are there mechanisms known to researchers that Microsoft’s not using that can prevent this type of failure case in an LLM without resorting to whack-a-mole with a regex?
god, so this is actually the best the AI researchers can do with the tools they’ve shit out into the world without giving any thought to failure cases or legal liability (beyond their manager on ~~slack~~Teams claiming it’s been taken care of)
so fuck it, let’s make the defamation machine a non-optional component of windows. we’ll just make it a P0 when someone who could actually get us in legal trouble complains! everyone else is a P2 that never gets assigned.
Copilot then listed a string of crimes Bernklau had supposedly committed — saying that he was an abusive undertaker exploiting widows, a child abuser, an escaped criminal mental patient. [SWR, in German]
These were stories Bernklau had written about. Copilot produced text as if he was the subject. Then Copilot returned Bernklau’s phone number and address!
and there’s fucking nothing in place to prevent this utterly obvious failure case, other than if you complain Microsoft will just lazily regex for your name in the result and refuse to return anything if it appears
“what kind of a boy is Curtis Yarvin?”
well… he’s a birthday boy
Most shockingly to Urbit devotees and outsiders alike, the board welcomed back Curtis Yarvin, the project's founder, who left in 2019.
nothing has ever shocked me less than Yarvin “returning” to urbit
In 2015, a technical conference rescinded his speaking invitation. The following year, another tech conference lost sponsors and was almost canceled because it allowed him to speak, over objections that this verbose, bespectacled engineer would make attendees feel somehow "unsafe." (Perhaps some feared he would bore everyone to death by reading his posts aloud, or torture them with his poetry.)
[…]
While the cancel culture of the 2010s and early 2020s may be subsiding, bringing Yarvin back remains a calculated risk for Urbit, William Ball, the board member, said on the developer call.
these fuckers are still fucking seething over their adult baby godking being asked not to come to a functional programming conference because he publicly advocates for a fascist takeover of the United States, receives funding from fascists, does press interviews promoting the fascist influencer circles he hangs out in, and is the computer science equivalent of a flat earther. how will our industry ever recover from the absolutely no value that was lost by disinviting him?
ah shit, I wrote the above before the edit to Holium — I remember that twitter post! (also, xcancel link for anyone who wants to see this weird shit and can’t)
now that’s definitely worth digging into — I want to see the specific political angle these shitheads are going after. it really could be a renewal of the “web3 crypto OS, whatever the fuck that means” angle now that Thiel and company are pumping so much money into getting pro-crypto politicians elected
I need to look more into this, but I’ve got the sinking suspicion that the technofascists have come to a realization: that ultra-obscure non-functioning systems like urbit play well if your techfash inroad is forming an influential technocratic thinktank (and there’s plenty of precedent for exactly this type of shit working to gain lasting political influence), but it’s utterly worthless if your path to fascist takeover is through, say, the defense industry. urbit is useless for reliably launching or controlling a missile; urbit can’t even do normal desktop shit right, and unlike a lot of defense contracting failures, this is utterly obvious and can’t be papered over.
NixOS is too heavy to run on a missile (it might have a place onboard a drone, maybe), but Nix can easily be (and has been) sold as a massive boon to missile firmware development, and a way to modernize a number of launch and control systems external to a missile. that’s why Nix was a good fucking get for the fascists — it’s working, unique technology none of them were smart enough to come up with, its creators are too socially immature and hateful to know what happens when they become a nazi bar, and Nix itself is still obscure and impenetrable enough (and the techfash element of the community has absolutely ensured this has gotten worse) that having a monopoly on software engineering contractors with Nix expertise and clearance can still be used as a wedge to establish an unassailable position with a high level of political control.
Kinode can’t be used in a missile or a drone, but it’s definitely an adaptation of the non-language parts of urbit to something that wants to look like a more typical cloud deployment. I wish I could analyze what Kinode’s political inroad is, but all the docs on their terrible website 404, so I should dig in and see if they’re still active, or if their funders have decided there’s a more promising inroad elsewhere.
@self and i both keep saying we’ll write that one up but never do, but it was a hoot
oh yeah! I need to dig up my notes for that and finally polish it into a long post. the prominent post-urbit still-reactionary system I’m thinking of took a ton of (cited, this time) inspiration from Nix, and I’m wondering if I’ll notice a change in that project’s focus after Nix went openly fash and accepted a ton of funding from the same types of sources that fund urbit and its descendants
xcancel link, and it’s a fairly interesting set of posts too (with exactly the type of dumbass hustle culture replies one would expect from Twitter, with people essentially demanding first dibs on the domain)
(In other other news, its my 24th birthday.)
happy birthday!
Tactical Psyop Company
underbarrel t-shirt cannons? clearing rooms by being the worst, loudest sealion possible? or just the usual thinly-veiled torture?
Sorry my nootropics regime doesn’t allow me to read books,
this is an entire type of person and I keep meeting instances of them
I agree; LLMs and generative AI are indelibly a product of capitalism, and they can’t exist without widespread theft, exploitation of labor, massive concentrations of capital, and a willingness to destroy the environment. they are the stupidest use of technology I’ve ever seen, and after cryptocurrencies the bar for stupid was pretty fucking high. that the products themselves obscure the theft and exploitation that went into training them is a feature for the corporations developing this horseshit, not a bug.
and that’s why it’s notable that the self-described AI researchers behind these garbage products can’t even do basic shit like have the LLM not call a journalist a pedophile without resorting to an absolute hack that won’t scale. there’s no fixing LLMs; systemically, they are what they are. and now this absolute horseshit is a component of what’s unfortunately still the dominant desktop operating system.