It’s April 17th. On this date in 1907, Ellis Island saw its busiest day ever, more than 11,000 new Americans coming through.
Coming Attractions
You can find me next week at Harvard’s Institute of Politics on Wednesday with Margaret Spellings.
On Monday, I’ll be at UVA talking with Amber Northern about her new report outlining a path forward for IES. I liked it in the original Robin Lake, which in the spirit of DOGE was more efficient? Amber’s report is comprehensive and we’ll discuss the substance and prospects for action.
On April 29th I’ll do an exit interview with Dr. Sonja Santelises on a Decade Leading Baltimore City Public Schools as part of a joint Bellwether – The 74 production. Sign up to join via that link.
ICYMI – Todd Truitt, Jaime Osborne, and I talked about curriculum, knowledge, and pedagogy from the perspective of parents, teachers, and policymakers.
ASU+GSV
I’m just back from a few days in San Diego at the annual ASU+GSV conference. The conference is a madhouse — a feast of information hurled at you amid a swirl of receptions, happy hours, side meetings, and dinners. Yet it’s an event everyone seems to love to hate. One attendee told me their resolution this year was to stop opening conversations with how much they dislike it. Oddly, liking ASU+GSV is sort of punk.
If you don’t like it, don’t go? It’s exhausting, but I like it. Over the years I’ve presented, participated in panels, done side events, and taken a lot of meetings — so I’ve experienced it in many different ways. Here’s how I see it: if you’re looking to sit under a tree and have a deep conversation, then go to Aspen. But if you calibrate your expectations and want to learn a ton about what’s happening on the industry side of education — and make no mistake, this is a one-trillion-dollar industry in K-12 alone, and ASU+GSV covers p-career — then it’s a valuable few days.
I go in accepting that I’ll feel bad about all the people I see at a reception, event, or party but can’t really talk to. That sucks. But ASU+GSV isn’t a dinner party. It’s a rave. Make peace with that and you’ll have fun and learn.
The conference itself has also evolved. The business of the conference used to be business: deal flow, all of that. Increasingly, the conference itself is the business. And that’s fine, because there is real value in convening — especially right now, as we’re being pulled apart on an industrial scale.
Want Eduwonk.com in your inbox via Substack? Sign up for free here.
I did an adjacent event at ASU+GSV for one of our clients, MetaMetrics, to discuss a forthcoming analysis about Lexiles Bellwether is working on. You can watch it here. I think Lexiles and Quintiles may be a tool to universalize how parents understand learning in an increasingly unbundled and diverse delivery system where differences between state proficiency definitions is only one problem policymakers must help parents navigate.
I also brought Rick Hess, comet-like, to ASU+GSV, where we did a discussion with Penny Schwinn about education politics. Getting him there surprised some people…
Three ASU+GSV Observations
I was surprised by the lack of attention at ASU+GSV to the new federal education tax credit scholarship program, ECCA. This could be highly disruptive and will certainly create some opportunities for those selling education services. In the next 12-24 months ECCA could be the largest federal education program – and it’s direct to consumer. That energy was not evident.
At the same time, it also seemed to me that the industry is minimizing the growing backlash against technology. There were a few protesters, but that energy will grow — concern about social media, anxiety about AI, and skepticism about ed tech are colliding. These should really be three separate conversations, but they’re increasingly moving as one. The lack of collective action across the ed tech sector looks like a real vulnerability.
It’s cheap to say everyone is pivoting to AI the same week a shoe company did. But it’s real. So many people in this sector who have been service providers, run schools, other things outside of the tool space are now building AI tools. And many of them are good! Some are investing in them, too. There is a lot of activity.
The Weather Doesn’t Care What You Think
When something is new, analogies and metaphors can help us explain and understand it. Here are a couple I find useful for thinking about AI and education — and how I think about AI and schools specifically.
The first is weather. AI is like rain, snow, heat, or cold — it happens whether you like it or not. You can move, sure, but fundamentally you can’t escape the weather. Although some people believe AI can be banned or contained, I don’t see that happening in an interconnected world. Attempts at containment, outside of truly authoritarian regimes with broader restrictions, have failed. That’s good. I’d argue even if you have concerns about this stuff, that’s still good. And every time there is a new technology, from the printing press, through the radio, adn email, there are efforts to regulate or ban it that are disproportionate.
So for schools, whether you like AI or think it’s good for education matters less than how you respond to it. In the same way that whatever you think about cold matters a lot less than dressing appropriately when it’s going to be freezing out.
The second is carpentry. I’m a mediocre carpenter. I have friends who do it professionally or as serious hobbyists and they’re not mediocre. They take shortcuts, move quickly, and can skip steps. I can’t do those things. I take my time, I follow the sequence. The same is true of cooking: amateurs follow exact recipes, pros can improvise. To some extent this is also true in music, which is why pros can sit in and play together in ways amateurs cannot.
You can see where I’m going. The same logic applies to learning in school. If you already know how to write, you can sometimes skip steps. If you’re a skilled reader, the same is true. And in areas where you’ve accumulated a lot of knowledge, you develop a kind of “horse sense” that translates to efficiency and allows you to use some tools differently. But none of that is helpful while you’re still learning. You need to follow the steps — and the shortcutting we now call “cognitive offloading” doesn’t help you learn. It makes you worse off.
So here’s the problem in our sector. You can think AI is a novel technology or not. You can think schools will adopt it in novel ways or not. (I tend to think it is a genuinely novel technology given what it’s capable of now and will be capable of in the future. But I’m skeptical that schools will adopt it in novel ways.) So on a 2×2, I’m lower right corner. You might be realy bullish and be upper right. Or, some people don’t think it’s novel at all. Others believe it is, and that it will transform schools or at least move a great distance toward what enthusiasts now call H3 learning.
Whatever you think, a few things are worth keeping in mind.
First, the education field, ironically given what we’re supposed to be doing, absolutely loves substitutes for content and content-rich learning. It’s like a drug around here. The 21st Century Skills movement was in practice a major push to devalue content knowledge. Without close attention, “Portrait of a Graduate” approaches risk going the same way. The “durable skills” packaging en vogue today gives the game away even if you think those skills are important. Lots of people casually say it’s “both,” but when you scratch beneath the surface or push them they think content is less important and are unaware of the science about how people learn domain knoweldge. Because they don’t care. They don’t call them “mere” facts because they think they’re important! (And yes, some of these same people are the quickeset to wring their hands about misinformation and truth).
The idea that we live in such revolutionary times that you don’t need to know anything — you can just look it up — is a seductive, stubborn, and bad, idea that lives zombie-like across the sector. So if you’re hoping for a vigorous defense of content, knowledge, and facts from sector leaders, prepare to be disappointed. And that’s a big problem.
Second, ed tech 1.0 didn’t go that well — if by “well” you mean things like ROI for schools, broad learning efficacy, or genuine transformation. There are some great ed tech tools out there, and some not-so-great ones, but I’m not sure the track record bolsters confidence that implementation and adoption of AI will be rigorous, systemic, or anything other than haphazard. The lack of support for teachers and change management in most places points this way.
Meanwhile, AI is more powerful than ed tech 1.0 — and more powerful than anything that’s come before it. As a result I’d argue actual content knowledge is more important than ever.
Reader Feedback (sorta).
I recently asked why Donald Trump is not talking about the education tax credit more, given it’s a pretty good deal? Dale Chu asked ChatGPT. Perfect way to land a post that started with ASU+GSV. Here’s what it told Dale, verbatim:
It’s a really good question—and the silence is actually more revealing than it looks at first glance.
A few plausible (and not mutually exclusive) explanations:
First, school choice is a governing achievement, not a political identity marker for him. Much of what drives his communication style is not “what passed” but “what produces visible conflict or loyalty signals.” School choice—once enacted and delegated into implementation—doesn’t generate ongoing, high-drama confrontation the way immigration, trade, or cultural issues do. It becomes, in his mind, “done,” and therefore not worth airtime.
Second, he tends to talk about issues where he is the sole protagonist in a visible struggle. Even when policy is significant, he gravitates toward narratives with a clear antagonist (media, bureaucracy, foreign governments, political opponents). A federal school choice program—especially one that is broadly popular or structurally diffused through states—doesn’t naturally lend itself to that kind of story arc.
Third, credit diffusion works against presidential messaging incentives—but he’s unusually sensitive to credit ownership. Ironically, school choice is the kind of policy that could benefit him politically, but only if he feels he can ownit. If implementation is happening through states, private providers, or a sprawling federal-state apparatus, it may not feel like a clean “Trump accomplishment” he can repeatedly brand in a simple way.
Fourth, it may not be part of his “issue set” in the way people assume. Even when presidents sign landmark legislation, they don’t necessarily internalize all of it as core identity issues. His durable communication portfolio has been relatively stable: immigration, tariffs/trade, elections, cultural grievance politics, and institutional distrust. Education policy—even major structural change—has never been central to that narrative.
Fifth, and more strategically, talking about school choice might not buy him much marginal gain. If polling already shows it as popular, there’s less incentive for him to invest rhetorical energy there versus issues that move persuasion, turnout, or media cycles.
So the paradox resolves like this: what looks like a “no-brainer talking point” from a policy or governance perspective is not necessarily a “high-yield attention asset” in his communication system.
In short, it’s less that he’s overlooking it—and more that it doesn’t fit the pattern of what he repeatedly uses speech for.
Want Eduwonk.com in your inbox via Substack? Sign up for free here.
View the original article and our Inspiration here