Linkblog | Google apologizes for ‘missing the mark’ after Gemini generated racially diverse Nazis

from the are-we-the-baddies? dept

What we currently call “AI” isn’t actually intelligence; it’s just pattern-matching. A colleague of mine calls it “spicy autocomplete.” Unfortunately the rapid development of this spicy autocomplete has led to an expectation of intelligence that doesn’t actually exist.

Take this article, for example. It’s pretty clear what happened: in an attempt to address the issue of systemic racism in AI, Google has added in some sort of “prompt salting” (there’s probably an actual terme d’art for this, but I don’t know what it is and I like that this hearkens back to “spicy autocomplete”) that automatically appends “…and don’t make them all white” to prompts asking for pictures of people.

It’s the naive solution, but it also probably solves the problem decently well for most requests.

Still, the AI doesn’t know what a “Founding Father” is, any more than it knows what a hand is. The LLM doesn’t even know what a “picture” is; it just knows that there’s a strong pattern of the word “picture” being associated with some other format it can’t read. And the image generator doesn’t know what “diverse” means, it just knows that the neurons matching pictures of non-white folks should light up when it sees that word. Both of them could probably tell you a pretty convincing fiction or draw you a pretty convincing picture about their understanding of both of those concepts, but only because we’ve drawn diagrams and written definitions that they can scrape and pattern-match.

And so, when it tries to generate a picture of “German soldiers from 1943,” the LLM (or some dumber algorithm) recognizes a pattern that it determines is asking for a picture of a group of people and appends “…and don’t make them all white.” The image generator, not knowing any better, dutifully does so. It doesn’t know anything about revisionist history or whataboutism, it just pattern matches a photo of Nazis with the added instruction of including diversity.

All of this is why AIs, while impressive and even convincing, are still not a replacement for humans; they don’t have any true understanding of context. Just educated guesses and pattern matching. And since the internet is the source of those patterns…well. Let’s just call this a reminder that it isn’t ready for anything more than helping us out with simple tasks.

Linkblog | Jon Stewart Tackles The Biden-Trump Rematch That Nobody Wants | The Daily Show

from the importunate-widow desk

Note: moderate language advisory for this video

Jon Stewart Tackles The Biden-Trump Rematch That Nobody Wants | The Daily Show
Jon Stewart kicks off his Daily Show Monday night residency by coming to grips with the reality of America's two chronologically challenged presidential cand...

Jon Stewart is back on The Daily Show. For one day a week, at least. His first show was this past Monday, and he came out swinging with a piece about the upcoming presumptive election battle between Joe Biden and Donald Trump. The whole piece is good (it’s worth asking why we have to choose between two very old men of whom the question of capacity at least has to be asked, when there are plenty of capable and qualified people under the age of 50 who could be on the ballot).

But his closing words are a tour de force.

So what’s the good news? …That was not rhetorical framing, I’m literally asking you.

Look, the next nine months or so (and more than that, depending on the coup schedule)— they’re gonna suck […] it’s all gonna make you feel like Tuesday, November 5th is the only day that matters. And that day does matter, but man, November 6th ain’t nothin’ to sneeze at, or November 7th.

If your guy loses, bad things might happen, but the country is not over. And if your guy wins, the country is in no way saved. I’ve learned one thing over these last nine years [note: since his first run on The Daily Show ended]—and I was glib at best, and probably dismissive at worst about this: the work of making this world resemble one that you would prefer to live in is a lunch-pail [bleep] job, day in and day out, where thousands of committed, anonymous, smart, and dedicated people bang on closed doors and pick up those that are fallen and grind away at issues ’til they get a positive result—and even then have to stay on to make sure that result holds.

So the good news is, I’m not saying you don’t have to worry about who wins the election. I’m saying you have to worry about every day before it and every day after, for-ever.

Although, on the plus side, I am told that at some point, the sun will run out of hydrogen.

If you’re a Christian like me, please resist the urge to mentally tone-police or redraft Stewart’s statement here. Yes, the word “worry” is a little bit outside our theological preference. Yes, the nihilism of the sun’s eventual demise is far from the actual hope we have in the future. Yes, the Lord is sovereign over our elections.

But the message he’s sending here comports with Christian theology, as well as probably every theological framework people in our world subscribe to. Namely: that the fate of the world isn’t decided on a single day, and the outcome of even the most fateful days is less important than that of the days that precede it and follow it; and that our role in the meantime is to keep trying to make this world resemble the one that you would want to live in (Jeremiah 29:7), to pick up those who are fallen (Matthew 10:7-8), and to grind away at issues ’til you get a positive result (Luke 18:1-8).

And Stewart is framing this as good news! Because it’s funny, yes—but also because it brings things to within our sphere of influence. We can affect the way this world works, and in fact we should. We have the ability and opportunity to make changes.

So let’s do that.

Linkblog | Cory Doctorow: How I got scammed

from the mothers-maiden-name desk

Cory Doctorow regales us with his story of being phished, being scammed, being almost scammed, and being almost phished again, and uses his $8,000 fleecing as an opportunity to teach us about swiss cheese security, when everything lines up perfectly for a scammer to get through.

It makes me think about a Hollywood heist film, where George Clooney waits patiently for the exact right amount of money to be in the vault, for the people who are responsible for its protection to be distracted or harried or frustrated in a specific way, for the business interests of the target to be specifically aligned away from what he wants to go after; so that when he makes his move he only has to actually put his people in a particular place and let them go through the holes that have already opened up.

Good reminders for individuals and for companies who rely on strong security to maintain their business or reputation.

Linkblog | Ava’s Demon

from the Tuls-and-Nevy-shipping dept

Ava’s Demon

Ava’s Demon is a web comic about a girl named Ava and the demon haunting her.

The spectacular sci-fi webcomic Ava’s Demon first hit the internet in 2012, and quickly made a name for itself with its lush, luminous painted panels, its occasional splashes of animation, and its cinematic one-panel-at-a-time cadence. Written and drawn by Michelle Fus, an animator formerly of Pixar and DreamWorks, it’s probably still the most beautiful webcomic online.

Until recently, the comic updated more or less weekly, dropping a handful of new panels about Ava, her titular “demon” tormentor Wrathia, as well as Ava’s…”friends” Maggie, Odin, and Gil; but despite putting in twelve years and over 2,500 pages of kinetic action, Ava and co. have only just finished up their first action-packed day as they run from the forces of the intergalactic god/despot/CEO Titan.

I lost touch with the comic by about 2017, when keeping up on webcomics proved to be somewhat incompatible with parenting multiple young children; and since then, Ava’s Demon has continued its steady drumbeat. Right up until November of 2022, when after a decade of fairly regular uploading, the comic just…stopped.

Webcomics lapse. They rarely come back. But Ava’s Demon did.

It lapsed for a good reason…and came back for a better one. If you haven’t read the comic before, I recommend you start at the beginning; but if you’re unconvinced, take a look at these incredible pieces of art, and imagine how much better they are in context.

Linkblog | Ars Technica: The amazing helicopter on Mars, Ingenuity, will fly no more

from the red-planet-whirlybird desk

The amazing helicopter on Mars, Ingenuity, will fly no more
Ingenuity has spent more than two hours flying above Mars since April 2021.

Well done, Ingenuity. The Martian helicopter—well, okay, it’s an Earth helicopter on Mars—flew its final mission this month, breaking one of its rotors during its 72nd out of five planned flights in the thin Martian atmosphere. The four-pound aircraft flew a total of 11 miles (17 kilometers) across the Martian landscape; by contrast, the first wheeled rover on Mars, Sojourner, only traveled about 330 feet (100 meters) before we lost contact with it.

It was originally planned to fly between 1 and 5 times; it flew 72 times. Its maximum operational duration was 30 sols (Martian days), but it made its final landing 977 sols after its first takeoff. Its design expected a maximum altitude of 12 meters (39 feet), but it doubled that expectation, traveling as high as 24 meters (79 feet). Having exceeded nearly every design parameter, I think it’s earned its rest.

The whole mission was impressive, but if you’re interested in the design of this incredible aircraft and spacecraft, you might enjoy this video from the YouTube channel Veritasium, made over a year before Ingenuity even reached Mars:

Congratulations to NASA, the JPL, and everyone who worked on this impressive piece of technology.

Linkblog | The Verge: I’m sorry, but I cannot fulfill this request as it goes against OpenAI use policy

from the the-tech-isn’t-the-problem-the-implementation-is-the-problem desk

I’m sorry, but I cannot fulfill this request as it goes against OpenAI use policy
“Our [product] can be used for a variety of tasks, such [task 1], [task 2], and [task 3], making it a versatile addition to your household.”

The Verge posted this delightful article about how…uh…we’ll call them “ethically-challenged” companies have shown their hand on Amazon by wiring ChatGPT directly up to Amazon and telling it to generate whole product ideas for them.

Companies hiding behind multiple nonsense names has been a problem on Amazon for a while, since listings for multiple copies of the same item are cheap and can create the illusion of competition; which is why you’ll see Amazon listings for NKNZLY, JXMOX, BKAYP, MUXA, IANOPE, and USAMS all selling the same charging cable. They’re probably all the same company. It’s not technically a *scam*—you’ll probably get your product, and it might even work!—but it also isn’t completely above-board either. With such a public screw-up, though, confidence in both Amazon and unknown sellers is sure to go down, which can only mean one thing: Amazon is doomed.

Haha, no, of course not. No, they’ll just update their policies to restrict or prevent AI listings, and begin the moderation whack-a-mole, while the sellers find new and exciting ways to trick the system, and legitimate sellers get caught up in the game as false positives.