Andy Rotherham: Why AI Makes Knowing Things Matter More

It’s an interesting inversion. You haven’t seen a situation where there’s been this kind of pressure on computer science graduates and where you’ve seen in some cases humanities graduates getting hired at at higher rates. I told my kids they needed to learn to code and that may have turned out to be really bad advice.

Andy, thank you so much for joining me in my my my home version of AID Studios. Andy Rotherham is the co-founder of Bellweather, one of the elite consultancies that works in the education space. He also writes on Substack under the moniker Edgiwonk.

And yeah, you’ve been you’ve been an adviser, a a thinker, a leader in the education space for for some time. And we’ve we’ve been really I mean I think we’ve had a lot of these conversations where we were just sort of chatting and I think it was in at Lone Rock in Colorado where we had this long conversation and both of us were like dang it I wish we’ recorded this so we could just sort of like transcribe it and like turn this into an op-ed or something. Maybe we still do that.

I was thinking you know let’s do this podcast. I would love to do some writing with you and I know this is something that we’ve been toying with for for some time. I wonder if the stuff that we talked about a year ago are still relevant.

And I that’ll be what I think that’s one of the things I want to really draw out today is are any of your perspectives shifting based on how the space has been moving. But to get started, I actually I did a little bit of an experiment in radical transparency. I’m going to share my my cla screen.

Oh man, you make me you make me put on my my readers. Oh yeah, I can I can zoom in. So Oh, interesting.

I uses that has some extra context. Ignore that. Okay.

So I dropped in this is like cra is I promise you this is not actually cr this is for science. But I just I just gave it your your substack bellweather. It’s okay known for heterodoxy empiricism.

I don’t what is heterodoxy? Empiricism and student focused pragmatism. This is this is this is quite I don’t I don’t even use I think this is actually way better than the bio that’s probably on the website for me.

I should screen gra I should screen grab this and I should have said it’s great to be here. It’s great to see you and and we do think alike a bunch of stuff and I have a lot of regard for your work. So I was I’ve been looking forward to doing this.

I didn’t realize you served under the Clinton administration. Yeah. Yeah.

A long long time ago I was I was a lad. Is this right? Are you an AI skeptic in the best sense?

I I guess I’m a sk, you know, it’s interesting. I’m a skeptic of how it’s going to be used in education. I’m not a skeptic of the technology.

I think it’s a very powerful technology. I’ve been playing around with this and sort of asking people. I’ve been interested where you are just a 2x two of is AI a novel technology and then is AI going to be implemented in novel ways in the education sector?

I think it’s certainly a novel technology. I think people who think it’s not haven’t really probably used more of the premium tools or or haven’t been using them for the kinds of things it can do. But I do think it’s a real open question.

Will it be implemented in novel ways in the education sector or is it going to be sort of edtech 1.0 all over again? And and and so that’s I think where I’m more I think that that may be what Claude was picking up and where I’m a little bit more skeptical. Yeah.

I wonder what it would say about me about whether I think people often assume I’m an AI skeptic, but I think it’s more along the lines of what you just shared, but we’ll get into that. Let’s just see. Okay, so this is all just factual stuff.

I’m curious if it can actually pull push on his AI skepticism productively or unproductively. I think this is fine. You know, I think people appreciate when other people say, "Alex, you know, you’re you’re such a gifted speaker." And I really don’t feel like that.

Actually, that sounds like I’m like sort of like this false humility or something. I do I often get I often get compliments for my speaking style, which which is it’s it’s very hard to replicate. I was just talking to my we just hired a comm’s director and you know, she was like, "Oh, can you send me your like talking points or sort of like a script from your latest for your most recent talk?" I’m like, "I’ve never had a script.

Like, I don’t I’m really bad at memorizing and if I try to like recite anything, I really I mean, you should see me. I was recording like some 4-minute video and it was scripted and it took us like four hours to record it because I’m very bad at memorizing." So, I have to kind of get into the flow. The minute I have some structure, it just like throws me off.

So, my role here is really just I want to I want to te te tee you up to just opine on the moment because I think the goal of this podcast is truly to just make sure that, you know, for folks folks in education are not necessarily always, you know, in the spaces where they’re hearing about the cutting edge. I mean, they’re they’re in spaces where they’re hearing about the cutting edge of edtech. But I think Andy what you and I have like one one thing we share is that we have this like deep curiosity about like the frontiers of the AI technology itself and I think that to the extent that we can create a bridge for folks that maybe just attend you know like like you know education conferences well even I mean most teachers don’t even get to attend the conferences.

I think what one of the fascinating things about our sector is just how little support teachers get in general relative to the kind of support that even like a junior analyst in our sector is getting what they’re getting exposed to and then particularly on AI just the acute lack of PD and this keeps coming up in surveys again and again the teachers are just feeling completely unsupported and sort of thrown into the wolves on this and and so people go to conferences we see stuff you get product demos all the time. People want to show you their stuff and and they’re not getting any of those kinds of experiences in any way like on the regular. And that’s that seems to need to be a huge problem, especially with something like this that’s moving so quickly.

And one of the things I’ve liked about your work is as that we’ve learned new things and as things have evolved and this technology is evolving pretty quickly, you sort of updated your your mental map and your writing sort of reflects that. And I think, you know, that’s something we need more of just given the velocity here. Thankfully, I didn’t have too many readers when I published my very first piece on Substack was it was a it was it was a piece about basically a lot of like warning that we might be hitting this sort of plateau moment and I think it was going into 2025.

I was I was preparing myself mentally for okay, you know, AI was sort of like the flavor of the day. Obviously, a nonprofit called the AI Education Project benefits from that. And I was thinking about what is it gonna like how am I going to make the case for our work if people are sort of just over it, you know, like if AI kind of is seen as this this fad that kind of fizzled out.

But we’re not there. I mean, for better or worse, it seems like the opposite. I don’t and I think actually I feel like there’s the opposite problem now.

So this is my first question for you is do you feel like people in education right now are are overestimating or underestimating the the the the raw potential of the technology notwithstanding its use cases in education but just like what it’s capable of because you had alluded to that people aren’t necessarily using the cutting edge tools. I think both and I think it depends on the person too. Like I think you get some people I mean the old sort of 21st century skills crowd is is they’re coming back out and they’re now now it’s AI and they’re they’re saying the exact same things and I think they’re overestimating and and also misunderstanding how people learn.

And then I think you do have some people I still run into a lot of people who say one of two things. A sort of it’s not a big deal which I think is a mistake. We could talk about that.

And then B sort of we should just ban it or get rid of it which even if you were to believe that I don’t think that’s on offer and places that have tried to play around with that it hasn’t gone well I don’t think so I think that’s not even if you think that’s the optimal policy which I don’t but even if you did that’s not on offer but the people who think that it’s you know not a transformative technology I don’t think fully do understand what it’s capable of you know particularly the way it can teach itself in development is how fast these tools are changing as a And I think they’re also misunderstanding some of the macro context that we’re going to be operating in. Two of which I think one is employment for kids who are you know now sort of our middle high school age kids and the kind of job market they’re going to be moving into. And then second just the effects of that.

I think there’s going to be a fair amount of economic disruption and that affects education in a whole bunch of ways including you know school finance pretty acutely and so I think there’s just a lot to pay attention to here and we’re living obviously at a time there’s a fair amount of stimulus coming at us. Says a lot of different things, but I think this is this is one I urge people to pay attention to and not sleep on. And I do think in in that way it is different than than technology 1.0.

Yeah. One other thing I guess I’ll say I’ve said this in a few settings. I think one of the things I worry about is I don’t I don’t want to sound like a Homer and that like Bill Clinton got everything right on technology because I don’t think that’s you can go back and you can look at examples of some of the stuff they he or or or Hillary Clinton thought was going to happen on technology that didn’t.

But I think something he did that’s really important and gore to and is not happening right now is when he was president, Bill Clinton spent a fair amount of time talking about the effects of technology, what was happening, the new economy, what this was going to mean, what it meant for, you know, in practical terms, sort of skills and education more generally in terms of society. And you’re just not that’s not happening right now. I mean, there’s a bunch of people talking about AI and then there’s a whole bunch of people who are sort of living their lives blissfully unaware of kind of what’s coming and what this could mean for sort of middle management jobs for just whole sectors of the economy.

And I worry about that a great deal. And I think that’s something we need to talk more about both in education and more generally is just what kind of leadership in explaining what this is going to mean and could mean not in sort of hysterical terms, not in terms of boosterism, but just helping lead the country. And I don’t mean that as like a partisan statement, but like that is something that Trump is just absolutely not doing.

He just doesn’t go out and sort of try to persuade or explain. And I think that’s a loss at a at a time like this with a technology like this this coming. Reminds me of the early days of COVID where really plugged in people knew what was coming and everybody else was about to get a big shock.

We ought to learn from that, not repeat it. Yeah. I mean, code keeps coming up which is a bit a bit scary.

Because it was one of those things where it was like it it was slow and then it was fast and it I don’t know about you, it feels like we’re in that pivot moment where it’s not slow anymore. If I mean I think some people might think that the chat GBT sort of like Cambrian explosion moment in 2023 felt really fast but I feel like that was child’s play compared to cuz and I know that Bellweather has been like on the on the cutting edge of really like leaning into understanding you know like the frontier models and their capabilities now. And we’ve had a lot of conversations with your team about this.

Most folks in education are not actually using AI. There like and I know this because you go into rooms and it’s like and I was just at South by Southwest and I you know and so this is first of all this is a biased sample set because people at I think essentially like a tech conference you know ask I always do this I ask for a show of hands if you ask people like who’s used AI like most people will raise their hand but the way to really draw out you know sort of like the level of depth is you know who’s used deep research and I just just say that without explaining it and I think in a room of like 100 plus people there was like maybe five hands that went up which is completely crazy because it’s like one button, you know, whether it’s Gemini or otherwise, you just like click a button and then you it’s two clicks, right? You click the plus button, a tools button, you click the research button and it’s to me it’s the most useful thing.

I mean, like there’s I’ve been playing around with like cloud code and all this stuff. But honestly, like just whenever I’m writing something, I’m going into a meeting, you know, just run some deep research queries. It’s like, you know, being able to go through a thousand sources and this is a very useful thing for people in the education space who want to be able to, you know, draw out u what is you know, ground make sure that their work is like grounded in in research and yet people aren’t using it.

And it’s not because they’re dumb. It’s not because they don’t have curiosity. It’s like there’s something else that just I don’t know that there is the the channel for them to even like it’s like for me I I learned about it cuz like someone sat I someone literally was sitting next to me and showed me they’re like here like let me show you how this works.

I think there’s a so I’ve mentioned your take on this. I think there’s a bit of an arbitrage play happening because I talked to some people who I will obviously not name on your podcast and out them who are using this quite a bit and they are working a lot less and they are now doing in like a couple of hours what it used to take them a day or two to do. They have tons more free time and their managers and so forth have sort of not picked up on that and I’m I I’m not using it at that level and some of what I do I can’t yet use it for like that.

So it to me it it’s an efficiency tool but hasn’t become like a hyper productivity tool like that. But with some other kinds of consulting there’s people I know who are able to do that. And so I think I I I I hear you but I think it’s a little bit but there are some real power users who are really out there.

And there’s a bunch of folks to your point who aren’t even like really using much beyond the the open source free tools and so they have a very different perspective of what this can do and the quality of it. Even though those tools I would argue are actually pretty good if you step back even like the free versions are like pretty remarkable relative to you know just a few years ago. Yeah.

Yeah, but the challen I mean so that’s interesting. Is there a I guess there is an advantage to if you are figuring out how to use the tools don’t don’t let anybody else know because then you actually you you basically want your organization unless you’re leading that organization you want your organization to basically be as laggered for as long as possible because because then yeah you have that you have that sort of internal advantage. Well if it’s a junior person my attitude is more power to you.

I do know like there’s there’s a consultant I know who had to shift. They don’t charge by the hour anymore. They charge by the project because hourly it was depressing wages too much because they were too efficient which is an interesting problem to have.

Let’s go down this path because I I I am fascinated by this the the game theory of consulting. So, you know, we I talked to a lot of people who had have like websites designed and I was talking to some one person who was like, "Oh, yeah." Like, I was working with this web designer and and they used, you know, they vibe coded my website for me and I was like, "Do you realize like I can show you how to it probably took them maybe like 30 minutes of actual time. I mean, like notwithstanding the time it was like, you know, actually going and co off coding and coming back." And there’s something about just like right now that that consultant is benefiting from the arbitrage of you know I guess people’s like lack of confidence and just trying the tools out for themselves but I think so the idea that you hire somebody to vibe code for you is like a very silly concept to me because it’s yeahiring somebody always happens right so that’ll that’ll kind of I think that’ll sort of work itself through to your point earlier about like what’s surprising I do think it’s an interesting inversion that we’re seeing is the rise of the creatives and some of these tech skills are the ones that are being replaced.

I I mean like in in basically my career since college, you haven’t seen a situation where there’s been this kind of pressure on computer science graduates and where you’ve seen in some cases humanities graduates getting hired at at higher rates. That’s a fascinating thing. But will that continue?

I don’t know. But like that’s an interesting trend. And I’m certainly of the generation where I told my kids they needed to learn to code and that may have turned out to be really bad advice.

So those kinds of things are changing and I think that should keep everybody a little a little humble on with a disruptive novel technology like this what we understand and don’t understand about it and what it’s going to end up doing. I hear lots of people being very confident about what industries are and are not going to get disrupted and I’m not I’m I’m I’m not entirely confident on those things. Yeah.

So but consulting is so interesting because yeah you have this this industry that is basically built on this model of like you know charging for people’s time and then like you have margin on top of each you know like like like each person’s time that kind of ladders up into the total cost of the consulting project. And for some there will I think there’s still some amount of runway where the clients the buyers as it were just not savvy to what the AI is capable of. But like sooner or later I’ll just actually describe the phenomenon.

So I was sitting next to somebody at this dinner. I think it was like a a deep mind and it was like some it was a CEO of this big communications firm like 200 people. So quite large and and we were talking about how oh you know like do you think this is coming for you know communications consulting?

And he said oh no I don’t think so. But in the same conversation he described this. He was like, "Oh, you know, I I was I I was like trying to I was like thinking about buying a house, a new house, and I wanted to figure out the the tax implications.

And so I asked my accountant, and you know, two days later, he sent me this like, you know, five-page memo, and he said, you know, I I I’m He probably just used AI to write it." And I think that was like that exact moment is what I think illustrates the game theory of this where where The minute your clients assume that you’re using AI, you’re sort of forced to use AI because why would like if you’re spending all this time writing this memo and then your client is basically valuing it like discounting the value to like well this is you know at best they used AI and then maybe reviewed it. You know at a certain point you’re going to be forced to just use AI because you know that the there’s there’s no benefit right to like the there’s no artisanship behind a an accounting memo right the differentiator is going to be I think for consulting it’s actually similar to what you’ve written about in schools your your recent report I thought was really good where you talked about like how this actually increases the cognitive necessity and the cognitive load that people need I think the same things be true in consulting the differentiation will be that body of knowledge You’ll hire consultants because they not because of the process they can do but because of what they actually know and they can bring to the process. So it’s going to be like like AI plus very similar to I think what you’re talking about with how schools need to approach this.

My concern in schools is like there’s a fairly pronounced anti- kind of cognitive bias in in in a lot of the education sector. And so I like your your recommendation I think is is correct and the sector sort of likes to do the opposite. So I worry about that.

We can talk about that. But I think in consulting and professional services that is going to be end up being the differentiator. It’s that it’s that degree of content expertise will then with coupled with these tools that will be the differentiator.

And if you don’t have that or it’s a part of consulting where you don’t need that then then I’d be I’d be concerned. But like what what we do is a combination of the process and then it’s also very refined expertise and I and I don’t think I think AI will impact that but I don’t think it’ll fully it’ll fully disrupt it. Yeah.

Because I guess maybe to put it like differently and like Bellweather’s a great like sort of example here if we could put put you under the microscope. I mean it’s you’re not that big. I mean, you’re big, but you’re not I mean, compared to like 100, you know, an extensure.

Yeah. We’re a nat compared to the really large ones. And we’re pretty big compared to most of the ones in our sector.

And I think I could accurately describe you as like a bit of a bit bespoke, you know, very, you know, you don’t hire Belleather to just do like busy work. But what you have is when you talk about your expertise really this is like a essentially like a proprietary data set that lives within I’m sure you have documentation internally but I think a lot of this actually like it lives within the brains of the experts that you employ and that’s something that you can’t just get from Claude right like I could go and say oh and look at everything that Bellweather has ever published but my my sense is that that’s not really reflective of the type of expertise that you would bring to the people and working on something and yeah I mean I’ll be I mean being realistic about I would say Chad GPT claw Gemini to some extent can do large parts of it but it can’t do it all and the the all matters so this is not a business you know this is not a business where getting 70% of the way there kind of gets you there you need to get all the way there and so that again that’s going to be that’s going to be the differentiator an interesting thing that you point out a challenge for a lot of firms challenge for us is like knowledge management And people invest, you know, a fortune in it, whether it’s Salesforce or other tools to try to download that and codify it. And every firm struggles with it in different ways because it turns out it is particularly among your senior people.

It’s that experience they have. It’s their ability to call on those things that ends up being a differentiator when you’re actually working with people, advising, trying to come up with ideas or solutions and and you know, so yeah, you’ll get part of the way there, but you won’t get all the way there. And this is again, you know, this is a business where to really differentiate, you got to be able to get all the way there close doesn’t work for more complicated projects.

I was talking to this a senior executive at like one of the biggest companies in the country and they I asked them, you know, they were we were talking about their AI transformation and and they had said it was kind of amazing. And I wish I’d recorded it, but I asked them like, you know, they were they were talking about all this like internal training and like upskilling programs and and I asked like, have you like what are the sort of AI leaders that you’ve like what are the the competencies or what’s the persona of like somebody at your company that’s an AI leader? And and they said, you know, it’s we thought it was going to be the technical expertise, but it turns out that like that’s like a necessary but not sufficient.

It’s like yes, you need like the technical people to create the structured data sets and do some of the clever stuff on the back end to enable the AI work, but but actually it’s, you know, it was he literally said verbat like it’s communication and collaboration skills, which is such a powerful validation, right, of like something that, you know, the the 21st century skills folks. I’m actually going to I’m going to I I need to pull up this photo. Because I because you said they’re coming out of like the woodwork again.

And it it reminds me of this. Have you seen this? The bear coming out of hibernation.

Just like the Yeah. It’s like it’s like Yeah. No, they’re bad.

For better or worse. The good news and the bad news is it’s like yeah, we’re actually this is not this is not a terribly there’s not a terribly novel answer to this question of like what are the skills of the future. I mean there is like some AI skills layered into there, but you know this is a company that is solely focused on profit.

Like they have no interest besides maximizing their return to their investors and the fact that they are honing in on durable skills like to me is like a signal of okay well this is actually the best information that we have we don’t know what the future holds is that I mean would you say that you’ve seen the same thing at bell weather I mean when you’ve cuz you all are in the midst of your I’m sure your own whatever your own version of our durable skills are different though like you can I mean in this sector I find pretty quickly you can you can sus out people who have actually really been around are in the public sector. They’ve been in government roles. They’ve served in different ways.

That kind of thing. They understand nuance. They understand like there’s always like a public narrative, then there’s what is actually going on.

And the action is sussing out the what’s actually going on like those things. And that’s that those are those are skills that you have to have had various kinds of experiences. I don’t see AI replicating that.

Anytime soon. What is the underlying skill there? Like sussing out the It varies.

We do a bunch of different things with bellweather. You know, we do academic program strategy work that’s actually in schools helping schools with school operations, culture, curriculum, instructional management, all of that. We do strategy consulting that would be familiar to anybody who’s been on the receiving end of that sort of strategic advising.

We do lots of policy research, both proprietary and outfacing. So I think one of the interesting things with bellweather is that though we have different kinds of people with different kinds of backgrounds and who are going to be working on different aspects of that work. The secret sauce is we can bring them together on unique projects and really capture that synergy.

But that those skills are going to look different in different parts of the organization. So like I don’t do a lot with the instru you know instructional management side. That’s not my area of expertise.

And that’s a different skill and background than than what I do over more on the on sort of policy and advising side. So, it’s so I’m not trying to dodge your question. It’s just a different it’s a different question.

It’s a different question for different roles. One place I will just if you want to riff on where I see it having an impact that’s kind of interesting is you will have you it can do a lot of what a junior level analyst can do. You were talking about deep research a minute ago.

There’s a bunch like if you’re you want to pull lit reviews, you want to do basic analysis, you want to work with data sets, you want to you know I I had a question recently. I was curious after one of these teacher strikes settled on on school district’s final offer versus the actual settlement. So what was the best offer the district made before the strike?

What did the union get after the strike and and what the delta was there? You know that would have taken and you would have had a research assistant on that and you would have been pulling stuff. Now, you can use AI and you can get a pretty good answer to that pretty quickly.

A very serviceable answer if you’re trying to figure out like the impact of these things. But here’s the thing. I wouldn’t send what AI did to a client, but I wouldn’t send what that research assistant did to a client either without looking at it.

And people think this is a big dunk. They’re like, well, you still have to have a human look at it. And it’s the wrong counterfactual because we have to have a human look at nothing at Belleather is always a human in the loop.

Nothing goes out that was generated by AI without somebody with that content domain expertise looking at it. And so I think that’s the kind of place you’re going to see these people think are underestimating the kind of enhancements that we’re going to see. And again that’ll I think that has huge labor market implications that we can talk about.

But that’s kind of the direction I see things trending. Yes, it’s that feels like what I think what has changed over the last six months for me is particularly with like cloud code and Gemini 3.1 and now I guess it’s I haven’t used codeex yet but apparently it’s quite good. Yeah, the people in Silicon Valley are like very like very loudly shouting now from the rooftops that like yeah that the it’s nobody’s going to be it’s going to be really really hard to justify employing juniors.

And it’s not that all you’re not going to have any entry- level people, but I mean I think there’s and at the same time I was just talking to a large teacher training organization and I was asking about like who are the teachers that are like that you are seeing be the change makers in schools and what they told me is it’s it’s the the teachers that are right out of you know that that like just started. They are the ones that are not embracing AI but truly actually like the power users. Yeah.

And so I I don’t know I feel this tension of like I’m not totally sold that like you know people like you and me are going to be the people that figure out AI. I mean I think there is like the domain expertise that’s important. I’m pretty sold I’m pretty sold we’re not Yeah.

Like it’s it’s I mean how how can we get to some level of advice to like a college student or a parent who’s like watching what’s happening and is like legitimately concerned and trying to figure out like what am I supposed to do to be on the to avoid being on the wrong side of this and well this may be wrong but I’m at domain knowledge and I when I talk to students it’s it’s learn something about something and because then you can learn about other things the the ability to learn about something allows you then to learn about other stuff. This again, this is what we don’t do in K-12 when we focus on skills or we say it’s both. When people really mean it, they they diminish content.

Content and domain knowledge really matters a great deal. And so that’s my that is my core advice to young people is learn about something, learn how to learn as a result of that. And that will serve you well because I do I don’t think you can fully game out where this is going.

And I think that can be a little bit of a of a fool’s errand, but like it’s hard to go wrong knowing things, understanding the world. Say you study biology. And even if you’re not going to be a biologist, you understand biologic processes, you understand a bunch of stuff that’s going to serve you well in other things.

Obviously, history, political science, English, all these things. And so, I think I think I think we have devalued domain and content knowledge at our peril. We’re living in a sort of postfact society as a result.

And just re-anchoring in that. I think that’s actually be more important in the AI era. Now, obviously I have a bias there.

I’ve thought that preAI too. But I think it’s I I I think it’s objectively going to be going to be more important. This idea that you can just look stuff up and all that.

I think we need to make sure people have those skills. Also like probability, math, we need to we we we do a terrible job teaching math. And my friend Kareem and you know, founded Citizen Math.

He talks about like everybody thinks math sucks and they’re right. To think that because the way we teach it, but like helping people understand large numbers, probability, all these things are really important. So that’s that that that’s my answer, Alex, is like we don’t know.

And so trying to figure out what’s the thing, you know, what’s the process degree you should get seems you can learn a whole bunch about data science. It’s going to be, you know, completely rendered you know, render not super valuable by AI. Like it’s figuring it’s knowing about things and then applying it to things like data science.

I could be wrong. Save this clip. In 10 years that may look like a super foolish prediction, but 10 years I mean we’ll be lucky if that’s like two years.

Yeah. No, but that that’s my that’s my own read. That’s my own read on it.

I was it was it the CEO Palunteer? There was a there’s one of the one of the tech CEOs was was basically saying that you know this is this is a moment where you just need to work really hard and there’s going to be this like you know the the the the alpha of like you know either you need to work hard to get like as much domain expertise over the next couple years you know before AI really takes off or you need to really like work and understand AI. But I think the the the idea that you can just the way to get ahead of AI is just to be like a power user.

I’m not sold on that. Like I I think because I think about the people who spent all this time learning prompt engineering and like you know memorizing different prompts and all the prompt prompt bibles and stuff that came out and like you know that stuff is basically now all moot. I mean I suppose there’s some skill that comes from just like the process of learning it and maybe un like but the AI is really prompting for yourself and yeah I I I’ll say this at AID we have we’ve we’ve been hiring for a handful of junior roles and we went through this like process of like okay are we hiring someone that’s just like for the AI skills because we do expect these roles are going to be leaning heavily on using AI tools and we kind of landed on no I mean you you I mean the one thing that changes is like the take-home test is now an AI test.

It’s it’s more like, you know, and it’s like we want to look at the actual prompt history of like how are you using Claude, let’s say, or Gemini to get to the end point as opposed to just like, you know, cuz previously it was like write a, you know, but like we’re hiring like a fundraising lead or development lead and it was, you know, it would have been like, you know, write a a brief for like a proposal for a for a potential funer and like, you know, provide a little bit of research and and now it’s like that should be something you can do in like 5 seconds, right? So like what what does sort of the new takehome test look like? But otherwise I mean we’re basically looking at the same stuff.

And are you I mean I encourage young people also like it’s show how you differentiate. It’s you plus the AI. Like I don’t think a really compelling pitch in a lot of the kinds of work we do or you know media any of this kind of stuff is is you instead of AI because it is AI is coming.

It’s being used for a lot of stuff. It’s it’s the it’s it’s the you plus. Why is that a differentiator relative to your peers?

So, it’s knowing how to use it in a skilled way, which also means knowing its limitations and so forth. And then why you are the human who should be in the loop. I think that’s like if you’re a young person, that’s the compelling pitch.

Now, that’s the micro. The macro. I do worry and I don’t know if you’ve like what you’re where are I worry where is our future talent pipelines if we just decide to obliterate entry level jobs because you can where do people go to get these skills I think over time you can certainly see the market taking care of that but like you know over time doesn’t like help you if you’re caught up in the in the in this regum while that’s going on and so I worry like where where where will you get sort of your tier of middle and senior people.

This idea that you just invert the pyramid like they still need to come they they need to come from somewhere. Yeah. I mean I think I don’t have a good answer to that.

I think it’s going to be there will be like a transition period that might be quite rough. I do think that like all things considered, you know, first of all, I mean, you’ve actually written about this like we’re like right now we’re worried about there not being enough jobs, but we’re not too far off from worrying from, you know, like when I go to I was I just was speaking at this event in Japan and and I was sort of like prepping with one of our partners there and you know and I was like and she was like, "Alex, like basically everything you’re talking I had to like completely redo my my talk because she’s like we don’t nobody cares about jobs running out. Like there’s there’s no anxiety about there not being enough jobs.

It’s a completely reversed, right? Like we’re worried about there not being enough people to fill the roles. And so I think there’s also that dimension that it’s it’s less acute in the US.

So we’re not it’s not part of the political discourse really. But yeah, I mean sooner or later organizations will be clamoring for, you know, young blood and and and dynamic talent that has the ability to learn. And so yeah, but there may be this period of where like how do you make sure that you’re building that domain expertise if it’s hard to get employment in, you know, certain certain careers while you kind of wait for everything to settle out.

But you said something that I just want to double click on, which is it’s it’s not it’s not the AI or it’s not even how you use AI, it’s like how you can use the AI to do certain things. And is that is are you basically because is it is it that AI is essentially just like a a multiplier but everybody has the same multiplier and so it’s like what is like the what is like the base factor that you’re multiplying like is that is that framing right? No I think it’s both your your base matters and people’s base varies based on their experience education all that and then I think your ability to use it in a skilled way and it sounds like you guys are testing for that in hiring.

We are like how do you use it? And are you able to because again AI can be everything from something that helps really generate and create new content and analysis to a glorified search engine depending on your skill in in using it. So I think I think if I’m understanding your question I would say both there’s there’s variance around both those things and the real sort of to me if the higher you’re going to be a higher multiple on both of those if you learn a bunch of stuff and you learn how to use AI you’re going to be you’re you’re going to be differentiated particularly right now if you’re a young person maybe relative to a lot of your peers I do one of the things I worry about this whole labor market thing is this generation’s coming out right now we really like gave them a triple whammy we decided let’s just see what app is with like unregulated social media for young people.

Let’s see what the effect is. It turns out not good. We also were like let’s what happens if we close our schools for a year.

And then they’re coming out into this economy that’s being disrupted by AI. That’s a that’s a really tough three things. And so I particularly worry about that.

This will work itself out over time, but I worry about that cohort a great deal. And then and then there’s the fourth thing now, which is like the AI companions themselves. Yeah.

Which feels very different from social media, right? It’s like so yeah, you and I have talked about that. I worry about that a lot.

Just in terms of like human functioning, human relationships, society, we’re not meant to have perfect friends and AI is sort of a perfect friend in some ways or because none of us are perfect and so you need your friends imperfections to help you reflect on your own and you we’re meant to be in community like that. And so this is a I think this for young people that I worry this is existential. We couldn’t keep them off social media, the addictive properties of like Tik Tok and Instagram and this kind of stuff and this is next level what we’re seeing out there with with this and you’re already starting to see it show up in some ways.

I’m I’m very concerned. Julia Freeland Fischers, you know, is doing a ton of work on this. The Catholic Church is interestingly and the last two popes have had a lot to say about it which interesting institution that you know to suddenly be opining on sort of frontier level technology but they actually have some interesting important things to say about community.

So I don’t know it’s coming from different places but I it worries me. Yeah. One of my adviserss met with the pope’s AI adviser which I he must be the first pope to have somebody dedicated to AI.

Do you think that is is there like an argument that that consulting is actually maybe the best pl okay counterintuitively consultants are maybe the most at risk because they’re it’s easy this it’s very easy to get you know to shed your consultants as opposed to maybe FTEES but if I like I I have a bunch of friends who are sort of between jobs and one thing that I’ve seen is that the folks who have gone into consulting just like you know small projects or like often building AI tools for different clients. They are learning so much like they they’ve been put onto this because it’s actually this like very dynamic space where there’s like a ton of competition but also a lot of like institutional and legacy players that are like slower moving. Being in consulting forces you to understand the market and the landscape just much more urgently than if you’re, you know, if you’re like at a Fortune 500 company, like you might have 10 years of runway before you actually really need to to pivot if you’re like an HVAC, you know, supplier, let’s say.

But am I glorifying this idea of I don’t think we consultants are at risk. I think bad consultants are going to be at risk. I think you’re going to see that across a bunch of industries because AI is going to it’s it’s going to pick it’s going to pick up some of the slack.

The thing about consulting I joke about it like is you you’re getting paid to learn. It’s an that’s like that’s the that’s the really fun thing about it is you’re you you have to learn and stay current so you can help clients and then as you go through the course of of various projects and advising and so forth, you’re you’re learning and so you’re just you’re you’re consuming so much information which is fantastic. And I think that will that that will continue.

It’s a learning field. It’s I think it’s the consultants who aren’t good at that or doing sort of things that are easily replicable. Those are the ones who I I if I were I I would be concerned, but like for a firm like Bellweather like we’re pretty differentiated in what we do.

So I’m not you know that that’s not an area of like extreme people are going to need that. But and I think you’ll see that look I think you’re going to see that in sort of you know AI is going to disrupt law a lot. But there’s gonna be certain kinds of law that you’re still gonna you are going to be heavily human focused just given the nature of it.

I think you’ll you’ll sort of see that across the board. So I don’t think you’re like romanticizing it, but like good consultants like it sounds like the people you’re describing who are constantly in that learning cycle. Like I think there’s going to be plenty of work for them potentially more because it’s these sort of disruptive times where there’s a lot of turmoil that ends up creating more opportunity to help and advise.

I wish there was a little I wish there was a little I wish there was a little less of that just now. There’s a little more a little more turmoil around government than than I would prefer. But it it but it does create it creates opportunities to do various kinds of work because people need help.

Yeah. I’m trying to get to like advice for a kid in college. I think irrespective of what their major is and I guess working backwards from you like you need to be learning because like the biggest risk of like the next like 5 years is is you being cut off from opportunities to learn and so consulting feels like a I it I feel like someone like if you’re like a freshman or a sophomore in college I think you should be like challenge accepted like take two years and become a like an expert at, you know, applying AI in whatever domain it is that you’ve chosen, whether it’s, you know, computer science or business or, you know, media.

I feel like across the board, you can look at every single sector of the economy and what you will find is that companies are clamoring to figure out like how do we find people that actually understand how to use these tools? And so, you know, I think that’s now that’s not a solution for everybody, but I think right now if you’re trying to figure out like what can I do it, but you have to be learning. And so trying to do projects for I mean even like the local pizza shop is going to be really interesting learning experience where it’s like okay like how can I use AI to you know help them do more social media or marketing or I go volunteer for a nonprofit.

I mean I think there’s help you improve your writing. So when I teach, I always encourage my students to get a writing buddy and give each other feedback. When I talk to students, I do some work down at UVA.

I’m a visiting fellow at the education policy works down there and with the students and and you tell them go get a writing buddy, but half of them, you know, it’s it’s a little awkward to walk up to somebody like, "Hey, do you want to be my writing buddy?" So some of them do it, but some of them don’t. AI now means you don’t need that. That’s your writing buddy.

It can give you if if and I tell them if you have it doing your writing for you, first of all, people are going to be able to tell because AI writing, anyone with a trained eye can spot it. Second, you’re not learning anything. You’re you’re not helping yourself.

But it it’s great on giving feedback. It’s really a good tool to be like, "Hey, give me feedback on this and so forth." And so suddenly you have a writing buddy who’s available 24/7 whenever you want that feedback and you want to do that. And so, like, to your point, if you’re an undergrad and you’re not sort of taking advantage of that to hone your skills, like you’re you’re going to be falling behind because other people are going to be doing that.

They’re going to be using that to get better. They’re going to be using it to get that kind of feedback. They’re going to be using it for interviews, to prep and so forth.

And so, it’s going to come in at like a higher level of of preparation. They’re going to up their game. And those are the people you’re competing against for jobs.

So, I think I think you’re in the nail on the head. And again, like while you’re doing that, like learn something, major in something and learn something, even if it’s not what you’re going to do for your career, like the having a body of knowledge. I I just can’t like I’m halfway through a piece I haven’t finished writing just on why we’re going to need a better everyone’s going to need a better detector in this new era.

And the best way to get that is to know something. It just helps you think more skeptically and critically about other things. And that’s the big fallacy by the way is that you can teach critical thinking absent that content.

You teach it by you get the content then you learn to think critically then you can start to apply that elsewhere and in education we get that backwards. We talk about let’s teach that skill but like you know give the kids much to actually think critically about contentwise. Yeah, that’s so eloquently put.

We because we cuz our journey was started out really focused on AI literacy and then we quickly realized that like you can understand how AI works or what it is capable of but really the most important skill is critical thinking and and then you get to well how do you teach critical thinking and it’s like well you don’t just teach critical thinking you you build those skills in English and social studies and math and science and and that was sort of our journey to actually like school transformation as the the end point as opposed to, you know, sort of like discrete like AI set of skills. But so what you’re saying is if you want to build critical thinking skills, yeah, you just have to like learn about something and that and and and be in an environment where you apply it. And it that’s that sounds to me like quite a strong endorsement of like for those who are wondering whether they should go to college, you know, maybe don’t take out like a $200,000 loan to go to like a tier three private school, but but I think yes.

Right. It sounds like I’m I’m a proponent of college. One of the things you see like people who go to college have better outcomes across the board.

And it doesn’t matter what kind of job. College educated plumbers have better outcomes than non-ol it doesn’t mean that you can’t like these other things. But it it there there’s a there’s a premium there.

If you really don’t want to do that kind of work and you do want to do the skilled trades, you should go do that. I think that’s really important work. I do think we do an incredibly poor job helping kids understand what these careers involve.

I’ll tell you a funny story I was thinking about maybe writing about this. We had a duct problem at our cabin this weekend and where we are it’s pretty remote and it’s a little difficult to get repair people up. So my wife and I are like we’re just going to have to fix this ourselves.

So we were down in like a 2 and 1/2t crawl space fixing ducks all Sunday afternoon. Like that is miserable work. I mean I that you’re wearing a respirator cuz there’s like all sorts of stuff that’s bad for you and like it’s really bad and we did that for an afternoon.

It’s kind of a funny story. You know and you’re a little your back’s a little sore afterwards. Like doing that every day that is brutal work.

And I have all of our friends in Washington who are like oh you should go do HVAC. None of them would like make it a day doing that let alone like on the regular. It’s brutal.

And and so I just think like we we we need to in general help people make we we do a really poor job helping kids actually understand what these different opportunities, what they’re going to look like, what the career paths are, and then let them make. If you love working on cars and you love engines and stuff, you should do that. Like forcing you to go to school and get a degree in something you’re not interested in is is stupid.

And to your point, going to a school where you accumulate a lot of debt for a less valuable credential is a really bad decision. But helping them understand and weight that I think is like the best thing we can do rather than what we do now which is all these various kinds of social engineering that everybody should do a trade or everybody should go to college or whatever it is like empowering people to make choices in their own lives like they’ll make mistakes. We need to make sure we’re paying attention to switching costs but like I think that’s the best thing we can do than than what we essentially do right now which is like deacto make the choice for people in different ways.

Anyways, that’s a that’s a little bit of a riff off of AI, but I don’t think AI sort of changes that fundamental equation that we don’t give kids good exposure to what jobs are actually like. And we need to do a better job of that. And a and AI, by the way, can help with that.

Like I think that’s a place where AI can play can help expose kids to things, help them that that there’s there’s clear so many of the use cases interesting, Alex, are not that I find most compelling are not the actual teaching and learning. This idea that we’re going to replace teachers with AI. It’s all this stuff around it that AI can do that’s actually interesting and compelling.

Whether that’s like what we’re talking about here or special ed, executive assistance for teachers, helping teachers communicate and handle paperwork. There’s all these very interesting use cases that aren’t the actual immediate studentf facing core piece. Yeah.

Well, let’s talk about that because I feel like it’s a very lonely time to be a sort of qualified AI skeptic. And I feel like it’s I was talking to one funer and they were like, "Alex, I feel like you’re I I know I know that you’re like the you know, you’re one of like the vocal AI skeptics in the space." And I’m like, "I’m really not an AI skeptic. I’m like I’m I’m a power user of AI.

My organization uses it. I’m a like you know vendor skeptic I think like right like I I’m skeptical of some of the current modalities of how AI is trying to be sort of like forced into schools but you know like when you close your eyes like 10 years from now you know I is like and and and and and you’re imagining like the the best schools in the country public or private you know like like how prominent is AI’s role in those schools like is it is it there is it it’s a great question I think prominent it’s going to have to be because its role in our lives is going to be prominent but prominent how and I think less in terms of this idea that we’re going to replace teachers with AI and kids are going to be learning that way and more all the various kinds of supports instructional coaching again things like IEPs helping teachers with paperwork helping teachers grade now an interesting thing is there’s a culture shift that needs to happen all these not all but many of these really compelling use cases are things that parents right now are actually the most skeptical of. So when you poll parents, they will tell you they think if a teacher is using AI to grade papers, they’re not doing their job and they’re against that.

That’s actually a pretty compelling use case that could really help teachers ease the paperwork burden, allow them to use their time differently and potentially more effectively. But we’ve so we’ve got some hurdles to get over just with some of the the cultural aspects that people have around this. But I think it’s going to be those it’s going to be those kinds of things and and a little bit if you think to edtech 1.0 like the there’s some cool tools in the classroom.

I think people who like this is not binary and people are like oh it’s all garbage or it’s all great like you know or most of it’s great. It it ranges but less compelling to me is is some of the classroom stuff than all the things we can now do that tech it’s that is is a technological advancement. When I first started doing this work, we didn’t know how many kids graduated because we couldn’t track them.

We had no denominator to figure out. And so you’d make these crude estimates based on senior year and so forth, which were wildly inaccurate, you know. And Mark Warner when he was governor of Virginia and head of the NGA came along and and and started to do work to address that ultimately led an effort to to start to be able to you to compute these things accurately.

That was because of technology that didn’t previously exist. You can see that sort of across the board in the way we we do a number of things. I think the lesson here may be the same thing.

A lot of these performance and support tools may end up being more powerful than some of the studentf facing stuff. That said, I mean, if you got Diane Taber in here, she would like disagree strenuously with me on that. She thinks like there’s a whole, you know, so there there’s there’s I think there’s room for lively disagreement.

I may be lowballing what we can expect. I’ve I have been surprised by some of the people in this field who I really respect on instruction who are just really deep on instruction who are fairly bullish actually on AI in ways that is has has sort of startled me. And so I think I think there is it is it is that that could be an issue that’s evolving or we may all be, you know, kind of updating our our thinking on that before too long.

But right now I tend to think it’s these wraparounds, these supports is more where the action is. And I think you’ll see that and you and I have talked about this. My dystopian concern is like what happens if you still have and we sort of have some of this now.

You know, you get affluent kids in private schools. They’re getting content. They’re reading books.

Are doing all this and poor kids are getting taught much more by AI and other things like you know I’m not I’m not sure that portends a great future and we’re already we’re going there a little bit like you talk to private school kids they tend to be reading more books and they’re reading the public schools right now because there’s been all the push back on this not every private school and I don’t want to romanticize it but higher quality ones yet smaller group instruction it’s still a little more text based like seeing that play out you know at at some scale that that is troubling to that that seems like you know I want I want you know the basic idea of John Dewey you know you should want for all kids what you want for your kids like that is to me sort of a that’s not a bad northstar in this sector yeah that I mean that’s that’s that is truly like how I have formed my like that’s my heristic is what are the elite private schools going to look like and even if that’s not necessarily entirely feasible like that should be what we are trying to to unlock and yeah most of the private school I mean I was talking to this 17-year-old girl who is in I mean surely one of the top private schools in the country and they’re very anti- AI like very anti-AII to a point where like it to to their detriment but it’s interesting because she she was actually at our South by Southwest workshop where we were doing like vibe coding and she built by far the coolest tool. It was like a it was like a therapeutic game for kids with ADHD and it was like based on all this research that she had done in a class that was like basically just a class where you just conduct research on a topic and in about 20 minutes she built the coolest game because basically she uploaded all her research and she now had domain knowledge of this of this of this area of like non-farmaceutical therapies for ADHD and was able to build something a lot more interesting than like some of the gimmicks that other people were building. And I was interviewing her afterward and she was talking about, "Yeah, my school is like really anti-AII." But I was reflecting on if their school hadn’t been so adamant about like keeping AI out of school, you know, would she have had this opportunity to like spend a whole year doing research?

Like, you know, I I worry about a situation where a teacher says, "Oh, you can just let’s just teach you how to use a deep research tool with AI." And like I think you or I can really leverage that tool for benefit because we’ve kind of gone through the revs of actually doing the manual research and like understanding what does an even like a good source look like versus a bad like I can you know roughly speaking go to a source on a website and you know ascertain like is this something that I can trust and you talked about that earlier about sort of just the ability to to develop a critical perspective on things. It it it has sort of like it it’s crossf functional beyond between domains, but you have to kind of do it at least once. You got to know how to do things the right way before you learn the shortcuts.

And I think that’s true of everything. And like if you like like carpentry, I’ll use like I’m a fairly mediocre carpenter, right? Like I need to do things very carefully in the right way.

People I know who do this either professionally or just much more skilled, they’re doing shortcuts. I watch them and they’re doing but it’s because they have that background in training that I don’t have and they can do the shortcuts. I can do shortcuts with things some things I do professionally that somebody just coming in would not be well served by doing.

It’s because I’ve been doing it for a long time and and so there sort of analytically there’s tricks and so forth. They can save you a little time and and and and just various kinds of shortcuts. And and knowing when not to use them is also important.

And so I think that’s the with kids you got to learn this stuff. And I worry again in schools we’re going to jump to the shortcuts before we teach you the ways to do these things. And I think there’ll be a real price to be paid for that.

I I don’t think AI obiates some of the core ideas of education that have proven to be effective and and and so going back we were talking about like again the 21st century skills or whatever. To me most of the stuff isn’t actually new. What’s new is the need to democratize it.

But most of the things we talk about like these are skills that people have had literally for millennia. But it’s the need to democratize people more more and more people need that those skills with the economy that we’re moving into. One thing just on schools I just want to pause on for a second.

Something you said I think is interesting is I think we’re at this weird time where concern about social media, concern about screens and concern about AI are all getting conflated in really unhelpful ways and school is going to end up bearing some of the load on that in ways that might not be productive. It’s I think it’s like it’s entirely possible really concerned about social media and its impact on kids and not think we should just be like banning all screens in school or getting rid of all ed tech. There’s there’s there’s a lot of nuance but it’s all getting kind of mashed up in this political moment around this stuff.

The narrative around edtech cuz I’ve heard this before. It’s like well we’ve had all this edte thrown at schools and like look at the nape scores. But if you actually look at the NAPE scores, it’s like, you know, right around the computing revolution, they actually were going up pretty significantly.

Like it wasn’t until like the the 2010s where you started to see like a little bit of like a leveling off and it started to deteriorate. But I think that we’re I think at tech is not just like a one point in time. It’s like it was this sort of 30-year process of technology evolving.

Is that is that your I think there’s correlation causation fallacies all over the place in this and so people can you the the scores have gone down so people can pick whatever their thing is. I think there’s circumstantial evidence around certain things that I think is stronger than than other things. I think some aspects of public policy and all that but yeah some of the evidence on edtech is is positive.

I mean the real story is with all this stuff is variance. It’s all this stuff’s extremely high variance and the results we see and then also what’s the counterfactual because people always like assume like you’re replacing like a super high level of instruction with a super low level of edte. Often you’re replacing a low level of instruction and we we have a problem delivering powerful instruction.

A lot of schools you’re replacing that with edtech that may be marginally better. Now is that optimal? You know that’s a good that’s a very legitimate question.

Is it marginally better? That’s a different question. So I think a lot of this stuff gets it gets it gets all bound up in in in unproductive ways.

And my like I’m watching politicians to me the real action with kids and screens and technology is like what happens when they leave school through the afternoon evening till the next morning especially overnight. You know a lot of them are on on it and politicians can’t do a ton about that in in in reality. What they can do a lot about is what happens in between like eight and three, you know, or seven and three when kids are in school.

And so schools are going to then disproportionately bearing the load because a politician wants to show, hey, I’m doing something. I’m aware of this, that’s where they’re going to go. And I’m not sure, as with some other things that are sort of more general society things we decide to fight out in schools.

I’m not sure how productive that’s going to be for educators. And I wish look an interesting thing out there you mentioned earlier some of these teachers these younger teachers it’s younger it’s they’re not all young there’s some mil there’s let’s let’s let’s give a shout out to the olds like like me u it’s it’s a mix but one of the things when you talk to the edte the companies the AI companies and you’ve probably seen this they they’re getting these teachers who are virally and sort of organically signing up for these things. So, you’ll talk to a company and they’ll be like, "Oh, yeah.

We’re in we’re in 20 school districts and we’re doing this." And then they’ll tell you how many teachers they they have. And you’ll be like, "Well, that doesn’t like add up." And they’re like, "Oh, yeah. We’ve got like another like 1500 teachers, you know, we’re adding, you know, 10 a day or whatever.

Just kind of find us." That’s interesting. Those teachers are the ones I want to hear from because they are obviously like the power users. They’re out there.

They’re seeking out these tools. They’re trying to figure out how to use them and do stuff with them. Like, we should be like harnessing and capturing that.

Those people have like learning to share and so forth and we need to we need to find them. We need to hear from them. We need to support them and and you hear this across like different vendors just these sort of free you know I don’t know what you want to call them little free radicals out there in the system.

And we should I think we should be doing we should be doing more we should be more supporting and learning from them. It’s so interesting because you started out with with like a pretty strong opinion that I that I share that with which is concern about the about social media about you know like frictionless AI relationships. But then you kind of ended on this point of like and and there’s going to be this like reflexive push to remove technology from schools to to the detriment of kids.

Like how like what should what is the position that that you think like how how can we articulate this position because like I think that like sitting with attention I find really hard and when people ask me like what is your opinion on screens and it’s just like well that’s how do I even answer that that’s I don’t have an answer because it’s too broad of a question I’m excited let’s look I think an interesting thing that’s happening is we used to think of states as a laboratory of democracy now we’re going to have nation states or to some extent so Australia is banning it you know for younger kids That’s like I think it’s going to be useful to watch see how that goes in different places and figure out what we I don’t think we I don’t think we know yet. And I think if you start in a place that like all edtech is bad or we can’t get edtech to kids fast enough because of the world they’re going in. I think those are the wrong places to to start from.

We there just needs to be a lot of learning. The evidence I don’t you know I don’t think Jonathan height is wrong. The evidence on sort of social media and kids is bad and the evidence on kids being distracted is really bad.

And I wrote a while ago, I felt like these cell phone bans were sort of a white flag of surrender that like we can’t compete with any of this stuff, so we’re just going to ban it. Which isn’t inappropriate. We ban there’s there’s other stuff we ban for kids for various reasons, but I think there’s just still a lot of learning to do and I think it’s important to just disentangle these things.

AI is distinct from edtech is distinct from social media that kids are accessing on their phones and we need to talk about E2. If we lump these together, I don’t think we have a great conversation. But to have a good conversation, you need folks who are like informed at least curious and like to your to something you said earlier like have this posture and orientation towards learning and like how do you like maybe thinking about like whether it’s clients or just schools that you’re familiar with, districts you’re familiar with that have successfully created this like culture of innovation and a culture of learning not just for like the students but for the teachers and maybe even the parents.

How how do we how can we guide districts to sort of bring their communities along for what is unfortunately a journey and not an answer, right? Like it’s not like you can tell parents what the solution is, but but I think if parents aren’t engaged in the question, you know, they’re rightfully going to just want they’re going to want answers sooner or later. And the you know, they may not necessarily be pushing for answers at the right time.

If you go to Bellweather’s website, you look under emerging approaches, there’s a whole AI section and some of the more interesting stuff there. There’s some stuff on some case studies of how AI is being implemented, which I think, you know, there there’s going to be a halflife on that because there’s so much going on, but right now those are very those are very valuable. I’d urge people to check them out.

There’s some stuff just on measuring and evaluating and making sure we’re actually setting up ways to learn from this and do what you’re talking about with innovation. And there’s not enough of that. And and in education, obviously, you know, innovation and and fad travel pretty closely together.

And so, you’ve got to be evaluating and and and being rigorous about it. I think you see this in different places. The problem is it’s not sustained.

You get a superintendent, they come through, then they move on. You get a school that does it, and then the leadership changes. And so we sort of what we have in the sector in general is sort of random acts of of innovation and then random acts of learning but we don’t have sort of a systemic approach.

If there’s any, you know, if we ever get a good federal R&D function back, that’s a place where we need we need a lot more. But I can’t point you to a bunch of places and say they have it right because again, it is so episodic. Places try different things.

They move on. And then look, the the charter ventures, you know, I think there’s a lot of learning from what happened with with Summit. There’s, you know, there’s a lot there’s there’s learning there’s learning good and bad with things that have succeeded and things haven’t succeeded.

Like we have a bigger problem I think with with that stuff not traveling and people learning than we do with sort of people trying sort of different things and trying to figure out you and I you know we’ve been there’s some interesting stuff happening in Ohio they’re trying to take like a statewide kind of approach on this and support their educators and you know you and I are friends with Nick and we’re watching kind of what’s happening there you know stuff like that I think is it should all be statewide probably not but what Ohio is trying to do I think is really interesting let’s watch and see what what what comes out of that so I think it’s a it’s there’s a bunch of different places, but I can’t point you to one place and say if everybody did this, we’d be we’d be better off. I do think there’s just, as I said, the lack of support for teachers and a lack of sort of clear rules of the road, and that is a big problem, particularly again, I think we’re going to see a real focus on policy here that may be kind of card ahead of the horse stuff before long. I mean but I think this is like I think you just articulated why you know when funders ask me well you know what is the like what is like the case study that you are trying to scale and unfortunately like I I I agree with you.

Think there are I think there are actually some charters that are that are doing amazing work and I think of like like dream or intrinsic or you know design tech. The challenge is like I was like well sure it works in this like small school but that’s not necessarily going to work in a big district which I think is probably right but but at the same time we’re all sort of grasping in the dark right now. Yeah.

So like I think that’s they need to shine some lights. We need more people shining shining lights. That’s where the action is it seems to me.

And just to bring us home like you talk about this like learning agenda. I mean you you’ve been talking a lot about just a posture towards learning. Where where do you feel like you are learning the most as you kind of try to understand this landscape whether it’s AI or education or anything else just you know and any any content medium you know this is dealers.

It’s it’s three things. That’s a great question and I think everybody’s doing it differently. I’ want to want to turn this around.

Gentleman. I want to hear your answer. One is is there’s a lot of good content out there.

Yours among them. I read your stuff. I read critics like I think Ben Riley people if if you’re not reading Ben, you’re missing out on a smart critic.

He he’s stridened on this and some people don’t seem don’t like that. But like he’s got a point of view and he’s not afraid to say it and he’s raising criticisms. Obviously Robin Lake, the work we’re doing at Bell and we have an a specific newsletter on AI.

If you go to our website, you can sign up for it as well as the work we’re doing. You know, Robin’s stuff is great. John Bailey, there’s, you know, Dan Meyer on the math pieces, he’s been talking about AI.

So, there’s a lot of there’s a lot of content out there. And then there’s more general interest AI stuff. And on the sort of sector in general, I don’t know, I’d be interested your take.

I think the economist’s coverage is the best right now of sort of mainstream sort of legacy media. Just their coverage on AI in in in general. Is just is just excellent.

The second way is convenings. You just got to be with people. It there you said it’s like there’s it’s it’s we’re sh trying to shine light in the darkness.

You got to talk to people, hear what they are saying, and that’s when I’ve gotten some of my biggest aha when people say things and you’re like, "Okay, wow." Particularly people you like regard who have real subject matter domain expertise on certain things and they’re pushing your thinking. So like any kind of new thing you’ve got to get out there and talk to people and then third educators and that’s at all levels of the system like how are they experiencing this and that includes like you know we’re talking about all the people who are power users or hyper adopters but people who really aren’t thinking about that much too that’s valuable also just to try to get a sense of what’s happening around the country because I think we can easily overindex on places that are sort of doing a cut and miss sort of the whole landscape. So those those three things combined and I think again given the velocity of this I think those three things combined they give you like a head start but I still don’t think that gives you a complete field of view.

I don’t think there are enough right now hours in the day even if you’re going to be entirely focused on this and especially if you’re not to sort of stay up stay up because there’s so much but it get that those things I think they at least give you some kind of a head start but what about you like how would you respond to that question? I mean, I think you you basically nailed it. I think you you closed with the most important piece, which is, you know, you you have to be like steeped in the cutting edge perspectives, but if you’re not also talking to, you know, teachers on Main Street, it’s very easy to get caught up in the this assumption that like everybody is sort of like having this conversation.

Most teachers are not actually talking about AI more than once a week, once a month even. Like for a lot of teachers it’s just not something that is that is top of mind and and that’s why when I saw that when I see the data like the the latest Walden Gallup survey and it’s like you know skepticism about AI is increasing completely unsurprising to me because like to your point most people are not going to the ASUGSV summit or South by Southwest and so I think but you have to balance that with that’s why there’s even more value in folks understanding the cutting edge and I don’t just mean like technology itself but like the research and like the learning sciences but also Oh, the one thing I would add is you have to use the like you really have to use them. And it’s like a because they are genuinely helpful, but also because what are we doing if like we’re talking about this technology as if like and it’s mostly through hearsay.

Like it’s just weird to me that like so many people that are supposed to be like literally shaping our strategy about this and you ask them like, "Oh, have you been, you know, trying have you tried vibe coding yet?" And they’re like, "Oh, I haven’t gotten around to it." And it’s just like, "How are you?" And there’s reporters even I mean it’s like just my new thing is I’ll like and when reporters are interviewing me I’ll be like well have you have you used deep research and they say if they say no I like will do run a research query on the story that they’re writing and send it to them. You know who’s great on that is John Bailey has been like so and and John he writes you can find him and he’s out there advising on folks. He uses all of these tools.

He does all and when they come out, he does different stuff with them. Even if it’s not related to his professional work and he’s just like a constant learner and it’s kept him really at the sort of ragged edge of of what they can do. I totally agree with you.

And use it for stuff that’s not your work. Like use it around your house. Do have figure out tasks that you’re that you could see if that that’s I’ve learned a lot just using it for other stuff and and seeing how it performs or doesn’t perform.

Yeah. Yeah. And I almost feel like we who is the who is the equivalent of like you or or John Bailey for a school because I think the beauty like like the the future I envision is one where every school has a couple teachers that are like whoa have you guys tried the latest blank and you know they’re like in the teachers lounge like sitting next to some folks at a laptop just like sharing knowledge organically.

I think that’s something that if you’re if you’re surrounded by, you know, really smart people, that’s something that we take for granted. But that feels solvable. Like that feels like I don’t know if we can get to every teacher in the next couple of years.

I do think that we can get to every school and make sure that there’s like there’s like one person that is just like learning and figuring out how to translate that learning. And so Andy to to to to end on that, I mean, I just want to encourage you please keep writing. I’m really interested.

You mentioned two different pieces, both of which sound really interesting. So, I’ll we’ll share the link to your to your Substack in the in the description and yeah, let me know if and then we should all we should write something. I would like that.

You and I have had a couple of ideas. I think it’d be great to revisit. I’ve been I haven’t been publishing much lately.

I need to get I need to get back on. I’ve been busy with a few other things, duck work, as it turns out, but more more stuff more stuff than that. But yeah, I need to give I’ve got a few things I so I I previewed too that I’ve been kind of noodling on.

I really do appreciate your writing. I look forward to getting your newsletters. They’re thoughtful.

And I appreciate just what you’re doing out there. We talked about shining a light like you are doing that and that’s that is super important right now and and you know people are stepping up to do that in schools. It’s at the system level.

We talked about Nick earlier like super valuable. And thank you, Alex.