A social media post from Donald J. Trump: "A whole civilization will die tonight, never to be brought back again. I don’t want that to happen, but it probably will. However, now that we have Complete and Total Regime Change, where different, smarter, and less radicalized minds prevail, maybe something revolutionarily wonderful can happen, WHO KNOWS? We will find out tonight, one of the most important moments in the long and complex history of the World. 47 years of extortion, corruption, and death, will finally end. God Bless the Great People of Iran!"

This threat is outrageous, dangerous, and unhinged.

Trump’s illegal war in Iran has already led to enormous death and destruction, including a school bombing that killed over 100 children. Congress must immediately act to rein him in before more people die.

U.S. strikes Iran’s Kharg Island as Trump warns ‘a whole civilization will die tonight’ The president had issued a deadline of 8 p.m. Eastern time for Iran to open up the Strait of Hormuz, pledging destruction by midnight if leaders don’t comply.

These are the rantings of a bloodthirsty lunatic. It is sick beyond measure to threaten to wipe out an entire civilization.

This is not what the American people want.

It is well past time for the Republican Congress to join Democrats and put an end to this war.

Day 8. Continuing the programming unit with "copy a drawing challenge". Give them a set of drawings to choose from and they work out how to replicate them, as precisely as possible, using python turtle. Tons of fun and they do some great work to figure out the angles and sizes.

Day 7.2 Then pivoted into our short python turtle programming unit. Designed to be as fun as possible to offer a pathway into our programming courses, and to open the discussion of how we get computers to do what we want in the programming mode vs the LLM chatbot mode

Day 7 Warmup on tradeoffs of an automated "tardy detector", the errors a teacher might make vs the errors a camera system might make. Planted the idea of disparate impact ("What if the system marked shorter people absent/tardy more often, even if the overall error rate was lower than the teacher's?)

Day 6 Returning the quiz, full discussion of medical diagnosis tradeoffs, why most people really hate false negatives and why "the professionals" will tend to shift the line a bit differently as they think about all the costs of false positives. Great discussion, lots of student thinking/processing

Day 5.1 Quiz on the target, training, testing, deployment vocabulary, followed by beginning an exploration of tradeoffs and the false positive / false negative distinction using a sim I built: itmeson.github.io/ai-explainer...

Placeholder People Ed-Tech Criticism. AI Refusal.

"Learn to listen, and not feel like, in your smallness, you must compensate by filling up the space around you with slop." ... @audreywatters.bsky.social

The subscription level of Audrey Watters' Second Breakfast newsletter is very much worth it. 2ndbreakfast.audreywatters.com/placeholder-...

Day 4.1 Discussion of AlphaGo and introduction of vocabulary - target, training, testing, and deployment. We're trying to communicate the idea that most/all "AI" systems have some common elements, even if the way we use them or the kind of thing they produce feels fundamentally different.

3.2 It does a great job of showing how fast change can occur and the impact that has on people whose identity is threatened by the changes.

Also serves as our intro to how AI systems get built and a lot of vocabulary: reinforcement learning, training vs inference, etc.

Day 3.1 We're getting started on the documentary AlphaGo. I have mixed feelings on it -- too much of it feels like an ad, though at this point the fact that it's 10 years old helps reduce that and I stop it a lot to point out some of those issues.

MENACE MENACE is a machine learning machine built from 304 matchboxes. On this page, you can play a virtual version of MENACE.

Then we discussed how the two bots work, how they mimic learning w/o understanding, and why Bot 2 still makes mistakes. (They use the MENACE algorithm www.mscroggs.co.uk/menace/ in js, without taking advantage of symmetry, Bot 1 is a beginner, Bot 2 has practiced against another Bot 2 many times)

Day 2.2 And then they got to move on to Bot 2. "This one's better", "It's trying to tie!", "Oh weird, it lost after I started on an edge", "It's trying to setup a fork!" "This one plays good, but sometimes makes mistakes".

Day 2.1 itmeson.github.io/tictac-demo/ They played against bot 1, beat it soundly "Your bot is trash, Mark!", then they speculated about the strategy the bot was employing. "It's always going in the bottom", "It's focusing on winning instead of defense", "It doesn't seem to be paying attention"

Day 2 Tic-Tac-Toe day! A classic in the history of AI development. We played a bunch of paper tic-tac-toe, wrote down general strategies, discussed how difficult it is to put strategy into words even when we are able to execute it reliably, and then I challenged them to beat a bot that I had written

Day 1.5 "This class is going to be in part about helping you figure out what those other reasons are so you can make good choices for yourself, and justify those choices using the real, underlying reasons. And it's also going to be about how the tech works, so you can understand that part of it too"

Day1.4 No one changed their opinion on whether this was a bad idea or not!

Which got me to my point -- "This is *ethics*. We're not done when we answer whether it works, you need to know if you want that thing, EVEN if it works perfectly, and it sounds like for most of you, the answer is NO."

Day1.3 Their questions focused on verifying the correctness of the grades and feedback.

So I showed them some of the scores GPT generated (I have an education account with all privacy settings turned on) and asserted that the scores were ALL correct (they would have matched mine).

Day 1.2: The strong consensus was that it would be a bad idea, so I asked them to identify questions they could ask a teacher (me) to help them see why it would be a bad idea.

(I made clear at this point that I wouldn't actually assign grades this way)

Day 1: I showed them scans of student work in my physics class (de-identified), then a prompt for chatGPT asking for grades and feedback I could give to the physics students. Asked students to discuss and vote in groups, "Is this a Good Idea or a Bad Idea"?

Started a new term on Monday, which means a new group of 9th and 10th graders in the newly required "AI & Ethics" course. I keep failing at keeping up with posting, so I'm going to try to just keep a thread for this class, this time.

#ethics #teach180

What work is this from?

Even if you include *all* energy usage (per capita, even though that includes industrial uses), the human still uses 1/30th as much energy. He's right that the energy for a single query is probably less than the energy for a human to give the same response - but it's still the wrong comparison.

Even accepting his premise (which I don't), the scale is still completely off -- estimates of gpt4 training cost were 50GW-hrs, while a human operates on 100 Watts, or about 20 MW-hrs up to age 20 (when we apparently "get smart"). That's about 2500 times more energy to train the LLM than a human.

Another Killing on ICE The right wing propaganda machine can’t spin the killing of Alex Pretti

Within 4 hours of ICE (actually CBP) killing Alex Pretti, the data suggested that the right wing spin machine was having trouble gaining traction, even on their home turf over on X.

Here's a Substack I wrote analyzing X data from that time. katestarbird.substack.com/p/another-mu...

Just a random thought after some conversation about AI, "workplace readiness", and education.

Learning isn't a transfer, it's an effort. It does not work without effort.

It's wrong to quantify that effort as a measure of the learning that has happened, but it's also just misguided to imagine learning happening without effort or eliminating the effort.

Speaking for myself, as a teacher, the kind of AI I want is the kind that would make the copier and printers work every time I want to use them. THAT would free up time to focus on what matters.

Why Didn't It Stick? Students forget things. What can we do to help more students remember?

fivetwelvethirteen.substack.com/p/why-didnt-... An awesome collection of wisdom about teaching for memory and transfer from @dylanpkane.bsky.social

Yeah -- I think "useful for what?" is an important question, and the answer will change w/ development. I focus on the conceptual a lot because I think it will be a better guide in the future when the teacher isn't there and the problems are real and unstructured, about what calculations to do.