eclecticca:

mostlysignssomeportents:

mostlysignssomeportents:

Humans are not perfectly vigilant

ALT

I’m on tour with my new, nationally bestselling novel The Bezzle! Catch me in BOSTON with Randall “XKCD” Munroe (Apr 11), then PROVIDENCE (Apr 12), and beyond!

ALT

Here’s a fun AI story: a security researcher noticed that large companies’ AI-authored source-code repeatedly referenced a nonexistent library (an AI “hallucination”), so he created a (defanged) malicious library with that name and uploaded it, and thousands of developers automatically downloaded and incorporated it as they compiled the code:

https://www.theregister.com/2024/03/28/ai_bots_hallucinate_software_packages/

These “hallucinations” are a stubbornly persistent feature of large language models, because these models only give the illusion of understanding; in reality, they are just sophisticated forms of autocomplete, drawing on huge databases to make shrewd (but reliably fallible) guesses about which word comes next:

https://dl.acm.org/doi/10.1145/3442188.3445922

Guessing the next word without understanding the meaning of the resulting sentence makes unsupervised LLMs unsuitable for high-stakes tasks. The whole AI bubble is based on convincing investors that one or more of the following is true:

1. There are low-stakes, high-value tasks that will recoup the massive costs of AI training and operation;
2. There are high-stakes, high-value tasks that can be made cheaper by adding an AI to a human operator;
3. Adding more training data to an AI will make it stop hallucinating, so that it can take over high-stakes, high-value tasks without a “human in the loop.”

These are dubious propositions. There’s a universe of low-stakes, low-value tasks – political disinformation, spam, fraud, academic cheating, nonconsensual porn, dialog for video-game NPCs – but none of them seem likely to generate enough revenue for AI companies to justify the billions spent on models, nor the trillions in valuation attributed to AI companies:

https://locusmag.com/2023/12/commentary-cory-doctorow-what-kind-of-bubble-is-ai/

The proposition that increasing training data will decrease hallucinations is hotly contested among AI practitioners. I confess that I don’t know enough about AI to evaluate opposing sides’ claims, but even if you stipulate that adding lots of human-generated training data will make the software a better guesser, there’s a serious problem. All those low-value, low-stakes applications are flooding the internet with botshit. After all, the one thing AI is unarguably very good at is producing bullshit at scale. As the web becomes an anaerobic lagoon for botshit, the quantum of human-generated “content” in any internet core sample is dwindling to homeopathic levels:

https://pluralistic.net/2024/03/14/inhuman-centipede/#enshittibottification

This means that adding another order of magnitude more training data to AI won’t just add massive computational expense – the data will be many orders of magnitude more expensive to acquire, even without factoring in the additional liability arising from new legal theories about scraping:

https://pluralistic.net/2023/09/17/how-to-think-about-scraping/

That leaves us with “humans in the loop” – the idea that an AI’s business model is selling software to businesses that will pair it with human operators who will closely scrutinize the code’s guesses. There’s a version of this that sounds plausible – the one in which the human operator is in charge, and the AI acts as an eternally vigilant “sanity check” on the human’s activities.

For example, my car has a system that notices when I activate my blinker while there’s another car in my blind-spot. I’m pretty consistent about checking my blind spot, but I’m also a fallible human and there’ve been a couple times where the alert saved me from making a potentially dangerous maneuver. As disciplined as I am, I’m also sometimes forgetful about turning off lights, or waking up in time for work, or remembering someone’s phone number (or birthday). I like having an automated system that does the robotically perfect trick of never forgetting something important.

There’s a name for this in automation circles: a “centaur.” I’m the human head, and I’ve fused with a powerful robot body that supports me, doing things that humans are innately bad at.

That’s the good kind of automation, and we all benefit from it. But it only takes a small twist to turn this good automation into a nightmare. I’m speaking here of the reverse-centaur: automation in which the computer is in charge, bossing a human around so it can get its job done. Think of Amazon warehouse workers, who wear haptic bracelets and are continuously observed by AI cameras as autonomous shelves shuttle in front of them and demand that they pick and pack items at a pace that destroys their bodies and drives them mad:

https://pluralistic.net/2022/04/17/revenge-of-the-chickenized-reverse-centaurs/

Keep reading

Gonna go watch Modern Times now…

(via mostlysignssomeportents)

deeisace:

deeisace:

neil-gaiman:

In the olden days they did things so sensibly. Page 8 of The Liverpool Daily Post, 29 March 1937

ALT

ID:

Barmaid Becomes Barman

Girl Re-Registered

The re-registration of a seventeen-years-old Ellen Caldwell, a Crewe barmaid, residing at Martin Street, as a male, and her appearance at the hotel where she is employed smartly attired in man’s clothing, has brought to a conclusion a remarkable episode of transformation.

Last summer, she sought medical advice, and an operation was performed at Manchester Royal Infirmary. The operation was successful, and her father has taken the necessary steps to have her re-registered, this time as a male.

She has changed her name to Alan, and Ellen the barmaid is now Alan the barman.

/End ID.

I was worried for a second, but it turns out I was looking in the wrong county -

Here he is!

Part of the 1939 register, a census showing the Caldwell family living at 46 Martin Street, Crewe.

John Caldwell, head of the family, a 54 year old general labourer. His wife, Edith, also 54 years old and listed as “unpaid domestic duties” - a housewife. Alan, spelled Allen here, is 20 (birthday 12th May), and working as a garage hand. A closed record - either someone who is still alive, and/or someone who was a small child at the time of the census, or,, idk some other reason, but it’s usually those. Last is John Junior, who is 12 and still at school.

It turns out he stuck with the spelling of Allen, because here look -

He got married! Spring 1941, to a Miss Sylvia Copp.

I can’t see the details without ordering the record itself, and I don’t hugely have that money spare atm, but I hope they had a lovely day of it, even if I can’t see who were their witnesses

coraliewhodraw:

Me making an AU? Haha imagine that

No, but really, I think Martha deserves to have a good team and go on scooby doo-esque adventures. I’m just a sucker for the found family trope, sorry.

So here is the team : Doctor Martha Jones, Tallulah, Hath Peck, and Captain Jack Harkness! Maybe others will join? Who knows?

coraliewhodraw:

Starting the year with my wife. My poor girl cannot have a second to study. The world is always ending :((

Btw while I was trying to find what kind of phone she had in the show to be accurate I stumbled on a blog listing every type of phone in the show and it’s so fascinating to me, what an icon

(via burntlikethesun)

cwalshuk:

doctorwho:

Video: BFI ‘Doctor Who: An Unearthly Child’ Event

with Waris Hussein, William Russell, Carole Ann Ford, Mark Gatiss, and more.

I was lucky enough to be at this event and, having never seen An Unearthly Child, really enjoyed both watching it and hearing from the people who made it.

Thank you BBC and BFI!

Do you have any more video of the Q&As ?

Also, great write-up by a much more knowledgeable Whovian, Patrick Mulkern, here http://www.radiotimes.com/news/2013-01-14/the-bfis-doctor-who-at-50

Eleven years ago today!

I didn’t realise I’d been here that long!

I ought to return properly, if only for all the new Doctor Who fanworks.

Here’s another video from that event, with some of the stars of Doctor Who in 1963: https://youtu.be/_iSA7PRs_fU?si=38Oe__vdwc2hhXKa

And one from a later event about the drama about making Doctor Who in 1963:

https://youtu.be/vDaoR81b0DI?si=vbGcLwKrjmrG7IK1

(Source: youtube.com)

cardboardlife:

Crochet cat!

kiran-wears-science-blues:

thatwhichresemblesthegrave:

smolbeerke:

redvelvetrevolver:

thanook:

inthesensethat-deactivated20240:

Incredible

wait so

all this time it was not normal

because all the time I have something playing in my head

(Most recently its usually Lagtrain, Loopspinner or a remix of the two)

its completely normal. 92% of people get songs stuck in their head.

#next on tiktok: did you know having THOUGHTS is ABNORMAL and a sign of BPD???

this feels like a good time to plug this study:

The Reach and Accuracy of Information on Autism on TikTok - PubMed

The informational content about autism made available on TikTok reaches a wide number of people. Most of the information provided, however,

PubMed

“An examination of the top 133 videos providing informational content on autism, which totaled 198.7 million views and 25.2 million likes, showed that 27% of the videos were classified as accurate, while 41% were classified as inaccurate and 32% as overgeneralized.”

Here is a link to a free full-text version of the article mentioned above!

(via dungeonnerd)

thelvadams:

60 YEARS IN TIME AND SPACE

Yes, it all started out as a mild curiosity in a junkyard. And now it’s turned out to be quite a… quite a great spirit of adventure, don’t you think?

lizclimo:

my first attempt at making these guys move

(original comic here)

elodieunderglass:

wild-saber1337:

MY GOD WHAT HAVE I FOUND XD

I’m not sorry for having you know this exists XD

My sibling sent this to me with no explanation it’s how we communicate