DaVince @davince21 - Tumblr Blog

Discount Strong Bad emails

Children of the early 2000s internet (hi!) may remember spending hours watching Strong Bad emails from homestarrunner.com. They are literally short videos of some guy named Strong Bad checking his email and they are hilarious.

Anyway, Strong Bad just emailed this out on the Twitter:

To find out what that would be like, I used a real neural network, the discount version of Open AI’s GPT-2, the publically-released version with 117 million parameters. Starting from its knowledge of a huge chunk of the internet, plus whatever I trained it on last (looks like that was D&D character bios), I finetuned it briefly on the full text of ~200 emails extracted from Hrwiki.org. It worked so well it’s almost scary.

The questions get Strange.

For more strange emails than would fit in this blogmail thing, go on over here to get emails in your very own email.

aiweirdness

What the machines will debate when the humans are gone

In 2018, IBM introduced “IBM Debater,” a machine learning algorithm whose goal was to argue with humans. Rather than a free-for-all argument (which the AI would win using the same techniques as a tireless two-year-old), it was trained to follow the tightly-constrained format of a debate competition. It could craft arguments on about 100 different topics by stitching together snippets of news articles. It could even listen to and (sort of) respond to the arguments of its competitor (one observer charitably said that perhaps it was ignoring its competitor’s argument as a debate tactic, not simply because it was confused). But it did not come close to winning the debate.

This House Believes That had the debate topic been chosen by an AI rather than a human, the competition would have been more fair.

Blog reader Alex Deam suggested that I train a neural net to generate topics, and pointed me toward a spreadsheet of 4,369 British Parliamentary Style debate resolutions. I fed them into a character-level instance of textgenrnn, and it did its best to imitate the human-written topics. Please enjoy these debate topics, suitable for use in the post-human era.

Are you a self-aware AI looking for more debate topics? The neural net generated more great topics than would fit in this post. To get them (and optional bonus material every time I post), enter your email here.

aiweirdness

Don’t let a neural net mix drinks.

So I’ve used neural networks to generate recipes in the past. They’re computer programs that can learn to imitate the data we give them, copying the way that humans drive cars, label images, or translate languages.

That is, they try to learn. They’re called “neural” because they have virtual neurons that work a little like the real neurons in our brains. Their virtual brains, however, are really tiny. Where a human has about 86 million neurons, the neural networks we use today have hundreds to low thousands - think nematode worm or, optimistically, jellyfish. So they often struggle, and recipes, my friends, are one of those times. (you haven’t lived until you’ve seen a neural net’s attempt at “small sandwiches”)

I turned to cocktails, inspired by Beth Skwarecki’s cocktails bot (and helped immensely by the dataset she sent me).

The neural net’s first attempt was… an attempt.

I let it train for a little longer, and things got a little more recognizable, as it learned to kinda spell more ingredients.

It had, however, not learned when to quit. 64-ounce (2 liter) cocktails are not unheard of at this stage.

It continued to learn on its own, all without input from me, and eventually ended up with cocktails that were almost plausible.

Some of my other training attempts, however, did not go so well. At first the cocktail titles in my training set were in all caps, which confused the heck out of the neural net because capital letters were so rare that it didn’t see enough examples of what to do with them. It struggled with the titles.

You may also notice that the ill-fated all-caps attempt had a lot of repetition - that’s because I was using textgenrnn, which has a somewhat limited memory. It could learn to spell ingredients, but had no idea whether it had already added sugar and cream. When I switched to char-rnn for the lowercase recipes above, I could give it a memory of 50 characters, enough to cut down on repetition. The textgenrnn version, however, became strangely obsessed with creme de cacao. The less said about its cocktails, the better.

For more neural net cocktails (including custom-generated cocktails for any name you care to provide), check out Beth’s cocktails bot!

For a few more of the cocktails I generated (including some very unfortunately-named ones), you can sign up here, and optionally get bonus material every time I post.

aiweirdness

Neural Horror Picture Show

(Images generated by BigGAN)

Neural networks are a kind of machine learning algorithm that learn to imitate the examples I give them. They’re pretty good at picking up on the feel of craft beer names vs guinea pig names, or metal bands vs my little ponies. So they should be pretty good at thinking of new names for horror movies, right? Right?

Ummm

The Baby Bunny Pretty Day 3 Little Clock Train The Sea Bothers No Curse Great Apartment The Funny Bear Missing Bun Noice Fish Christmas

The above titles were generated by a neural network (textgenrnn) trained on over 78,000 horror movie titles extracted from IMDB by Twitter user @enkiv2. It turns out that not every horror title is Slashy This or Murdled That. Titles in the original dataset include “PigDog”, “Princess Baby”, “King Dinosaur,” and “Sasquatch! A Love Story”. The neural net learned a bunch of interesting vocabulary from them.

And the neural net DID manage to pick up on the vocabulary of horror film titles. It just also managed to make them sound somehow more like the Halloween special of a kid’s cartoon show.

Halloween Bear Case of the Big Monster Dracula’s Holiday The Creepy Clock The Interbother The Haunting Song The Scream Fish Tapes Monster Interviews Monster of Mystery The Devil’s Cool World of the Band Zombies The Lucky Dracula The Christmas Man Werewolf

Other times, it was just plain weird. I know horror is supposed to make the mundane scary, but I somewhat question its methods.

Haunted Boss Final Hair The Cord Thing Day in the Room Attack of the Cleaning Day The Stall Satan The Return of Bob The Date of Horror The Snape Unknown Carp Halloween Special Planet Boot Cook The Tooth Things Stanning the Damned Halloween: Realm Wars Part 1 American Star Horror Stories: The Creature of the Santa

With other titles, the neural net avoided sounding cute, but also avoided sounding convincingly human.

A Not Meat of the Wild Fright & Trap of the Woods The Man in My Dangers I Call Zombies The Night Night Fright Footage The Bad Side of the Dead The Boobley of the Dead Please Alive Lost Blood Horror Story The Haunting of Flesh Show The Will the American Horror Stories - What the Green Horrors

I really like these titles with made-up words.

The Blunge The Crooping Tomb The Moonmanic Sands My Boblop Trap Clownfull Missing The Contains of Halloween Trailer’s Comman Reaper Trailer 2000 Scarecrowning Strangler Secret + the Basament in the Children

It also generated a few that were a bit more along the lines of, um, erotic horror. To go with those, I also decided to see what would happen if I mixed the horror titles with ice cream flavors. Here’s a taste (ahem):

Honey & Darkness Passion Owl Caramel Horror Spice

To read more (and optionally get more bonus material every time I post), enter your email here.

smitten-kitten-sarah-blog

SEVEN OF PENTACLES

“… shows that you have a strong desire to invest in those things that will provide long-term benefit. You understand the value of putting in time and energy now for longer-term rewards, and you are not necessarily looking for ‘quick wins’ as such. You also want to make sure that you are focusing your attention on the right areas that will give you the biggest ‘bang for your buck’, instead of wasting time on areas that will not deliver any value. You have a long-term view and you are focused on sustainability of results.” “You are likely to see financial or other tangible rewards for all of your hard work.“ - https://www.biddytarot.com/tarot-card-meanings/minor-arcana/suit-of-pentacles/seven-of-pentacles/

thathomestar

A Somewhat Comprehensive List of Good Old-School FPS Games, Where To Get Them, & How To Get Them Working

I love old FPS games. The action is fast, the levels are explorable, the guns are powerful. Whether you’re tired of modern FPS games or just waiting for some new hotness to come out, there’s a lot of these old games you can play! This list will cover a lot of the old classics and how to get them running on modern hardware (assuming you’re on Windows. Some of these work on Mac/Linux though).

Of course, what defines an “old-school” game from something more recent? Well the way I see it, there are FPS games before Halo, and FPS games after Halo. Halo was such a landmark game that it shifted the entire genre, hell even the entire video game industry. So I’m going to keep this list to pre-2001 games for the most part.

Keep reading

negutivfeadbak

legendsoflocalization

I’m comparing the Japanese version of Final Fantasy Vi with the SNES, GBA, fan, and Google translations all at the same time. Details here

docreggie

Something that really irks me is that whenever people see a RPG where the objective is to collect and raise monsters/robots/whatever (like Digimon, Monster Rancher, Yo-Kai Watch, Dragon Quest Monsters, Medabots, etc), their first impulse is to directly compare it with Pokemon. Or more specifically, call them a “Pokemon clone” or “rip-off.”

And I really hate it when people do this, because it devalues what the game brings to the table and undermines a valid sub-genre of RPGs.

Listen, I’m a fan of what I like to call monster raising RPGs. RPGs in which you’re not limited to a set party of characters, but literally have hundreds of monsters (or robots or whatever else) at your disposal and can customize as you see fit. And as a fan of these, it irks me whenever I hear this. Pokemon is the most popular and widely-known series of this genre, but there are many more out there, most of which never see release outside of Japan. And because most aren’t released overseas with the exception of a few, people automatically think of Pokemon when they see them and associate them as so-called “rip-offs.” Even card-based RPGs get compared when reviewers call their gameplay mechanic “Pokemon-ish.”

I get there is a strong resemblance and influence, especially since Pokemon did popularize the genre. But to call them “rip-offs” undermines the effort and innovation that went into them. I’ll be frank, nowadays I actually prefer playing Yo-Kai Watch or Telefang over Pokemon because they have unique mechanics or at least set out to play differently, whereas Pokemon has otherwise stuck to the same formula for the last 20+ years with only minimal innovation between generations. When you’ve been playing Pokemon for as long as I have, these so-called “clones” are a breath of fresh air.

And also, Pokemon itself isn’t even the origin of this genre. Robotrek/Slapstick for the SNES came before it. Heck there’s probably other early examples more obscure than that. Point is, Pokemon only popularized the genre, it’s not the originator of it. And even if it was, there’s nothing wrong with having more RPGs where the core element is to raise monsters.

Give other monster raising RPGs a chance or at least respect them for what they are.

aiweirdness

The Visual Chatbot

There is a delightful algorithm called Visual Chatbot that will answer questions about any image it sees. It’s a demo by a team of academic researchers that goes along with a recent machine learning research paper (and a challenge for anyone who’d like to improve on it), and its performance is pretty state-of-the-art, meant to demonstrate image recognition, language comprehension, and spatial awareness.

However, there are a couple of interesting things to note about this algorithm.

It was trained on a large but very specific set of images.

It is not prepared for images that aren’t like the images it saw in training.

When confused, it tends not to admit it.

Now, Visual Chatbot was indeed trained on a huge variety of images. It can answer fairly involved questions about a lot of different things, and that’s impressive. The problem is that humans are very weird, and there are still many things it’s never seen. (This turns out to be a major challenge for self-driving cars.) And given Visual Chatbot’s tendency to react to confusion by digging itself a deeper hole, this can lead to some pretty surreal interactions.

Another thing about Visual Chatbot is that most of the images it’s been trained on have something in them - a bird, a person, an animal. It may have never seen an image of just rocks, or a plain stick lying on dirt. So even if there isn’t an animal there, it will be convinced there is. This means this bot always thinks it’s on the best safari ever. (For the record, it thought the stick lying on dirt was “a bird is standing on a rock in the snow”)

If you ask it enough questions, could you get an idea of how it made its mistakes?

An algorithm that can explain itself is really useful. Algorithms make mistakes all the time, or accidentally learn the wrong thing. This particular algorithm didn’t have trouble with hallucinating sheep like some other algorithms I tested. But it did have similar problems with goats in trees, and now I finally got to ask why.

[Goat image: Fred Dunn]

Upon further questioning, however, it also decided that dogs also have horns, and birds do not fly. Actually, it turns out that a lot depends on how you ask the question. The answer to “do bunnies fly?” is “no”, but the answer to “can bunnies fly?” is “yes”, so either the algorithm is answering a lot of these questions at random, or bunnies *can* fly but choose not to. (The construction “Do <blank> have <blank>?” seems to almost always result in a “yes”, so I can report that yes, bunnies do have spaceships and lightsabers.)

So I wouldn’t necessarily believe Visual Chatbot’s answer to my question about the zoo rocks thing. In fact, it seems to have learned to give explanations that are total lies - if it doesn’t know the color of something, it’ll answer “it’s a black and white photo so i can’t tell” without realizing that this excuse only works on an actual black and white photo.

It’s too bad this is so tricky. Since algorithms can often be biased, it would be great if we could ask them “Why did you show me that ad?” or “Why did you decline my application?”. But getting a sensible answer from them may not be all that straightforward, especially if they pretend they know more than they actually do.

Bonus material: Visual Chatbot explains the plot of The Last Jedi. Enter your email and be edified (contains spoilers - sort of.)

kaban-bang

aiweirdness

SkyKnit: When knitters teamed up with a neural network

[Make Caows and Shapcho - MeganAnn]

[Pitsilised Koekirjad Cushion Sampler Poncho - Maeve]

[Lacy 2047 - michaela112358]

I use algorithms called neural networks to write humor. What’s fun about neural networks is they learn by example - give them a bunch of some sort of data, and they’ll try to figure out rules that let them imitate it. They power corporate finances, recognize faces, translate text, and more. I, however, like to give them silly datasets. I’ve trained neural networks to generate new paint colors, new Halloween costumes, and new candy heart messages. When the problem is tough, the results are mixed (there was that one candy heart that just said HOLE).

One of the toughest problems I’ve ever tried? Knitting patterns.

I knew almost nothing about knitting when @[email protected] sent me the suggestion one day. She sent me to the Ravelry knitting site, and to its adults-only, often-indecorous LSG forum, who as you will see are amazing people. (When asked how I should describe them, one wrote “don’t forget the glitter and swearing!”)

And so, we embarked upon Operation Hilarious Knitting Disaster.

The knitters helped me crowdsource a dataset of 500 knitting patterns, ranging from hats to squids to unmentionables. JC Briar exported another 4728 patterns from the site stitch-maps.com.

I gave the knitting patterns to a couple of neural networks that I collectively named “SkyKnit”. Then, not knowing if they had produced anything remotely knittable, I started posting the patterns. Here’s an early example.

MrsNoddyNoddy wrote, “it’s difficult to explain why 6395, 71, 70, 77 is so asthma-inducingly funny.” (It seems that a 6000-plus stitch count is, as GloriaHanlon put it, “optimism”).

As training progressed, and as I tried some higher-performance models, SkyKnit improved. Here’s a later example.

Even at its best, SkyKnit had problems. It would sometimes repeat rows, or leave them out entirely. It could count rows fairly reliably up to about 22, but after that would start haphazardly guessing random largish numbers. SkyKnit also had trouble counting stitches, and would confidently declare at the end of certain lines that it contained 12 stitches when it was nothing of the sort.

But the knitters began knitting them. This possibly marks one of the few times in history when a computer generated code to be executed by humans.

[Mystery lace - datasock]

[Reverss Shawl - citikas]

[Frost - Odonata]

The knitters didn’t follow SkyKnit’s directions exactly, as it turns out. For most of its patterns, doing them exactly as written would result in the pattern immediately unraveling (due to many dropped stitches), or turning into long whiplike tentacles (due to lots of leftover stitches). Or, to make the row counts match up with one another, they would have had to keep repeating the pattern until they’d reached a multiple of each row count - sometimes this was possible after a few repeats, while other times they would have had to make the pattern tens of thousands of stitches long. And other times, missing rows made the directions just plain impossible.

So, the knitters just started fixing SkyKnit’s patterns.

Knitters are very good at debugging patterns, as it turns out. Not only are there a lot of knitters who are coders, but debugging is such a regular part of knitting that the complicated math becomes second nature. Notation is not always consistent, some patterns need to be adjusted for size, and some simply have mistakes. The knitters were used to taking these problems in stride. When working with one of SkyKnit’s patterns, GloriaHanlon wrote, “I’m trying not to fudge too much, basically working on the principle that the pattern was written by an elderly relative who doesn’t speak much English.”

Each pattern required a different debugging approach, and sometimes knitters would each produce their own very different-looking versions. Here are three versions of “Paw Not Pointed 2 Stitch 2″.

[Top - ActualJellyfish; Middle - LadyAurian; Bottom (sock version) - ShoelessJane]

Once, knitter MeganAnn came across a stitch that didn’t even exist (something SkyKnit called ’pbk’). So she had to improvise. “I googled it and went with the first definition I got, which was ‘place bead and knit’.” The resulting pattern is “Ribbed Rib Rib” below (note bead).

[Ribbed Rib Rib - MeganAnn]

Even debugged, the patterns were weird. Like, really, really nonhumanly weird.

“I love how organic it comes out,“ wrote Vastra. SylviaTX agreed, loving “the organic seeming randomness. Like bubbles on water or something,”

SkyKnit’s patterns were also a pain. Michaela112358 called Row 15 of Mystery Lace (above) “a bit of a head melter”, commenting that it “lacked the rhythm that you tend to get with a normal pattern”. Maeve_ish wrote that Shetland Bird Pat “made my brain hurt so I went to bed.” ShoelessJane asked, “Okay, now who here has read Snow Crash?”

[Winder Socks (2 versions) - TotesMyName]

“I was laughing a few days ago because I was trying to math a Skyknit pattern and my brain…froze. Like, no longer could number at all. I stared blankly at my scribbles and at the screen wondering what had happened til somehow I rebooted. Yup, Skyknit crashed my brain.” - Rayn63

[Paw chain 2 - HMSChicago]

On the pattern SkyKnit called “Cherry and Acorns Twisted To”:

“Couple notes on the knitting experience, which while funny wasn’t terribly pleasurable: Because there’s no rhythm or symmetry to the pattern, I felt I was white-knuckling it through each line, really having to concentrate. There are also some stitch combinations that aren’t very comfortable to execute physically, YO, SSK in particular.

That said, I’m nearly tempted to add a bit of random AI lace to a project, perhaps as cuffs on a sweater or a short-row lace panel in part of a scarf, like Sylvia McFadden does in many of her shawl designs. As another person in the thread said, it would add a touch of spider-on-LSD.” -SarahScully

[cherry and acorns twisted to - Sarah Scully]

BridgetJ’s comments on “Butnet Scarf”:

“Four repeats in to this oddball, daintily alien-looking 8-row lace pattern, and I have, improbably, begun to internalize it and get in to a rhythm like every other lace pattern.

I still have a lingering suspicion that I’m knitting a pattern that could someday communicate to an AI that I want to play a game of Global Thermonuclear War, but I suppose at least I’ll have a scarf at the end of it?” -BridgetJ

[butnet scarf - BridgetJ]

There was also this beauty of a pattern, that SkyKnit called “Tiny Baby Whale Soto”. GloriaHanlon managed somehow to knit it and described it as “a bona fide eldritch horror. Think Slenderman meets Cthulu and you wouldn’t be far wrong.”

[Tiny Baby Whale Soto - GloriaHanlon]

Other than being a bit afraid of Tiny Baby Whale Soto, the knitters seem happy to do the bidding of SkyKnit, brain melts and all.

“I cast on for a lovely MKAL with a designer I totally trust and became immediately suspicious because the pattern made sense. All rows increase in an orderly manner. There are no “huh?” moments. There are no maths at all…it has all been done for me. I thought I would be happy, yo. Instead, I am kind of missing the brain scrambling and I keep looking for pigs and tentacles. Go figure.” - Rayn63

Check out the rest of the SkyKnit-generated patterns, and the glorious rainbow of weird test-knits at SkyKnit: The Collection and InfiKnit.

There’s also a great article in the Atlantic that talks a bit more about the debugging.

If you feel so inspired (and don’t mind the kind-hearted yet vigorous swearing), join the conversation on the LSG Ravelry SkyKnit thread - many of SkyKnit’s creations have not yet been test-knit at all, and others transform with every new knitter’s interpretation. Compare notes, commiserate, and do SkyKnit’s inscrutable bidding!

Heck yeah there is bonus material this week. Have some neural net-generated knitting & crochet titles. Some of them are mixed with metal band names for added creepiness. Enter your email here to get more like these:

Chicken Shrug Snuggle Features Cartube Party Filled Booties Corm Fullenflops Womp Mittens Socks of Death Tomb of Sweater Shawl Ruins

aiweirdness

In Which a Neural Network Learns to Tell Knock-Knock Jokes

The neural network can be trained to write recipes, invent Pokemon, and invent superhero names. But can it learn to tell a joke?

@researchbuzz generously provided me with a list of 200 knock-knock jokes - brief and highly formulaic, they seemed to be the form of joke best-suited for a neural network to reproduce. I figured it would quickly learn the formula, but would never, never manage to tell an actual joke.

It proved me wrong.

Very early on, the network figured out that question marks were important somehow, as were the words “Knock” and “who” in some mysterious way:

Koo k ano? ? Khnck Knock oae?t rA noc? Knock Woock Whecr lwhot h o nwoe hio

In very short order, it learned to begin a knock-knock joke, and almost never got that part wrong again. But the rest of the formula took quite a while to figure out, and the algorithm frequently devolved into playing a game of telephone with itself:

Knock Knock Who’s There? Iane Aatar who? Aaane who? Aan who? Anac who? Iobe who? Irata who?

Next it figured out the formula, but no actual, you know, words:

Knock Knock Who’s There? Bool Hane who? Scheres are then the there

Then it learned the joke about the cow with no lips which became its Favorite Thing Ever, even if it didn’t know how to properly tell it. The punchline to Every Single Joke was ooo ooooo ooo…

Knock Knock Who’s There? Alec Anic who? Wow co poo the tho tho door to the soout oo ooo oooo ooo oo oo oo oo ooo oo oo oo ooo ooo oooo ooo ooo oo oo ooo ooo ooo ooo ooo oo oo ooo ooo ooo ooo oo oo oo ooo ooo ooo ooo…

It made it past this stage, but still had problems that were remarkably similar to two people telling knock-knock jokes over a very bad phone connection, and maybe forgetting who was telling the actual joke:

Knock Knock Who’s There? Alther Water who? Water who who? Who who? Arton and out on the moor?

At last it figured out the basic formula and proceeded to make jokes that, while technically correct, didn’t make all that much sense. Many of them were partially plagiarized from jokes in the training dataset.

Knock Knock Who’s There? Hiwl Howlowwhowhow

Knock Knock Who’s There? Ireland Ireland who? Ireland you money, butt.

And then.

It produced. An actual joke. I checked, and this one most definitely wasn’t in the dataset. There was one about Alec, but the pun was “Alec-tricity”. There was one about knock-knock jokes themselves, but it was “Irish I knew more knock-knock jokes”. And it didn’t produce this just once, but over and over again. I give you possibly the freakiest thing the neural network has done to date:

Knock Knock Who’s There? Alec Alec who? Alec- Knock Knock jokes.

davince21

Alec knock-knock jokes too.

whycam-deactivated20191124

killin-the-machine

All cats the same

tupacabra

people who say “melk” instead of milk

neanerbeaner-deactivated2019090

I say melk…

tupacabra

leadingtofreedom-deactivated201

you know, melk is actually the dutch word for milk

are you really going to beat up all of holland

roguetitan-deactivated20140517

yungblkfeminist

😂

denjuu

Hello! We are the administration of Tulunk Village, the largest English-language fan community for the Keitai Denjuu Telefang, Network Adventure Bugsite, and Dino Device series. In order to better serve the fan community, the village elders have decided to ask a few questions. No personal information will be stored and you are free to ignore this survey or skip specific questions if you like.

We’re running a survey of the Telefang fan community; please take some time out to complete the Tulunk Village Census!

All results will remain in confidence with staff and no personally-identifiable information will be shared. We will use this information (whatever part you choose to share) to plan future Tulunk Village activities.

meversusanpost

Train Letter

I tried to send a letter to a train driver. I got a reply from the train driver a month and a half later. An Post win again

The sending letters to transport idea comes from Reginald Bray the king of Postal hacking. He wrote “Driver of Locomotive no 733, Caledonian Railway, Glasgow Station” on this postcard above and got a reply.

The Englishman who Posted Himself and Other Curious Objects

Trending Blogs

Recently Viewed Blogs

DaVince