For a while, I must admit, it looked as if software developer roles like mine were done for. How could we fight against tireless robots? But our industry is slowly realizing that making truly cutting-edge software still requires humans to think and work together, to maximize their skill sets and to practice their respective crafts. A.I. can write very good software, but it also makes it easy to do someone else’s job badly, which is part of why all those projects fail. Now that everyone can code, it’s become clearer why many shouldn’t.
— Paul Ford, A.I. Was Supposed to Give Us New Killer Apps. What Happened?
Perhaps you’ve read Ben Riley’s exposé on AlphaSchool’s “boot camp” -- along with his follow-up in which the school provides even more creepy details on the bizarre eight-week training program that high school students enrolled in the private school must complete.
Perhaps you haven’t.
Perhaps you’ve blocked from your memory the whole AlphaSchool insanity: the promise of “two hour learning” – two hours a day spent clicking through AI-mediated educational programming – that seems to have captivated the imagination of the usual crowd of tech investors and evangelists (and sadly, a fair number of journalists). I don’t blame you. Indeed, I congratulate you if you've been able to ignore it, as the hype-sters won't let us go a single day without hammering us with how transformative (and terrible) the "AI" revolution will be.
Or perhaps, like me, you simply cannot believe that we’re still talking about this very ridiculous "AI" revolution and, in particular, this very ridiculous school. (I’m cited in Ben’s piece calling AlphaSchool “snake oil” and I stand by that – the “two hour learning” claim merely the latest in a long history of educational gimmicks and fads where some hustler claims their product will enable buyers to quickly, even instantly, acquire new knowledge.)
Honestly, I felt as though Dan Meyer thoroughly debunked the whole premise of the school – and not just the “two hour learning” with “AI” instead of teachers bullshit – when he first wrote about AlphaSchool over a year ago. Over a year ago.
But instead of letting a bad idea fade away – and yes, we could say that for this whole “AI” brouhaha – there’s a group of very committed people sold on, and in turn, selling the story. It's exhausting and it's depressing, particularly as too many decision-makers in too many places still believe there is a magic ed-tech bullet. AlphaSchool keeps expanding to new cities and new states, even though it’s never being able to show that it works for all (or hell, even most) students. Indeed, reporting from Wired and 404 Media (and elsewhere) has underscored that there are all sorts of bad things afoot: both in the AlphaSchool software systems and in the classrooms themselves.
(I’ve noted this before...) There’s a scene that’s been stuck in my head for over a decade now, from a 2012 episode of PBS NewsHour, in which journalist John Merrow explored the promises of “hot new startup/school” of that decade: Rocketship. The charter school chain was also technology-driven (and tech-investor lauded and funded); it also promised to replace teachers with algorithms. Its students, most of them Black and brown children from lower socio-economic backgrounds, sat in colorful cubicles for much of their day, headphones on, bored and isolated and clicking away at their lessons.
For its part, AlphaSchool also has plans to expand into the (historically pretty terrible) virtual charter school territory. But the schools that have made most of the headlines have been its private brick-and-mortar locations – those with tuition that ranges from $10,000 to $75,000 per year. The target population for AlphaSchool aren’t disadvantaged students; and I suppose it might be easy to argue that these more privileged students will be fine even if the two hours of mandated teaching-machine-time is less than ideal, as the rest of their school day (and, of course, the rest of their lives) will be full of fun, enriching activities and opportunities.
“What the hell are we doing to these children?” I thought when I watched PBS NewsHour. To me, Rocketship exemplified the ways in which – despite all the language of “progress” often wielded to make ed-tech seem appealing to schools, to parents, to communities – these technologies and associated practices served to reinforce regimes of surveillance and control, utterly dehumanizing students as it datafied them.
Many of the charter schools that were so popular – then and now, I suppose – with the ed-tech and ed-reform crowds also touted a focus on “character education.” They demanded compliance, indeed often silence, from students. These programs were obsessed with behavior, linking students' outward expressions to their inward development.
AlphaSchool is eerily similar, even though its private schools are not marketed to low-income or marginalized families. AlphaSchool too operates a regime of surveillance and control. As Ben’s reporting underscores, it too is committed to dehumanization, humiliation, shame. And personalized learning software – whether you market it as “AI” or not – always relies on behavioral engineering.
But the “character education” that AlphaSchool doesn't demand silence or compliance in the same way – again, the appeal here is to parents who believe their child is the next genius tech entrepreneur billionaire. But it is damaging in its own right. AlphaSchool, as we can see in the details of the bootcamp, actively demands students mold their behavior and their identity, in this case into some sort of John Galt figure. The school's programming leans into the kind of pseudoscience that litters the pages of bestselling “get rich quick books” – psychological pseudoscience, economic pseudoscience alike – and the ideology of Silicon Valley's startup hustle. All of this is the kind of messaging you can regularly find on social media, no surprise, since this seems to be a school fixated on the anti-expertise of influencer-culture, the sociopathy of tech-culture: move fast and break things.
“What the hell are we doing to these children?”
“Do not obey in advance.”
I often think about Timothy Snyder’s first lesson in his book On Tyranny – his guidance for protecting democracy. "Do not obey in advance." It works for fighting fascism; it works for fighting "AI" – they're inextricable.
This week it seemed as clear as ever that “AI” is being marketed, even by its makers, as a threat. A threat to everything and everyone; a threat to life, to the very idea of future itself.
And sure, it’s a lot of “misleading metaphors” as Melanie Mitchell put it. I don't think that "AI" is going to kill us all (although the rapacious, climate-destroying commitment to fossil fuels and data centers is sure gonna try). But when "AI" proponents say this, they really do mean it as a threat, even if they don’t necessarily believe it to be an existential one. "AI" is a threat, and they want us to comply. They want us to consume. They want us to obey. They want us to click. They want us to cower. They want us to fear. They want us to be reliant on their machinery, their answers, their ideology.
Of course, as Ryan Broderick put it, “AI is still boring and also bad and also expensive.” Resisting it really shouldn't be that hard. It sucks. Software sucks. The oligarchs' vision of the future sucks.
There isn't really conclusive evidence that asking "good questions" means that you get better answers from "AI." I know it might seem true. We're taught that asking good questions – learning to form and ask good questions – is essential to knowledge acquisition, to knowledge expansion. We glorify the Socratic method. We admire people like Terri Gross, for example, because as an interviewer she has a knack for asking very good questions that, in turn, elicit very interesting responses from the guests on her show.
But that is not how LLMs or chatbots function. "Good questions" aren't necessarily what elicits "good answers" – indeed no amount of "good questions" will consistently lead to a "good answer" from a probabilistic technology. Learning how to type "good questions" into the chatbot, teaching students to think of their inquiry as "prompt engineering” – these involve demanding students bend their thinking, in form and in content, to suit the interface, the machinery of surveillance and extraction.
Sure, asking questions is important. We should all wonder. We should be curious, and we should all have the right and the courage to question. But forming those questions – even outside the chat interface – is just one piece of how we think and learn. The questions themselves are never really the goal; but nor are some sort of finitude in the answers we uncover in turn. And questions are not always framed first or formed fully before understanding starts to happen.
We aren’t machines, so let’s stop instrumentalizing every part of our cognition, constrained by that terrible metaphor.
(We might stop to think too how, in light of all the compliance that the technology industry craves from us, “good questions” – again, whatever that means – are the one place where they promise we will still retain agency and control over "AI." See? Maybe “good questions” really don’t matter at all.)
“Comparable results have been obtained with pigeons, rats, dogs, monkeys, human children… and psychotic subjects. In spite of great phylogenetic differences, all these organisms show amazingly similar properties of the learning process. It should be emphasized that this has been achieved by analyzing the effects of reinforcement and by designing techniques that manipulate reinforcement with considerable precision. Only in this way can the behavior of the individual be brought under such precise control.” – BF Skinner, “The Science of Learning and the Art of Teaching” 1954
Today’s bird is the Namaqua dove. I’m also fascinated by which birds get the “dove” moniker and which are stuck being “pigeon.” The Oena capensis is a pigeon – a very small pigeon and the only species in the genus Oena. The bird is only about 22 centimeters in length (about the size of a budgie, for reference), including its very long tail. Its range extends from sub-Saharan Africa to the Arabian Peninsula.
Thanks for reading Second Breakfast. Please consider becoming a paid subscriber, as your support is what allows me to do this work. Normally, in these Friday emails, I just link to a lot of depressing news. Instead, you got many paragraphs of depressing verbiage written by me. Surely that's worth something, eh?
Various old computers had their keyboards adorned with unique symbols. Companies like Commodore, Atari, Amiga, or even – in its previous life – Apple chose to put their company logos on keys, and there were other weird and obscure keys on weird and obscure keyboards.
But it was Apple’s recent push to move their American keyboards closer to European ones by embracing more iconography, that made me think of forgotten key symbols less obscure, ones that belonged to platforms we still use today. Even on a Mac and a PC, some key symbols didn’t make it to modern times. So let’s start with the PC side today since that part of the story begins earlier, and do Macs in a follow-up post.
For a lot of 20th century, a battle has been waging between words and icons. The first salvo was, perhaps, the traffic signs: America embraced words, while Europe relied more on iconography. (As much as it looks like it, it wasn’t just “graphic design vs. not”; as a more varied continent with multiple languages, Europe needed a more universal visual language to help people travelling between countries.)
This, I understand, trickled down to other things: home electronics, and computers. There, iconography also made it easier to make one product and sell it across all of Europe, without needing to introduce many SKUs with different UI strings.
Here’s IBM’s Selectric typewriter from the 1970s, in its American and European edition:
(If you’re curious, Express was a very fast Backspace, and Index moved the page down; both were prototypes of future arrow keys.)
Here’s IBM’s early 1130 computer from 1965, which sported an unusual symbol for space:
Some IBM laboratory and scientific computers in the 1970s and even 1980s veered more into iconography, but eventually lost to text as office PC users rejected the confusing symbols. As their keyboards morphed into PC/Windows keyboards we know today, only four symbols remained and gained widespread acceptance: ⇧ for Shift, ↵ for Enter, ⇥ for Tab, and some version of an arrow for Backspace.
But let’s look at those old symbols, some beautiful, all interesting.
The two symbols below are: Print Screen (old CRT screen turning into a piece of paper) and key beep – popular when people were transitioning from loud typewriters to relatively quiet keyboards:
Here – on the front edge of the also-forgotten Reverse Tab – you can see Home, which historically meant “return to the top left corner of the screen” and sometimes even “clear the screen”:
But my favourites were these, for Insert (now gone) and Delete (still with us):
These seem inspired by proofreader marks, which feels wonderfully old-time’y:
Building on that visual language, one could also find invert/reverse video, blinking, and underline:
And this absolute beauty, which I think meant “delete word”:
The really interesting thing is that some of those symbols survive today in Unicode. I spotted at least ⎀, ⎃, ⎁, and ⎂. The last two are for contiguous and non-contiguous underline, which I feel is a story I should know, but I don’t (yet).
Games may be life with all the hard bits removed, but they provide a way to study why people make the choices they make. Traditional games are usually played against a static background: the rewards per outcome are constant. That limits their relevance to behavior because, in real life, the rewards and consequences of strategic choices are ever changing. Now, researchers have used a mathematical model to study a series of games that include evolving strategies and randomly varying returns.
A bit of history
Perhaps the most famous game-theory contest is the prisoner’s dilemma. In the prisoner’s dilemma, a pair of thieves have been captured and are being separately interrogated by the police. If both clam up, they will be punished for a lesser crime. If one prisoner makes a deal (defects) then that prisoner gets to go free and the other gets a heavier sentence. If both make a deal, they both get an in-between punishment.
The person running the game can start it with different rewards for cooperating and defecting to explore how the optimum strategy varies with reward and risk, which the players can figure out by varying the strategies across multiple rounds. Depending on the balance between the reward for staying silent (cooperating) and betrayal, the game stabilizes with everyone betraying everyone. In this simple situation, everyone loses.
I don’t think anyone particularly enjoys typesetting in monospace.
Regular text is okay – at least as okay as it can be:
Right aligning is also fine, as long as you don’t mind counting spaces, but centering already gets tricky, as you don’t have a half space to make things truly even:
And full justification is where things get particularly weird. The spaces are just too large, and cannot be distributed evenly, creating a really unpleasant feeling:
The solution used in typesetting elsewhere is hyphenation, but in monospace hyphenation also feels unpleasant, with the hyphens drawing too much attention to themselves (and subsequently messing up copy and paste):
This is why you don’t see full justification in text files very often.
But there is one more option. You can rewrite the text to choose only words that precisely add up to the line length to avoid any double spaces.
It’s astonishing, as it goes like this on for 17,000 more words, each right margin perfectly ending on a letter, no twin spaces in sight.
The author lightly covers in the FAQ at the bottom:
What program did you use to justify the text?
None. I just chose words carefully so that everything lined up on the right hand side. Everything was done with an ASCII editor.
I’m sharing this mostly as a curiosity; some rewrites for physical books are par the course to avoid widows and orphans, but you don’t see them as much in onscreen writing.
At the same time, who among us didn’t nod in recognition at least, having once spent hours massaging a button UI string or a tooltip just to get it to fit under the certain width in a densely packed interface?
I like pets, but, for various reasons, I can’t have one. So I flirt with other people’s pets. There is a black cat across the road who likes sitting in my driveway. So I had to buy a new car with a backup camera to make sure the cat is safe. As you can imagine, I couldn’t skip this problem about cats posted by Konstantin Knop on Facebook.
Puzzle. One cat eats one stick of sausage in 27 minutes. You have four identical sticks of sausage and five cats. Using only the cats and the sausages, measure exactly one minute. All the cats eat at the same constant rate, and all the sausage sticks are identical. You may measure time intervals only between moments when cats finish sausage sticks. At each such moment, you may redistribute the cats among the remaining sausage sticks.
Log in
