not-prosecutable behavior often resembles the no-win "double-binds" used by cults & psychological abusers.

eg: …

Zucker, Abrahams and Zucker bring you: "Vantablack Mirror!"

savoir commerce, the cousin of doux commerce

Meta is killing off its own AI-powered Instagram and Facebook profiles Instagram profile of ‘proud Black queer momma’, created by Meta, said her development team included no Black people

…
so in the nearish future, or for higher political/economic stakes, even an internet-savvy investigator could be fooled.

I mean, it's not like Meta was so reckless as to fill their platforms with thinly-disclosed AI-generated fake friends last year, right?
www.theguardian.com/technology/2... /

good to know other factors you've considered, & – as I mentioned – given Sep 2025 assumptions (plus low stakes here), pretty convincing.

but it's gonna get very cheap & easy to create everything you checked – including backdated if done by platform insiders, or one-person/one-click automated.
…

…
quite soon, pranksters or marketers will find it easy to render all the subtly-varying scraps about such an event, self-consistently, & syndicate through many platforms/accounts to create a false impression of diverse cross-reinforcing sources.

our sensemaking habits & orgs are not ready. /

…
generative capabilities race forward. Smith's team thought nothing of casually confabulating promo video.

afaict, all your A/V sources come via Meta Instagram & Google YouTube – conglomerates with leading generative video offerings, & services/partnerships with professional photogs & journos.
…

does an email interview provide extra marginal confirmation they're real, beyond prior work?

your original correlation of multiple photo sources, & plausible conjecture of how real photos fed video slop, give me high (but not 100%!) confidence - under September 2025 standing assumptions.

but, …

…
Signal could've let users consciously opt their data into native backup mechanisms: plenty of other personal media & info thusly backed-up is of same/greater sensitivity as Signal comms.

or, similar app-private exportable "recovery key" could've secured user-local backups at any time. /

…
until such user control someday delivered, still feels like Signal has made conscious 'enshittification' design choices, holding our messaging-history durability hostage, to prep for their upcoming recurring-revenue subscription service.

lost/broken phones never had to mean lost history!
…

(corrected reaction)
most interesting part of this announcement is statement "future plans include letting you save a secure backup archive to the location of your choosing".

that's long overdue - it's been 10y since users 1st told "import/export", enabling user backup, was "on the roadmap"!
…

the optimal amount of cultishness is not 0%

but, it is typically closer to 0% than to 100%

close-up of a detailed shiny-paper seahorse, held between a thumb & forefinger

"Why of course, the symbolic origami in 'Blade Runner' (1984) was a seahorse."