Arial: Classic font, looks good for the most part, but a little bit bland also. 6/10
Berlin Sans FB: There’s already a lot more character, but the mouth is simply too long and the lips don’t curl enough which takes away from the experience. 4/10
Comic Sans MS: Now THAT’S a colon-three. There’s so much silliness in that face it’s hard to bear. My only real critique is that the eyes are just a smidge too narrow, but that’s just a nitpick more than anything. 8/10
Courier New: Sleek. Professional. Big vertical eyes full of glee. Very solid choice for a colon-three font. 7/10
Jokerman: I feel bad for including this one. That font stood no chance against the others. This is the eldritch horror of colon-threes. These eyes are filled with nothing except murderous intent. The mouth is crooked with a sharp corner, but the most egregious part is probably the teeth-like protrustions from the bottom part of the mouth. 0/10
Goudy Stout: An interesting take on colon-three. I like the idea of having an incredibly thick mouth (even if it’s a bit too thick for my liking), and the eyes being big and centered is a big positive. Much sillier than most fonts, but I think struggles to beat Comic Sans in terms of silliness. 6/10
Consolas: Worse version of Courier New. The lips just don’t curl enough and it just ends up looking a little pathetic. 4/10
Fixedsys: Oh my god. Holy shit. What the fuck. 10/10
Chicago's Museum of Science and Industry has an unusual feature: injection-molded figurine vending machines. The museum is covered in them. Most of the biggest exhibits have their own custom figurine, and if you poke around, you can find a "Mold-A-Rama" which will dispense one for you, in around one minute, for $5 USD.
Make your own EXCLUSIVE PRODUCT... molded in COLORFUL PLASTIC in seconds!
These machines are really bizarre. They melt and cast low-density polyethylene beads, then spit out the warm, waxy, soft figurine into a box. You're supposed to hold them upside-down until they cool off, because "drips of plastic" may run off them. The whole time I held mine upside-down, I wondered: is this shit going to kill me?? Is this a good idea???
These mold machines appear at the Field natural history museum as well, and the Mold-A-Rama website contains a list of the seven total locations where you can find these machines, each with a list of all the models printable at that location. The MSI has 9 different models. They sell an extraordinary amount of non-machine Mold-A-Rama merch, and you can just buy the models off their website as well (which seems to defeat their purpose).
The MSI contains a whole exhibit about the history of the vending machine devices, which were invented in the 1960s and have a very distinct space-age, industry-loving vibe to them. All sorts of current and historic figurines are on display - it seems that they have periodically added new molds for new exhibits.
The exhibit about the Mold-A-Rama devices contains a lot of pretty conflicted text. I bet this exhibit predates the news from the last few years about how little plastic is really recycled in the US at all.
I also noticed the yawning rhetorical gap between the first paragraph and the second paragraph in this wall graphic. It's very common for rhetoric about societal and material harms in the US to be constructed in this way - for the "case in favor" and the "case against" to be juxtaposed without any connective material. This is particularly common in newspaper articles where the journalist has been culturally trained to present various facts without drawing any conclusions, in the optimistic hope that their audience will draw the obviously-correct conclusions themselves, without any guidance. It's a strategy of self-preservation - it avoids saying anything which might feel too opinionated to pass review by conservatives. You can say "Kamala Harris said X" and "Trump said Y" in back to back 'grafs, and avoid saying who is obviously correct, and cross your fingers in hope that the audience will guess the truth in the gap between these isles of fact.
In the MSI Mold-A-Rama exhibit, it was very funny to me to see the painful, obvious truth hovering in the gap between these two paragraphs: it is stupid to make "novelty" plastic shit for no reason. But so many billions of people on earth make, buy, and value dumb plastic shit that does nothing for them... including me, because I printed one of these dumb toys to see what the fuck it was like.
(Pointing and screaming, at myself:) Induced demand for LDPE!! I would never have created this Object if the museum had not put this bizarre, ancient vending machine in my way!!
So what do we say about that? Bringing up plastic bags at all was intensely funny to me, because there is a ton of really fascinating and interesting stuff to be said about disposable plastic objects-of-convenience-and-novelty in a science museum, particularly for kids. I saw plenty of science museum exhibits about conserving water as a kid... surely, an exhibit about plastics recycling would be the modern equivalent? Surely a modern museum exhibit about plastics should not be about their ease of use and utility for society, since this is now a truism proven by the physical reality of the world children live in. (The exhibit dwelled awfully long on 20th century enthusiasm for plastic, and even took time to talk about bakelite.) Shouldn't a modern science museum be about the challenges of developing cutting-edge plastic substitute materials, or about the conflicted role that plastics play in our society generally?
I am not a scientist or an engineer, but if I was designing a modern "science and industry" museum, I would expect to see a lot of cool exhibits about weird new materials, and shit like mass timber construction, and that biodegradable shit they make forks in cafeterias out of these days, and everything we are attempting to do with corn, or whatever. I am genuinely curious about that shit. I know little about it.
I am not sure that the MSI is ready to go there, though. The entire interior of the museum is covered in massive, ancient text pronouncements about science's utility to man, and to industry. In the room next to the Mold-A-Rama exhibit, we found the phrase:
Science discerns the laws of nature. Industry applies them to the needs of humankind.
... above a display box containing an Atari 2600.
The whole museum feels like a halfway-executed step away from an insanely industry-friendly vision of what a "science museum" should be. I got the sense that the museum's partnerships and supporters and historical interests were very industry-focused, and that the museum might be kind of stuck with a ton of these legacy-industry-centric exhibits for a very long time. There is a massive exhibit about farming machinery which was there when I was about 13-15 years old, too. I remember seeing its models of vegetables and its exhibit about fertilizer the last time I came here as a child.
There is a massive interactive coal museum exhibit which you can buy a tour ticket for. We did not get one. The installation is multiple stories tall and is a gigantic object in one of the two-story museum wings which you physically enter. Removing the coal mine exhibit would probably be as expensive as adding a new one.
I get the sense that when you install an exhibit like this in a science museum, you are stuck with its contents - and its rhetoric - for practically a generation.
Once you see that a museum like this is a project of rhetoric and indoctrination, it becomes a surreal, hilarious experience. You can genuinely enjoy the tactile and didactic functions of such a high-quality science museum while also thinking, right in the front of your mind: oh my god. The kids are so fucked. I grew up on this too and I, also, am so fucked. This dirty, life-sized fiberglass model of a cow I'm looking at was here when I was a child, and has been subjected to repeated milking by this "high tech" robotic milker arm for my entire adult life.
It would take tens of millions of dollars in research, design, fabrication, and installation to improve any of this... but the newer exhibits are also very much designed in partnership with industry. There was a balcony exhibit about wearable tech which was just a bunch of startup advertisements.
I don't know. Funny and weird as hell. The coolest thing in the museum is a gigantic (like, 10 foot wide) disc full of red sand which you can spin a zillion miles per hour. A visitor can control its speed using a gigantic wheel. I think it was supposed to teach me something about avalanches but I think it would also be cool to just put this in a weed cafe and let people go nuts with it.
If I had a kid I would probably be a complete maniac about this shit. I would be banning them from the MSI and taking them to the Field Museum and the Shedd Aquarium only. As an adult, I can enjoy its massive, tactile physics demonstrations, and its interactive learning experiences - undeniably some of the best I have ever seen - with two of my brains ticking and spitting steam at once. And surely this stuff didn't ruin me forever when I saw it as a child myself.
The problem is that I am 35, and I saw many of these exhibits as a child. They haven't changed since. What the fuck is up with that!!!
At a public meeting in Columbia County, Georgia late last year, Dr. Rick Richards explained that he had run a piece of software he created called EagleAI against Georgia’s voter registration records.
“When we reviewed the Secretary of State records, roughly 8 million registrations, we found 1.4 million typos,” Richards said. “There were 99,523 apostrophes where there shouldn’t be. There were periods in the initials […] There were 197,851 hashtags in the field for apartment number. Well, they won’t match.” Richards said he “offered to clean this up for the state for free, and they said no.”
Richards was suggesting that Georgia’s eligible voter list was a mess, and had come to the meeting to pitch the EagleAI tool as something that would help election officials. In Columbia County, where he lives, Richards had found members of the board of elections who were not only interested in his tool, but who actively defended it when challenged by their constituents.
Using a public records request, 404 Media has obtained and analyzed several hours of audio from Columbia County Board of Elections meetings over the past year that give a sense for how election officials in one of the most critical swing states considered using AI-powered software to assist with voter eligibility challenges.
Nancy Gay, the executive director of Columbia County’s Board of Elections, told 404 Media that the county ultimately did not use EagleAI this year because it ran out of time to get trained on it before the election. But the audio shows how the software was pitched, what voters in the county think about it, and, most importantly, show how some election officials have in some cases begun repeating and spreading ideas that are popular with election deniers.
EagleAI is a piece of software developed by Republicans in the aftermath of the 2020 election to make challenging voter registrations easier. It compiles various databases and web-scraped data (changes of address, obituaries, property records) to create lists of voters whose eligibility can be challenged en masse. Voter eligibility in Georgia is often challenged by arguing that they do not actually live in the state or based on clerical errors and typos. It was inspired in part by a person named Jason Frazier, who challenged the eligibility of nearly 10,000 voters by himself. The existence of a tool like this is concerning in a country where voting access can be difficult, where voters who temporarily move or are displaced or who live in group housing can sometimes be dropped from voter rolls, and where voter suppression efforts are often successful. EagleAI has been called “a voter fraud hunting tool” and “a new anti democracy tool.”
Excerpt 1 EagleAI meeting
0:00
/662.204082
An excerpt from a meeting where EagleAI was discussed.
The push by EagleAI has coincided with efforts from conservative activists in Georgia to make it easier to remove voters from voter rolls without formal challenges to their eligibility. The Associated Press reported that the eligibility of more than 63,000 voters has been challenged in Georgia since July 1.
Kristin Nabers, the Georgia state director of All Voting Is Local, showed up at one of the meetings and said that a current Republican strategy is to distribute lists of voters who they think they can challenge because their records have typos or otherwise seem challengeable.
“These lists, even as it is now, are shared with amateur sleuths who can file hundreds of thousands of challenges at a time,” Nabers said. “It’s my feeling and the feeling of others that EagleAI is likely to falsely determine people as ineligible. Somebody could temporarily move for work or school, property records online can be outdated. They can have typos. Multiple voters can share the same name. This program uses unreliable data. We believe, as others have said, that the risk of removing lawful voters is high. The state appears to agree. EagleAI is going to overflag voters as suspicious.”
In the meeting audio obtained by 404 Media, Richards called the characterization that EagleAI is meant to disenfranchise people “90 percent horse hockey” and said it was hurting his feelings: “I just can’t believe the voracious evil and lies that people are willing to do to achieve an agenda. I’m not gonna go into all that, it really upset me that people are gonna be this dishonest about something they’ve never even seen. They don’t know me, they don’t know the product. How can they do this? So I’ll hush, before I get started.”
According to Richards, EagleAI cross-references (at least) voter lists from state secretaries of state with voting histories, state corporation data, property records from county governments, change of address data from the post office, obituaries, funeral home data, newspapers, death affidavits, county coroners, felony record information, affidavit of deaths and other data sources to create lists of voters whose eligibility can be challenged more or less with the push of a button. Voter rights groups worry that it may also be collating social media posts and other information, and worry that voters may be challenged based on inaccurate or incomplete data.
In Columbia County, where Trump won by 37 points in 2016 and 26 points in 2020, Richards found an election board that was willing to entertain using the software to help it “clean” voter rolls and analyze challenges. In effect, the county government discussed using EagleAI to help it verify voter registration challenges that very well may have come from people using EagleAI to make them.
Challenging the eligibility of voters despite an incredibly low rate of voter fraud has become a key strategy of right wing organizations in the aftermath of the 2020 election. All over the country, state courts have had to decide whether to allow last-minute voter roll challenges in which conservative groups have asked states to purge hundreds of thousands or millions of voters in any given state.
In meetings in October and December 2023, Columbia County officials repeatedly stressed that EagleAI itself does not purge voter rolls, and that it was interested in the software because Richards is a Columbia County resident who convinced them that the software could streamline what happens after a voter’s eligibility is challenged in the county. According to election officials speaking in the audio, they regularly receive voter eligibility challenges, which can be filed by anyone.
But a scenario that the Brennan Center for Justice and other civil liberty groups have pointed out is that election boards that use EagleAI are likely to use it to research voter eligibility on challenges that itself were generated by EagleAI. “EagleAI’s backers also propose that local governments use it to resolve private challenges, which would lead to disastrous registration purges,” the Brennan Center wrote in an analysis of EagleAI. “This scheme is right out of the election denial playbook. Since there’s no evidence that illegal voting is a widespread problem, conspiracy theorists create the facade of a problem and then demand that governments use their solution, never mind the side effects of disenfranchisement and intimidation.”
Excerpt 2 EagleAI meeting
0:00
/731.506939
This is the scenario that Columbia County more or less wanted to use EagleAI for.
“Anybody in this room can challenge as many voters in the county as they want to. We have had challenges of a voter in the county filing a voter challenge against a single other voter, and we’ve had someone file a challenge against an entire family because they had five voters living at one address,” Larry Wiggins, a Columbia County election board member told meeting attendees about the allure of EagleAI. “And we’ve had people file over 6,000 names on one challenge list.” Wiggins said that when the elections board receives challenges like this, it researches a voter’s eligibility manually, or using another system called ERIC. “EagleAI would be inserted there as a tool to help with the research,” he said, adding that “EagleAI lets you access [various data and voter information] all from one screen.”
“People worry the software will change things on the voter rolls or that it will disenfranchise people or it will cause people to feel like they’re being singled out. Well, first of all, they won’t know they’re being researched,” he said. “Because they’re going to be researched because something triggers the research. A death notice, a record in the National Change of Address Database, something of that nature.”
Gay, the executive director of Columbia County’s Board of Elections, said that, in Georgia, EagleAI itself would not remove people from the eligible voter list. But it could be used to automate a manual research process that results in the challenged voter being sent a letter that tells them “‘our office has obtained information that you no longer reside in Columbia County. If this is true, please complete this form, send it back to us.’ And then we remove them. If they don’t [send the letter back], we don’t do anything to the record.”
“Over the years, we have online [voter] registration, I think our voter rolls probably could have some excess weight on it, because voters do not think about changing their address or voter registration until the day they need to go vote,” she said. “I just want to clean the voter list.” She said that EagleAI would create a “more streamlined and automated” way to review challenges.
Wiggins and Gay repeatedly defended the potential utility of EagleAI, while also repeatedly stressing that it would not be used to unilaterally remove voters from voter rolls. But they both dreamt up scenarios where it could be used to streamline the potential removals of voters as the result of a challenge.
It was impossible for me to identify some of the constituents who showed up to the meetings because there is no video of them, and the audio that I obtained for some meetings was incomplete, the county told me. But many of the people who showed up came to say that they did not want EagleAI used in Columbia County.
When one person asked a voter what the Brennan Center said about EagleAI, the voter said the Brennan Center said “EagleAI is highly unreliable.” Gay interrupted and said “And how would they know? I mean, I’m playing devil’s advocate here.” The voter said “they get people who research these things and do statistics.” Gay interrupted again and said, “But do they do the job?” The voter said “I believe that they and other people who have participated in the process of looking at EagleAI do the jobs and are experts in that field, yes.”
Gay laughed, then Wiggins launched into a story about how EagleAI had identified 503 registered voters who were older than 113 years old, the oldest known living person in the United States. A man who had shown up to the meeting to insist that EagleAI not be used in the county challenged the idea that any 113-year-old person would be voting in a testy back-and-forth with Wiggins.
“They’re not voting if it says they’re 113 years old,” the person says.
“Absolutely. But you can’t assume that. There’s one person who is 113. They may or may not vote. But it’s clear their data was wrong when they had 503 people who were older than 113,” Wiggins said.
“My point is that problem would be corrected eventually,” the man said “We’re just concerned people could fall through the cracks, and you could accidentally be removing eligible voters.”
In another meeting, Wiggins said that in 2020, Columbia County got a challenge on the eligibility of 6,800 voters; the county decided that it would not accept the challenge, and a judge later threw out the same challenge in other counties. Wiggins said that “on my own time, I did a statistical sample review of those voters. And I found 13 people that voted in Columbia County that should not have.” In one of the examples, a voter had temporarily moved to Texas and voted absentee in Georgia, not Texas; Wiggins and a constituent argue about whether or not this is legal.
Speaking broadly about the appeal of EagleAI, Wiggins said “it’s sort of like, winnowing the wheat from the chaff. You’ve got to narrow down your focus to deal with the real problems or issues correctly. We’re coming up on what is going to be by all indications the biggest election in American history as far as turnout, could be the most contentious although that’s hard to believe it would be worse than 2020 and it is definitely possible. So we’re looking at everything we can do to do our job better and more efficiently.”
Discussion about EagleAI largely disappeared from election board meetings in the immediate runup to today’s election, and Columbia County told 404 Media in an email that they did not actually use it to assess any voter challenges during this election cycle. A recent challenge of 1,550 people was thrown out “due to faulty or outdated information,” Gay told 404 Media in an email.
“The plan or intent for us to use EagleAI was as a tool—so if someone dropped a challenge of multiple voters (like the most recent one) we could research their database quicker than doing our own research on multiple public websites to try and aid us in research,” she said. “But we never got to that point where we could utilize it.”
Much of last month’s meeting was about the security of vote-by-mail ballot drop boxes, in which “poll watchers” discussed whether someone would need to sign an affidavit to drop off a ballot and asked whether anyone dropping off multiple ballots would be questioned. They also said the ballot boxes—there are only two in the county, the minimum mandated by state law—are being monitored by a human guard and surveillance cameras. They are “manned,” one person said.
“As opposed to four years ago, where they were unmanned in many counties,” Wiggins said. “If somebody walks in with a post mail tray [of ballots] like they did in Atlanta, that’s not going to happen.” It’s not clear what he was referring to.
referring to a place by an exonym is imperialist but using the local language for it is appropriative, so please be sure to refer to locations by coordinates only. unfortunately longitude is eurocentric so please instead use the new Coordinated Universal Meridian which places 0° latitude at the weighted median location of all of Earth's population, updated annually based on the latest available data and published in the Coordinates Universal Meridian Mandatory Yearly Supplement
Nadella's Arguments against Copyright Misrepresent both People and Computers
By Emily
According to Katie Prescott in The Times Microsoft CEO Satya Nadella is out there arguing that copyright laws need to be reworked to enable "transformative" technology. I wouldn't want to argue that US copyright laws are perfect or even good the way they are, but the revisions needed are ones that protect the interests of creative workers against corporations, not revisions that further the interests of corporations at the expense of creative workers.
But that's not even why I wanted to write a newsletter post about this piece. Both Nadella's arguments and Prescott's reporting on them embed some insidious dehumanization that I think is worth calling out. Prescott writes:
Speaking after Microsoft's launch of virtual employees at an event in London, he compared the situation to that of using information from textbooks to formulate new ideas. “If I read a set of textbooks and I create new knowledge, is that fair use?” he said.
Nadella is making a false equivalence between what happens when a person reads textbooks or other materials, understands the ideas there, and builds on those ideas, on the one hand, and what large language models (LLMs) do, on the other. LLMs process the text (literally as strings of letters) in order to build up representations of which word spellings are likely to show up in which contexts. People can then use these LLMs to create new sequences of word spellings that look plausible and pleasing because they reflect the co-occurrence probabilities. His argument is not only specious, but also rests on minimizing what it is to be human, have ideas, learn, interact and communicate, so that he can say that the theft by companies of creative works to train their models is simply analogous to the experience of creative works by people.
Meanwhile, in the very same paragraph, Prescott refers to "Microsoft's launch of virtual employees." In the article linked to, "virtual employees" is given in quotes. I'd like to think those are scare quotes, because it is not reasonable to refer to automated systems as "employees"—even when automation is effective enough to obviate certain job classes, the resulting system is a machine or a tool, not an employee. But in this piece, Prescott drops those quotes and thus is writing as if "virtual employees" were a sensible concept.
Another aspect of the dehumanization in both the practice of "generative AI" developers and this reporting is in the conceptualization of creative work as "data":
Intellectual property rights have proved to be a contentious issue as the development of generative AI has accelerated, pitting the creative industries against the technology sector. AI requires high-quality and reliable data in order to produce high-quality results, but that data is expensive to produce.
(Aside, because it's not really about dehumanization: if it's expensive to produce, it's valuable. Pay the people who produced it.)
It's not just Nadella who makes the arguments that creative work can and should be seen as just "data" and that statistical processing applied to that "data" is analogous to how people learn. We see it frequently from tech cos and tech bros who want to justify their data theft. It's important to be able to spot both the logical fallacies in this argument, to debunk it when it is used, but also call out and resist its inherent dehumanization. Let's not normalize these dehumanizing discourses, but rather resist them at every step.
Impatient for the next MAIHT3k episode? Join us for the livestream of Episode 44 on Monday October 28, 2024. Alex and I will be having a good laugh at OpenAI's claims about "reasoning with LLMs".