Greg Detre turned his PhD research on why we forget things into revolutionary learning products used by over 70 million people. From putting people in brain scanners at Princeton to building Channel 4's award-winning data science team, his rare combination of neuroscience and AI expertise has transformed how teams learn and grow.

As co-founder of Memrise and the brain behind Rehearsable.ai, Greg's methods have been battle-tested at Fortune 50 companies and startups alike. But this isn't just about fancy degrees - Greg openly shares how he learned to make training work in the real world, including the sessions that fell flat and the surprising power of deliberate mistakes.

Key Talking Points:

Why cramming a team into a 2-day workshop is scientifically designed to fail (and what to do instead)
How pair programming revealed the "blunt axes" holding back an entire data science team
The counter-intuitive approach that transformed Channel 4's data team in just 12 months
Why making more mistakes than everyone else can become your superpower in AI and leadership

Links & Resources:

Greg’s Blog
Greg’s Website
Greg’s LinkedIn
Lecture: "Gin and the Cognitive Surplus" by Clay Shirky
Rehearsable.ai

Today's Exercise: The Five Whys Post-Mortem

This blameless review process helps teams learn deeply from mistakes and build psychological safety. Instead of finger-pointing, it reveals systemic improvements that prevent future issues.

Steps to Apply:

When something goes wrong, gather the team
Ask "why" five times to dig deeper than surface issues
Document multiple root causes at different levels (process, culture, systems)
Choose 1-2 specific improvements to implement
Schedule follow-up to check if changes are working
Celebrate the learning opportunity rather than dwelling on the mistake

Automated transcription

Please note : This transcript is automatically generated and provided for your convenience.

Rob: Welcome to Superpowered, the unique stories of modern leaders. I'm Rob D. Willis, a storytelling consultant and speaker. And each week I sit down with leaders from inside and outside the business world to hear those unexpected stories and insights, which can transform the way that you approach work and life.

Today I am joined by Greg Detre, who has a fascinating journey which started in academia and ended up at the front lines of AI innovation. His background combines psychology, neuroscience, and technology, giving him a rare perspective into how we can unlock human potential. In our conversation, Greg shares some invaluable insights about how to upscale your team effectively, drawing [00:01:00] from his academic research and experience as a leader.

If you haven't done so yet, please make sure to subscribe wherever you listen to podcasts and get ready for this masterclass in learning and development with Greg Detre.

Greg, welcome to the show.

Greg: Hello, it's great to be here.

Rob: You have recently co founded a company called Rehearsable. ai. And from what I read, it's using AI to create. unique interactive practice scenarios, which for someone working in training facilitation, like myself is super exciting, but for listeners who don't know you, can you tell us a little bit about what you do and your background?

Greg: Yeah, well, I, trained as a psychologist, and psychologists often study their own deficiencies. And I studied why we forget things. I put people in brain scanners, and I used AI to try and read their mind. From there, I co founded a company called Memrise and [00:02:00] our aim was to help people learn 10 times more effectively than staring at a textbook.

It's still going strong with 70 million users and most recently I was their chief data scientist at Channel 4 and I spend most of my time now as a coach for CTOs and helping companies that want to try ambitious new ideas with AI.

Rob: Yeah, exciting stuff. And definitely the right time and place to be working in AI right now. I think but that is so exciting. And you've been. Working with learning a long time, you've got these different angles of understanding as a founder, as a technologist, as a leader, as a coach. So just as a brief overview, what core principles have you discovered that actually make learning stick?

And how does that differ from training in your opinion?

Greg: Hmm. Well, It's nice when we're [00:03:00] thinking about learning to pick a really bare bones example. So I often use vocabulary learning for a foreign language, because that was one of the focuses for Memrise and I know it really well. And there are a few tricks that maybe we know, but when you do them, each one probably doubles the effectiveness of your learning.

So the first one, actually, is just to pay attention in the first place. This is particularly important if you're trying to hear someone's name, and you're more worried about whether you have spinach on your teeth. next is to, Really try practicing recalling it. Often if we just stare at the word on a page, then we might feel like we know it.

But it's only when you try and remember it multiple times that you really, really bed it down. And ideally, rather than, repeating multiple times in a row, you want to spread those repetitions, those reminders, out. And as Tolkien put it, Deep roots are not touched by the frost. So, the longer you've known something, the [00:04:00] more kind of deeply rooted the memory is, and so the less often you have to remind yourself.

So for instance, you might want to remind yourself after a couple of minutes, an hour, a day, a week, a month, a year, a decade, and that that schedule, spacing them out, is a kind of optimal way to really internalize things efficiently. So those are some of the basic principles. There's one thing to do that when you're learning, foreign language vocabulary, which is quite kind of.

Stripped down, it's a whole nother thing to try and think about how to train people in more complex frameworks, techniques, and, skills.

Rob: Okay, so spaced recall, I guess, would play some factor in that. And that's kind of what we miss in training, because people come on a workshop, we do the workshop, and then we leave them to it very often, because that's the limitations of the job. recall must be a part of that as well, though, surely.

Ha,

Greg: it's interesting you mentioned that. I think having people work really intensively for a [00:05:00] day or two learning a new topic and then wander off and not practice it for six months. It's probably the regime I would design as a psychologist to minimize, how well they remember things.

But, you know, for all kinds of reasons, it's not so easy to to do things differently. And actually that was one of the core intentions behind Rehearsable, and I should say that Georgie Peak has led the development and she gets the credit for a lot of the ideas and all of the implementation. One of our ideas was how could we how could we provide a way to attend a workshop, say on difficult conversations or giving feedback or negotiation, and then be able to practice that in the privacy of your own bedroom. Imagine giving somebody feedback, especially if you know you've got to speak to, you know, Fred tomorrow and that Fred is really really ornery or prone to crying.

How are you gonna, how are you gonna handle that? Well it'd be great if you could have a practice chat with an AI version [00:06:00] of Fred the night before and work it through a few times. And that was the core inspiration behind Rehearsable, so that you would be able to get this practice effect spread out over time.

Rob: Absolutely. And the other thing that it does is it removes the stakes or makes them a lot lower. And you see this, and maybe an odd comparison, but you see it in professional athletes, when they begin to have a bit of a bad patch. they need to go to more simple or less difficult competitions, begin to build the confidence up using the skills, and then they can go back into those bigger arenas.

And the same thing, you're talking about low stakes for a difficult conversation to practice it, then you build it up over time and you're able to apply the the principles correctly. That's really interesting. Let's think though about how it looked in Real life and you were telling me about a time at Channel [00:07:00] 4 when you were a leader there leading data scientists Could you tell me a little bit about your experience at Channel 4 and how you went about?

applying these sorts of principles that you learned as a neuroscientist and Psychologist, how did you apply them in practice?

Greg: So I joined Channel 4 to an existing team that had been around for a while you. And it was a very young team. Everybody was in their early 20s. Really, really smart, lovely, motivated group of people. You couldn't ask for a better team. But somehow we were punching below our weight. We hadn't really released anything great in a little while.

And I was trying to figure out why that was. of the things that I, that really helped was I was pair programming with one of the junior data scientists which is basically when you sit next to somebody and one person has the keyboard, maybe you switch the keyboard back and forth, and you work on a project together.

You know, you could obviously do pair programming with [00:08:00] writing or almost any task. And it's really good because all It gives you a chance to see them, how they work and to notice where there might be gaps. and also take turns reflecting, giving each other feedback. Somebody, one person's driving, the other person's navigating.

It's, it's, it's a really lovely experience. I'm quite annoying actually to pair program with, but at least it's extremely diagnostic as a leader. And I noticed a few things. I noticed, for example, that even though he was really smart, he didn't know. How to use his debugger. In other words, when there was a problem with his program, he would add print statements that would display the status of the program at various points, rather than using a more sophisticated tool that enabled him to step through and interrogate things in a more refined way.

So it was a little bit like having an incredibly talented young chef who didn't realize Well, who was still cooking over an open fire. There was really only so far you can go if you don't know how to use how to use your oven. And so that was a first thought where I was [00:09:00] like, well, I wonder if the rest of the team know how to use the debugger.

And I chatted to a couple and was like, oh, wow, none of them do. And in fact, Nobody's ever taught them how to use the command line and they only learned to program in their masters last year. And so I started noticing all these places where they might be able to improve. And I probably just recently read this quote that is attributed to Abraham Lincoln.

If I had to chop down a tree in six hours, I would spend the first four sharpening my axe. And I thought, wow, we've got a bunch of brilliant people and a bunch of trees. And everybody's axes are blunt. So, what can we do to sharpen our axes collectively? And that was how the first sort of training session was born.

Rob: Okay. And the first training session was the debugging program, how to use that.

Greg: Exactly. So we set aside a day, actually. I managed to convince my boss that this was a good idea. And [00:10:00] I was aware that it would probably help to come at this from multiple pedagogical angles. So I picked two or three really good articles or videos and said, okay, everybody, I want you to read these beforehand. And then we sat down at the beginning of the day, and we had a conversation about those. What did you take from it? What did you agree with? What was surprising? And and then I did a demo. And then we would work on a, a kata, which is like a sort of martial arts word for a an exercise, a formal exercise.

So I get everybody to try and do the same thing, but to apply what they just learned and then give each other feedback on how they'd done it. And then finally in the afternoon, we did a big collaborative game. So imagine like a kind of three hour completely madcap hackathon that was a complete disaster, deliberately, because it's impossible but we learned, but you get to, like you say make a complete hash of things with very low stakes.

And so what I was trying to do was [00:11:00] to practice a set of skills through Multiple kinds of approaches starting with reading and gradually getting less passive more active Spiraling around the same topics, but in more and more detail each time Giving each other feedback and and also just providing a forum where we could where we could fail often

Rob: Yeah. And that was one day?

Greg: That was one day so basically did that every month or so for a while and I'd remember it was an enormous amount of work to prepare this and quite stressful. And I'd done two or three of them and I was like, am I, am I just wasting my time self indulgently? I'm not really seeing, I'm not sure how much effect this is having.

They all sort of, we have a good time, we go to the pub afterwards, but I'm not sure I've noticed any kind of meaningful effect. I kind of stuck with it because I didn't know what else to do. And a bunch of them were, were, fell flat or didn't seem to land in the way that I'd hoped. So it's not like they were all brilliant, but some of them [00:12:00] felt kind of positive.

And all I can say is that after a year, the effect was night and day. So it was something about it that compounded, that took time to really bed in. And yeah, after about a dozen of them, It was just a really, really high performing bunch of people. And increasingly taking ownership of their own training because I kind of run out of things that I knew how to help them with.

Rob: Yeah, I sometimes get the feeling that it's not necessarily just the content in the session. It's about the kind of environment that you create through simply having those sessions and through creating that belief that you can improve that giving feedback to one another and so on. But, okay, let's just say with the gift of hindsight, thinking back on the times that it really did land and did work and the times that it maybe people were not as crazy about the day.

Do you see any trends or patterns as to what [00:13:00] made the learning or training stick better on those particular days?

Greg: Well, there's some obvious ones when, when I pick really good articles that were genuinely clear they would do a lot of the work for me. Cause if I'm trying to teach myself, you know, Functional programming techniques or the scientific method. There's just no way I'll do as good a job as the best YouTube lecture or the best reading.

So investing in picking really good stuff. Actually what worked really well was when other people did it. took did the teaching or led it. Especially if I spent some time with the people who are leading it beforehand, just to to ask them to practice with me, to give them a little bit of feedback.

I think people learn much better from peers. And so that created a really great atmosphere. And I guess finally I was willing to make a mess. I was willing. for some of the sessions to not go very well. And that Because I was optimistic that if I was prepared to take [00:14:00] risks, the ones that did work would go really well.

So one of my favorites was when we did an emergency drill. So I deliberately broke something, but I didn't tell them what and said, Oh my goodness, CEO is doing a demo in three hours and we need to fix this. And so everyone scrambles and we talked a little bit about how to handle emergencies, how to break things down, how to structure things.

And I would pretend to be the stakeholder and like just kind of charge into the room and bellow and like demand to know how things are going and just try and create as much pressure and discomfort as possible. And their job was to try and kind of communicate regularly with me. That was just enormously good fun.

And at the end we went to the pub and I think and debriefed and probably the pub was, was maybe where all the learning really happened.

Rob: that is the, the, the English way, isn't it? The true breakthroughs happened in the conversation at the pub afterwards. But it's so, so cool to hear about and [00:15:00] working in the learning space myself, I do hear a lot of similar, I mean, theories that very much correlate with what you're talking about. Flipped learning is one of the key ones where, I mean, To use the example of a school, you, your homework is to learn the lesson, but the lesson with the teacher is where you do the homework.

So you're able to listen to the greatest speaker on X, talk through something, and then you do the homework with your teacher. So you're getting constant feedback and learning from them. Your other point about learning much better from your peers, I think it might actually also be a Tolkien quote about learning from those just ahead of you on the journey rather than the world expert because you've got so much more in common with them, of course.

Greg: And that sweet spot of difficulty is such a great point. There's lots of evidence, including from my own PhD, that if you can be in the [00:16:00] sweet spot where things are tricky enough that you're having to reach, but not so tricky that you fall flat, that that is the sweet spot where you, you maximize your learning.

And so I think a really big part One's job as the facilitator is to try and be adapting that level of difficulty. And indeed, if you've got a mix, a group of mixed ability or mixed seniority, then you can have the senior people helping the junior people and then everybody's. getting their own rich experience from it.

And in fact, I remember this from Memrise that we spent a lot of time thinking about engagement, and probably the single biggest factor was, if you can get the level of difficulty right, then it's just incredibly engaging.

Rob: people talk about a flow state when it's just difficult enough that it's a challenge, but not so difficult that it's overwhelming. All kinds of ways to think about it, but you need to be trying things that are still hard for you, but not so hard that they'll stop you from wanting to do them lots of times, I [00:17:00] guess.

One thing I'm interested about though, because this is not something I have any contact with, is dealing with arenas where there's new stuff coming out all the time. I mean, I work with storytelling and it's kind of been around for hundreds of thousands of years, or tens of thousands, but when you're dealing with something, You're learning new programming languages or new technologies, AI, fantastic example.

The stuff that is, people need to understand now, didn't exist when they were doing their computer science degree. How does a leader, or how did you deal with that kind of thing?

Greg: That's a great question. Well, so we could point to at least two resistances. There's resistance from the team, and there's resistance from me. So if we start with the team they might be resistant to learning about this for all kinds of reasons. Maybe just the fatigue of the new maybe because they feel threatened or, or any one [00:18:00] of Not even realizing perhaps what they don't know.

So one thing that what you can do is to start with a conversation about feelings, which is always a good place to start. And say, okay, well, you know, what are our concerns? Is there anyone who feels like this is going to be a waste of time? Let's talk about why. And hopefully one has a psychologically safe enough culture.

That you've already created, that you can have that conversation. And on a good day, I'm able to be sort of receptive and think this stuff through without just kind of pouncing on people and telling them they're wrong. And so that usually is a good basis for at least opening people up. And I'll often frame things as an experiment, but listen, I'd like you to, Having, having talked through all these reasons why we think it might be a waste of time will you at least try and experiment with me?

Either for a few hours or a couple of weeks. And if at the end of that we conclude it's a waste of time, then, okay, I'll be glad for you to tell me you told me so, and we'll, we'll decide what we can learn from that. So framing it in those [00:19:00] terms after having talked through the feelings as an experiment.

That helps sometimes get through the resistance from the team. And I'll say there's also resistance from me. I've been doing this a while, and sometimes, especially recently, I've been experiencing future shock. And one of the ways that I've been dealing with that is to try and well, to, to feed two birds with one scone.

So I might Ask one of the more junior developers to lead a particular section to do a little bit of research on a new tool and give us a three minute tips and tricks demo. And that's a great thing for them. It's quite good fun. And they get to be the central attention and the maven in the room.

And and it takes the pressure off me. And so whenever you can Yeah, whenever you can feed the homeless to the hungry and get a win win situation. Yeah, I look for ways to do that.

Rob: That's really, really cool actually. Can we just stick with the idea of resistance and experiments though? Because I also understand you've got this background [00:20:00] in neuroscience, you understand the mind, and from that you come up with principles about things that should work in terms of learning and training.

And I'm wondering, is there something or an approach maybe, that scientifically should work, but for some reason didn't work?

Greg: Yes. Ha ha ha ha ha ha ha! Well I can think about Memrise, where we spent a lot of our early energy on mnemonics. To help people learn. Yeah. Like if you imagine the Pope holding a tiny can of Coca Cola, that might help you remember that St. Paul, the Pope is the capital of Minnesota. Mini soda.

Anyway. So these are daft mnemonics that can be obscene or funny or emotional and They are very effective at scaffolding, but it turns out that there are a bunch of disadvantages to them, including they're a little bit effortful to create, and I think they're too slow to use in real world language [00:21:00] learning.

You have to keep being like, okay well I'm in a conversation, someone's talking about Minnesota, oh wait, well, there was the Pope, and then, oh wait, it's already, you know, and then the pub conversation's already moved on. So so mnemonics were one of these kind of early dead ends. I think the other thing realizing that, the stuff that works best is often more effortful and that maybe you're better off optimizing for engagement than efficacy.

And and so picking things that may be slightly less kind of efficacious in terms of the every single moment spent is going to maximize your learning bang for your buck, maybe is less painful. Find a way that's more fun, but slightly less efficacious. Yeah, those are two big lessons.

Rob: that resonates a lot with what I, my, my own experience because As a, let's say, a field expert, I go into a company with one particular thing that I talk about, and I'm crazy about [00:22:00] it, and I think it's really important, and I want people to get the best possible understanding of it, yet they do not have the same desire to understand the topic as deeply as me, and they don't need to, it's not their job. Optimizing to make people interested in something sometimes is a good tactic. Some people will want to go further with you, but not everyone will, and that's fine. So, I can't remember where I read this, but someone mentioned, I think it was LinkedIn, they came up with this cool, principles or pillars of learning, that it's a mixture of personal motivation, relevant content, and a supportive environment.

And you need the personal motivation, for sure. And this is as important as the most advanced neuroscientific learning methods, for sure. Just one more question I have about your own story. And that's actually from being an academic. And I'm wondering, having been in academia, and done academic learning, going as deep [00:23:00] as you can into a subject.

Are there any lessons do you think from that environment, which are relevant and applicable to people in a corporate job nowadays?

Greg: Maybe two jumped to mind. One was that I was quite disciplined at gathering useful materials as I went. if you're reading lots and lots of papers, keeping track of the ones that you think are particularly great and, and coming up with a system for Being able to find them again ended up being really handy when you're having hundreds of meetings and you're trying to keep track of what you've learned in each and how they might relate to one another.

So some of those systems, and I've been obsessed with note taking ever since, I think that ends up being helpful with the training because if I realized that we need to do a training session on, I dunno, data visualization, then I can look up my notes on data visualization. I'd be like three or four articles that I have read in the past.

And at the time thought, Oh, this [00:24:00] is particularly good. And so it just meant that by having a scrapbook of useful stuff it really helped when trying to gather. I think the other area that springs to mind is around lab meetings. So I was a member of a lab, a computational neuroscience lab, and we would meet every Thursday.

At lunchtime, bring food, and someone would give a presentation, and people would pepper them with questions, ideas, critiques and it was just an incredibly positive atmosphere, and everyone knew that they were going to get they were going to get critiques, but also suggestions and ideas, and it all came from a very positive place, and, It was a great way to notice problems early, to have the whole group, group mind focused on a single problem.

And, it was just enormously good fun as well, and a great way to learn. And so I ended up running lab meetings for years at Channel 4, where people would present on their latest work. And we would have [00:25:00] almost exactly the same atmosphere. And the hope was that we were all doing better work as a result.

Rob: And there's so many levels to this. The work is not just better, but it enhances the team. It enhances people's capacity to give and receive feedback. The benefits are enormous. And it's so easy, isn't it? A lab meeting. Any team can do that in any department, I would say. . It's a really cool journey.

Neuroscience to Channel 4 as a founder. If you were to write a business book about this journey, what do you think you'd call it?

Greg: I thought about maybe calling it make more mistakes than everybody else because if you can make more mistakes than everybody else and survive them and learn from them That can become its own kind of superpower

I suppose there is a saying that the I don't know. The wise man will, a person will learn from their mistakes.

The really wise person will learn from other people's mistakes. So it would be a finer thing to not have to make [00:26:00] mistakes in the first place, and to be able to rely on, on kind of careful reading of evidence. But yeah make more mistakes than everybody else might be. The ethos I've been living by.

Rob: That's, that's something I feel I'd write a book about. we can all benefit from. I even think for myself, one does sometimes play it a bit too safe and you're not finding out the new cool stuff that push your field forward. Just because you've got to deliver something, it's got to be good for the client.

So you don't try the new things. You don't have the capacity to have a dud workshop.

Greg: So do you have a forum in which you can experiment and potentially drop the ball? Hmm.

Rob: There are forums that do exist, and I could appear more in those forums.

Greg: Okay, don't need to put you on the spotlight there.

Rob: No, it's good. I need to hear it. So, in front of all our listeners, I shall appear more in aforementioned forums. Okay, let's make it practical. Let's just say that [00:27:00] we've got a friend here. They are a leader of a team, let's say there's 40 people in their team and they feel that they need to upskill the team. They haven't been too precise or about what they mean by that, but what would, in their situation, they've come into a role, big team, they feel it needs to be upskilled. How do you approach that in the first, let's say 90 days that you're in that role?

Greg: Well, the first thing to observe is that if you come in directly without any trust or much credibility, and immediately try and change how things work and tell people to do things differently, there'll be an immune response and that people will reject that. So, maybe step one is build some trust. Let's assume you've done that.

Then well working, pairing with people, which doesn't have to be programming, there's lots of ways to pair with people on a task, [00:28:00] is a great way to see what what skills exist and where there might be gaps in the, in the team. And so you might start to keep notes on that. of ideas for potential training sessions over that first month.

Because that first month is very precious when you have fresh eyes and you notice everything. So keeping a log of all the things you notice with your fresh eyes, because in six months they'll have become invisible to you. They'll have just become background that you no longer notice. And encouraging the team to start gathering a sort of scrapbook of useful materials.

Stuff they've read that they think is great that they'd like to share with each other. All of these are quite low friction things. No one's really gonna resist that. And then finally you could start to introduce some weekly rituals. Or monthly rituals, which could be a lab meeting, it could be some kind of journal club where you read a paper and discuss it, or you could graduate to a more intense thing where you set aside an entire day and everyone's focused on, on learning and [00:29:00] development.

But I wouldn't necessarily jump straight to that. I would look for I'd look to graduate it starting with, starting small.

Rob: So essentially, at first gathering the data to understand, gathering the resources and doing that as a group, what's cool about that is you also understand what are people interested in anyway, where is the motivation, and then building the habits. I think about when I started working with a trainer.

for fitness. And he was adamant, we start with 20 minutes that you can do at home. We don't start going to the gym, we do the easy bare minimum to get you into the habit, because I cannot do anything with you until you have the habit. And it took a long time to graduate to going into a gym and doing stuff like that.

It's got to be easy. Now this is all cool, and I think it builds momentum. Do you have a framework or questions that you [00:30:00] ask to identify what skills give the biggest impact to the business.

Greg: Well you asked that question as if I might have a one sentence answer for you.

Rob: I don't think anyone does, to be honest.

Greg: Okay, well, I think it is the 64 million dollar question because we can make the team twice as effective or even 10 times effective. But if we're working on the wrong stuff, then it doesn't make the blindest bit of difference. So there are two or three things that have helped in the past. One is if you haven't already Googled for product discovery conversations. So ways to ask people questions about what they would value, that don't presuppose the answer, and that encourage honest feedback, because most people will just tell you what they want, they think you want to hear. The second would be If you've got a team, we ran secondments.

So we would take some of the junior developers or data [00:31:00] scientists or whoever, and second them to other parts of the business to add sales or forecasting or scheduling. These are different departments in channel four. And they spent three months with those other departments walking miles in their shoes, getting to know them, going out and doing their jobs.

And then, as a result, they built really strong relationships with them, and they could come back and talk to us, as a team, about what were the real problems that those teams faced. And because they understood what would be valuable to fix, and because they were trained as data scientists, they also knew what was feasible, they could very, very easily, They could very intuitively give you a sense of there's, there's, there's, there's sort of juicy fruit that we could, that are low hanging in this direction.

And you've got the buy in because especially if you've got nice people in your team, then when they start to suggest ideas, it doesn't come across as somebody from outside telling people how to do their jobs. So that [00:32:00] ended up being extremely effective half the time. Which is not a bad hit rate.

And by extremely effective, I mean projects came out of nowhere and then landed enormous major projects that came directly out of, junior, junior data scientists hanging out in random departments.

Rob: That's, that's an awesome, awesome way to go about it. Just to stick with the point you made about someone external coming in and telling people what to do. So many people, when they think about learning in corporate is we have a budget, we spend the budget and that's learning. And I'm wondering, I'm sure you've You've used some external providers or you've bought programs or online platforms or whatever.

And I'm wondering, where do you feel is the most effective place to start allocating your budget? Let's say they've got the basics in first. Now they're thinking, okay, where do I spend money? How should they look at that?

Greg: So we can start [00:33:00] small and say that If you actually read a book rather than just sleeping with it under your pillow, then that 20 quid and those few hours are amongst the highest returns on investment. So picking really good books, and we can talk about how do you pick them, recommendations from people you trust is probably the best approach, but I would start there.

I've been on a couple of training workshops that have meaningfully changed the way I operate. And that was a particularly good one on difficult conversations. It really helped that I had a difficult conversation that I had to have two days later. So I got a natural spacing effect. And so a really great workshop, ideally one that someone you trust has been to and says, yeah, this is, this is super can be a good use of money, but only rarely, most workshops in my experience.

aren't. And so you've got to pick the good ones. And then find a way, perhaps if you go on them with multiple people, find a way to try and practice them afterwards. So when I run any kind of [00:34:00] training, one of the things I'll do towards the end is to ask people how are you going to apply this in your daily job?

Think of one or two examples. And then I'll even perhaps ask them to schedule in their calendar. Okay, on Tuesday, I'm going to try and apply this to that particular task that I have to do. Because there's evidence that this kind of scheduled pre commitment massively increases the likelihood that people do indeed follow through.

So I suppose I would be, I guess I trust recommendations and I don't have a strong view on whether it should be books or training or something else but I think there's ways to then make the most of those things you spend money on. I'm personally a bit skeptical about conferences other than for like meeting other people so I only go to conferences if I'm speaking but I'm sure there are people who get value from them.

Rob: Yeah, absolutely. Okay. And last question on this. Measuring training is a massive topic of discussion. And I'm wondering, what signals are you [00:35:00] looking for beyond happy sheets?

Greg: So I spent a long time thinking about this and made very little headway. I suppose if it's training that I'm running, then I have a pretty good intuition for whether it's landing. And indeed I can then see in general whether there are effects on people's behavior over the next month or two.

They're often not immediate. It takes a while for things to sink in. And ultimately that's the litmus is effect on behavior. I suppose If people are able to concretely generate examples where they are going to use this in their, in their work soon, either during or right after the workshop, that's a good indicator, I would have thought.

Rob: That sounds perfect. And I feel that workshops, if it is a workshop, need to have that space to generate examples and assign moments. [00:36:00] and schedule calendar events to do things. And it's also why I think you shouldn't wait for your team to not be busy to schedule training. In fact, you want them to do it while they are kind of busy, because you want them to use what you're, what they're learning, and then apply them and reapply them using all of the principles that you've spoken about.

Awesome. Right. I, I, I pushed you there because this is something I'm so interested in myself and I work with so much. I'm going to go to some rapid fire. So just short answers, whatever you feel and I'm know you've got expertise in AI and learning. So let's start basic Chachi, BT or Claude?

Greg: Claude,

Rob: Okay, cool, best resource for learning AI prompting

Greg: the anthropic prompt engineering docs are pretty good. Yeah start with those.

Rob: Okay, I'm gonna check that out because I feel my [00:37:00] prompting needs to get to get better Have you is there a skill that's going to become more critical as AI advances?

Greg: Managing AI is a lot like managing Another human being in the following respect.

So as Chief Data Scientist, I was responsible for the output of my team. I was responsible for making sure that our models were correct, that the hundreds of millions of pounds worth of advertising and predictions they were making, the forecast, that they were correct. And if they were wrong, it would fall on me.

And yet, there were too many people, too many lines of code for me to be able to look at each of them. And so I felt kind of terrified that we might be churning out something and I couldn't I wouldn't notice it. So practicing ways to manage in the face of that uncertainty ended up being a really useful skill.[00:38:00]

For example, I got really good at figuring out, how will I know if we've succeeded? I write tests for what the correct behavior should be? Can I help the team think through how we're going to evaluate our success? And if you do a really good job of all of that, and the model or the output succeeds on your evaluation criteria, and you've thought through the edge cases and the nuance, then at that point you can feel pretty good, probably.

about what it's actually doing, even if you've never looked inside the black box. So that was a digression, but I tell you that because taking some of the same views about AI, how can I evaluate whether the output is good, is probably skill that we learn with humans, with managing humans, that ends up being extremely helpful with AI as well.

Rob: That's fascinating and this might be you might need to consult your scrapbook, but is there a scientific paper that [00:39:00] Changed how you think about learning?

Greg: I'll go with the very first thought,

which is actually a lecture by Clay Shirky. And he started to count up the number of hours that we spend watching advertising in a given weekend. As a nation or a or as humanity. And I forget how many hundreds of millions of hours of advertising we watch each weekend.

And then he counted up how many hours it took to write the English language Wikipedia. And he basically concluded that we could be producing mini Wikipedias per month. If we, if that's how we were spending our time instead of advertising, watching advertising. And that was one of the inspirations for, for Memrise.

And how can we crowdsource this enormous human potential in a way that will help us all benefit a little bit? How can we direct the flow towards something that, that benefits us all rather [00:40:00] than fritters it away?

I think it was called Gin and the Cognitive Surplus. And so it's not a scientific paper, nor is it, directly about learning. And so I feel like I'm not really answered your question, but that, that was the thing that most directly changed my, my thinking about the potential for, for, for products that could, and ended up becoming a product about learning.

Rob: Awesome. Now I'm definitely going to watch that for sure. Last point to come to is the listener challenge. So in this part of the pod, we give listeners a short exercise or a ritual that they can try out for a week to get a little bit of your superpower. So I'm wondering, what would you recommend people try out?

Greg: So my, my favorite Keystone habit for teens. Is to run a really good blameless five wise post mortem. And I wrote a post about this on, making data mistakes. com. And [00:41:00] the reason I'm so in favor of these is because a really good post mortem where I don't know, maybe there was a problem, something went wrong.

Customer's unhappy. It could be anything. A really good post mortem. will help you get an understanding of all the 19 things that probably were inadequate about your system, your processes and it's never just one. And there are multiple levels. Yes, this person made a mistake, but maybe that's because we have the wrong culture, or maybe we don't have enough training, or maybe the system should have caught it, or maybe it should have been automated.

There's always multiple reasons. And if you pick one or two. Of those, those causes and fix it and commit to fixing it. And you do that every time something goes wrong for a while. You will very quickly see that the amount of firefighting you have to do goes down, and it's just an incredibly beautiful way of learning more about whatever it is that you're trying to be good [00:42:00] at as well as creating a lovely team atmosphere of continuous improvement and supportiveness, and psychological safety.

So that would be the, I think, most transformative Keystone habit for teams that I could think of.

Rob: Awesome, the five wives postmortem. Greg, where can people go to find out more about you?

Greg: MakingDataMistakes. com is where I try and compile all of my favorite mistakes and hopefully what we've learned from them.

Rob: Awesome stuff. It's been really fun talking to you, Greg. Thank you for coming on.

Greg: Thank you.

Rob: That was Greg Detra. Fascinating stuff. And what struck me most from the conversation was how it's often the elements around the learning that matter more than the learning intervention itself.

Because that science of memory and recall is important, but the foundation of trust, the sense of community, and creating a [00:43:00] culture of feedback within a team, that is what creates lasting change. If you found the conversation valuable, please be sure to write a review wherever you listen to podcasts. And if you know someone who's looking to upskill their team, share this episode with them.

I think they're going to get a lot of value out of Greg's insights. My name is Rob D. Willis, and you have been listening to Superpowered, the unique stories of modern leaders. Thank you so much for joining me, and I will see you next time. Goodbye.