Miscellany

[BrotherMan]

That really hits me in the feels, y'all.

[Ymir's blood]

Originally Posted by Stormlight

wah wah wah wah!

Yes, Mr. Stormlight.

[Kamilah Hauptmann]

I don't know where else to put this.

Deep Empathy

[mickthinks]

... that doesn't look like AI ("teaching" neural networks is not AI) nor does it look at all likely to achieve the goal of making those lacking empathy when seeing images of bombed out Syrian cities feel more empathic. So it's basically bullshit, presumably trying to sell us something.

[Kamilah Hauptmann]

¯\_(ツ)_/¯

[lisarea]

It's only not AI in the sense that AI is a moving target, but not long ago, most projects universally considered AI were built pretty much like that. So I would argue that, while it's not state of the art AI, it could still be classified as AI.

[ceptimus]

Some people like to define AI as, "Something that humans can still do reasonably well, but computers can't do yet."

[Kamilah Hauptmann]

Increasingly sophisticated autocorrect schemes means a shift from decipherable typos to alien turns of phrase.

iPhone just changed what I was saying to “concubine enamel”. I’ve lost any memory I once had of what I was trying to type.

— Tim Maly (@doingitwrong) December 27, 2017

[Watser?]

I can think of no situation where that particular combination would be useful.

[Kamilah Hauptmann]

I’ve learned to never underestimate fetishes or names for punk bands.

[Watser?]

Punk band yes, but enamel concubine would be so much better.

[Sock Puppet]

Sounds moar like a drag queen troupe.

[Kael]

I must play too much exactly the right amount of D&D - my first thought was an obscure alchemical ingredient.

[ceptimus]

[Kamilah Hauptmann]

[Kamilah Hauptmann]

[Crumb]

[Kamilah Hauptmann]

Works on two devices with different viewers on two different networks. :dontknowwhattosay:

[Kamilah Hauptmann]

[Stormlight]

Doncaster's new gritters named David Plowie and Gritsy Bitsy - BBC News

Two road gritters have been named Gritsy Bitsy Teeny Weeny Yellow Anti-Slip Machiney and David Plowie in a council-run online poll that saw a flurry of witty suggestions.

The council currently has five named gritter vehicles: Brad Grit, Gritney Spears, The Subzero Hero, Mr Plow and Usain Salt.

[Ymir's blood]

Those Brits and their cutesy names for salt trucks.

[ceptimus]

Allegedly the output of a bot fed on a diet of Harry Potter books.

Botnik Studios

I'm suspicious that it may have had a little human assistance. If not then these bots are getting rather good.

[specious_reasons]

Originally Posted by ceptimus

Allegedly the output of a bot fed on a diet of Harry Potter books.

Botnik Studios

I'm suspicious that it may have had a little human assistance. If not then these bots are getting rather good.

I'm pretty sure the bots had help. If it's anything like the rest of the website, the predictive text gives choices, and a person picks the choices for maximum effect.

[erimir]

Predictive text tends to have a short window. The main use case for predictive text is things like text messages and Twitter, so you're not expecting someone to write a whole book chapter with it, but mostly short text segments. Usually it's something like a trigram model, since it's not meant to write a whole sentence for you (bigram models used to be more common, but I think they've all gone beyond that now). Sometimes it can suggest multiple words, so I guess it may extend to a 4-gram or 5-gram model. It doesn't need to be longer because it merely suggests words with the expectation that you will very frequently not use them. A word being the most likely next word doesn't mean it's likely - most likely in this context often will only mean 5%... Even "the" only follows "of" about a quarter of the time, and that's because those are both function words. Individual content words are far less frequent than function words. Anyway, whenever you pick a different word, the previously generated suggestions are useless.

So those reasons are why you don't need to bother using a very long window for "predictive text".

And the thing about the window is, once you're outside the window, it has no memory of what was written. Because of this, it's very repetitive. There's no reason why it wouldn't suggest the same words in one sentence starting with "the" that it did in another sentence starting with "the", although you could introduce some randomness to counteract this. And since it has no knowledge of what was written before, it won't tend to be very coherent. The subject of one sentence won't have any connection to the previous sentence.

You could counter this by having the window cross sentence boundaries, but you'd still need a long enough window to keep the topics coherent, and at that point, your window might be so long that you start repeating sentences from the Harry Potter books wholesale. And either way, a simple predictive text model can't manage coreference - it cannot remember who is referred to by "he" or "she" because such information is not built into the model. If it were built into the model, it wouldn't be accurate, IMO, to call it "predictive text" anymore.

But anyway, that writing sample is not very repetitive, is too topically coherent, and doesn't seem to have pronouns that are difficult or impossible to resolve. All of which suggests it was not written purely with an n-gram language model aka predictive text.

And, most importantly, the authors say they "collaborate" with machines in the about section

[Kyuss Apollo]

Historical generals pointing out the toilets: a short thread.

Historical generals pointing out the toilets: a short thread. pic.twitter.com/WYzsnRlszj

— Jonathan Healey (@SocialHistoryOx) January 5, 2018