Based Camp - September 06, 2024


Introducing the Collins Institute: Revolutionizing Education as We Know It


Episode Stats

Length

25 minutes

Words per Minute

199.46265

Word Count

5,073

Sentence Count

231

Hate Speech Sentences

1


Summary

In this episode, Simone and Malcolm talk about their new educational tool, The Kalanith Institute, and how it can compete with existing educational models. They discuss how to use it, how to get started with it, and why they built it in the first place.


Transcript

00:00:00.000 Hello, everyone. I am so excited to be here with you with Simone today.
00:00:04.160 This is going to be an explanational video, first on how to use the Kalanith Institute,
00:00:10.320 our hopefully sort of competitor to existing educational models, and then the theory behind
00:00:17.320 how we built it and why we made the design decisions we did, because what is coming to
00:00:22.460 the public is evolved a lot as a model from what we originally thought we were going to make.
00:00:28.640 So the first thing before we go into the school system itself, if you're like, okay, what
00:00:35.320 is this? Is this a replacement for public school? Is this a teaching aid for people who are already
00:00:41.640 in the school system? Is this just for kids? Or is it also for adult learners? The answer
00:00:46.680 is, in terms of how you contextualize it, the role it's filling in your head, is sort of
00:00:52.140 all of those. Think of it like Wikipedia, but transformed into more of a directed learning
00:01:00.920 engine. So a question with something like Wikipedia, is Wikipedia something that helps people who are
00:01:06.860 already in school? Yes, of course. Is Wikipedia something that would help a homeschooler? Of
00:01:11.300 course. Could a person realistically educate themselves to quite a high level using Wikipedia?
00:01:16.060 Yeah, they could. It's just not really designed to be used that way. So you can think of this as us
00:01:23.540 redesigning a Wikipedia-like system or collation of all human knowledge so that it can be used in
00:01:29.880 a directed way. And we have tried to include everything in this, you know, from how to lay
00:01:34.780 drywall, to aquaculture pharmacology, to tax policy as it relates to investments, to what you would study
00:01:43.000 in a subject from when you can read in that subject to midway through a PhD in that subject.
00:01:48.120 I'm just incredibly excited about this tool. And both Malcolm and I wish that we had access to this
00:01:52.700 when we were kids. Oh, yeah. Yeah. No, as somebody with an MBA from Stanford and an undergraduate degree
00:01:59.200 from St. Andrews, which are two really good schools, I would have been so much better educated had I been
00:02:06.360 using this platform than the educational system I actually went through. Yeah, to a great extent,
00:02:11.180 it combines a textbook with Wikipedia in the sense that when I used to study, I would go through a
00:02:17.340 textbook's quizzes before I would go through a chapter because I found that I was a little bit
00:02:21.160 more attentive in my learning when I was wondering if I got questions right or if I got a question
00:02:25.840 wrong. And that is something you can do with this tool. Our goal with the Collins Institute was to
00:02:30.100 divide all human knowledge into a skill tree like you might see in a video game. You progress through
00:02:36.080 this skill tree unlocking nodes, which gives you the ability to tackle more advanced nodes. Here you can
00:02:41.920 see the classic vertical style skill tree like you might see in a video game. However, if you prefer
00:02:47.840 other form factors, we also have a radial view where the skills all radiate out from a central node
00:02:53.680 and a collapsible view where you can collapse all of a single skill tree path using this double arrows
00:03:02.160 or specific pathways using this single arrow here. Today, we will be focused on the vertical view
00:03:09.040 because, well, frankly, I find it to be the easiest to use. To navigate this tree, you can click and hold
00:03:14.720 to drag up or down. You can scroll in and out using your middle mouse button. And if you are on a tablet,
00:03:21.360 you can use this at the bottom right of the screen to control how far zoomed in and out you are. And
00:03:28.000 on the far left bottom of the screen, you can use this thumb joystick. When you see a skill you want to
00:03:34.320 attempt to master, you can click the skill and then click see more to go into the skill. Within the skill,
00:03:41.120 you will see a description of what you need to know to pass the skill. And below it, you will see a list of
00:03:46.400 the best sources online for studying this subject. If you click add a source right here, you can add
00:03:52.000 your own. You can upvote and downvote sources here. If you notice that any of these sources have issues
00:03:58.000 with them and should be reviewed, you can click right here and flag the source for our admin team.
00:04:03.280 You can also offer yourself as a tutor in this particular subject if it is something you like
00:04:07.760 talking to others about. When you are ready to take the assessment, you can click the take the
00:04:12.880 assessment button up here at the upper right. It will give you eight questions that is not meant
00:04:17.680 as an absolute test of your skill, but more to keep you honest with yourself about whether or not
00:04:21.840 you actually took the time to study this particular skill. Right here, there is a question bank button.
00:04:28.240 This can be used if you wanted to edit the question bank for this skill. And up here, there is an edit
00:04:34.400 button which can be used if you want to edit the mastery requirements for the skill. Note if you are a
00:04:39.760 student, your edits will be put in a queue and then an admin will review them and approve them if they
00:04:46.880 are good. However, if you would like to become an admin, you can always just reach out to us and join
00:04:52.480 our team that is helping to improve this tree. If you want to, you have the ability to review skills
00:04:57.760 that you haven't unlocked yet. However, when you open one of these skills, instead of seeing an option to
00:05:05.280 take the test, you will see an option to go to the nearest unlockable skill. Just click this and it
00:05:10.880 will take you to the nearest unlockable skill within this test tree. Here you can take the assessment and
00:05:16.960 eventually unlock the skill that you were looking at. You will notice a few different types of skill
00:05:21.280 pathways. Some are simple like the ones we just looked at, but occasionally you will get to skills
00:05:26.960 that are called cluster nodes. Cluster nodes are when a skill involves a number of sub-skills to pass.
00:05:34.880 Here are some good examples here of the Roman Republic. So when you get to a cluster node,
00:05:39.760 you will notice this dot dot dot line that leads to the central cluster node and then the outer cluster
00:05:45.280 nodes. To pass a cluster node, you actually need to pass the outer cluster nodes before you can pass the
00:05:51.440 inner cluster node, which will include questions from all of the outer cluster node tests. This becomes
00:05:56.800 particularly relevant when a cluster node must be passed to pass the nodes above it as we see
00:06:01.520 right here with physics and chemistry 1. Here you would need to pass the properties of water, states
00:06:07.760 of matter, matter, basic concepts, basic reactions, volume, mass, weight, simple machines, and measuring
00:06:13.600 systems before you could take a test that covers all of those subjects and moves you to the physics and
00:06:19.120 chemistry 2 section right here. Once those are done, you can then unlock the center of the cluster node
00:06:25.600 here, which then unlocks the next stage of this particular subject, physics and chemistry 3.
00:06:30.400 Occasionally, you will come to a section in the tree where there is no node. Right here you can see
00:06:35.280 this at the branching point of physical science. When you reach one of these, these can be thought of
00:06:40.400 as nodes that are automatically passed and used mostly to help sort things. For example, right here,
00:06:46.000 you can see this sorts down into repair, building, and maintenance, which goes into things like
00:06:50.800 mechanical drawings, cutting things in electrical circuits, but is still technically under the
00:06:55.760 physics tree. However, it doesn't need basic physics like the properties of water and the
00:07:00.480 states of matter to move forward with it, so we separated it out. All right, we're excited to have
00:07:04.960 you guys back. I hope you enjoyed learning about how to use the product. There really isn't anything else
00:07:11.520 to it right now, unless you're an admin or an editor, in which case the user interface is a little
00:07:17.200 different, and you're going to have a few extra tools, but it's just as intuitive of what we just
00:07:22.800 showed you. I mean, I'm sure as we went through that, you're like, I could have figured all of
00:07:26.160 this out just by looking at it, and that's what we were trying to do in terms of building this.
00:07:30.480 The big design changes that happened as we were building it were a few fold. In the initial
00:07:36.800 system, we had tried to build out something that was very good at measuring a student's actual
00:07:43.280 progress in the moment. So by that, what I mean is like school. When you get a test at a school,
00:07:50.160 right, it's meant to judge your ability at that particular subject. However, what we realized is
00:07:57.280 that these are very not fun tests to take. Like a 50-question test or something like that,
00:08:02.320 or a long written assessment is just not very fun if we're treating this as a game loop.
00:08:07.360 Then the second problem is, is they're super cheatable on. Even if we use like an integrated
00:08:12.960 camera on a student's computer, which first of all felt like a huge invasion of privacy,
00:08:17.040 the student could just use AI to cheat on them. And they're moving into a academic environment where
00:08:22.640 that's going to be increasingly common. And so I sort of took a step back and said, hey,
00:08:27.600 okay, we're now in a world where it's very hard to judge somebody's objective understanding of a
00:08:33.440 subject. What if instead we made the test more about keeping a student honest with themselves
00:08:41.200 about whether or not they actually know the content of a node instead of us checking if they know the
00:08:48.320 content of a node. And then I realized that this completely changes the student's relation to the
00:08:53.280 educational system. It's not adversarial. It's not checking if they're constantly lying, basically,
00:08:59.440 or trying to deceive it. It is that they are to help them. And the relationship, the cognitive
00:09:05.200 relationship that is going to lead them to form to education as a process is going to be quite
00:09:10.320 different. Also, it allowed us to narrow these tests down into like short eight question tests.
00:09:14.640 That flexibility now makes the game loop of the testing process much more fun for students.
00:09:22.560 As we have seen in our own usage of this and much more viable for adult learners as well,
00:09:28.720 because now for an adult learner, you're not dedicating, you know, 20 to 45 minutes on a test
00:09:34.960 for every individual node you want to fill out. On the original system, we also had a decay
00:09:41.440 feature where the nodes, which actually sort of you'd lose them if you didn't keep working on nodes
00:09:47.440 within that tree or revisit them occasionally. And this again, just was not very fun from like
00:09:55.040 a game design perspective. There was a game recently, I was looking at vampire something
00:09:59.440 that had a similar feature with base building. And it was just incredibly unfun because you felt like
00:10:03.600 your accomplishments were not tangible. Like they just, they just fell away as time went on. And I
00:10:11.280 didn't want that feeling for students. If students want to redo a nodal tree, they can.
00:10:17.680 Might we add some voluntary decay feature eventually that teachers or students can add themselves?
00:10:23.920 Sure. But I think we also found that like making, making accomplishment ephemeral doesn't necessarily
00:10:29.360 help. Especially if someone's still advancing, because if you're advancing, you need to know
00:10:33.440 foundational knowledge anyway. It's not like if you get really far along in Roman history,
00:10:38.480 you're going to forget the very basics because they're going to be so solidified in your understanding
00:10:42.400 of more complex subjects. So, because most of the knowledge that you're developing through a
00:10:47.520 skill tree like this is going to be building on the more foundational elements, you don't have to
00:10:53.280 worry about decay as long as there's advancement. Yeah, you're absolutely right, Simone. Another
00:10:58.240 big thing that changed as we were going through this is we initially planned for students to have
00:11:03.760 like a bank of paid tutors that they could access through the site. And then the tutors are ranked
00:11:09.360 based on how much they helped students within an individual test. We moved away from this model
00:11:15.440 to the free tutoring model for two reasons, right? So right now, the way it works is you can just sign
00:11:20.560 up as a tutor and people can reach out to you until you want to take your name off as a tutor. And they
00:11:24.640 can also upvote and downvote you so other people can see this is one because you actually do help learn
00:11:29.440 more through tutoring other people to an extent. But I suspect the primary way that people are going
00:11:34.400 to interact with teachers who are using this platform is twofold. One, they're going to be
00:11:40.640 challenged by something and go to their parents or whoever is in any way moderating their learning.
00:11:46.000 So it might be, you know, they might be in a teaching group or something like that. They might
00:11:49.280 be in a homeschooling group or pod or something like that. Or they might go ask their friends who know
00:11:54.000 a lot about a subject. And that to me felt much more organic. Or they may ask an AI. And when I
00:12:00.880 look at the speed to which AIs are developing right now, I kept thinking to myself, like,
00:12:05.520 is a student going to get better instruction from an AI like perplexity even today than they are from
00:12:12.720 a low cost teacher that, you know, we were able to find and book at, you know, on demand on a platform
00:12:20.640 like this. And it was just like, obviously they will. And better yet, the AIs are on demand,
00:12:25.920 right? Like the teacher, you'd have to wait a day maybe for a reply. Whereas the AI, you don't need
00:12:30.320 to organize anything, you know, and it's consistently getting better. So we decided to
00:12:35.200 lean more heavily on students will figure out how to access instructors. And then we also thought,
00:12:40.960 but that's also an important part of the learning process. Like it gives students a lot more agency
00:12:45.520 if there isn't a, this is obviously what you do if you're struggling with something,
00:12:50.160 instead of a, you need to figure it out yourself. Because, you know, that's what we're teaching in
00:12:54.800 an educational system. And we also found that a lot of the advisors that we worked with
00:13:00.800 really found most of the students they tutored learn by teaching themselves. So the tutoring that
00:13:06.240 we offer is almost more as a helping you learn a subject kind of option than it is a, here's how you're
00:13:13.840 going to learn this subject. Obviously the tutoring is going to help both people involved in most cases,
00:13:18.080 but we've discovered that teaching people something you've just learned is also a really
00:13:23.360 important element to something to build into this, this platform.
00:13:26.240 Yeah. And if you are somebody who teaches like a pod of students or a group of students,
00:13:31.280 and you want to use this platform, we actually have a special type of account for you that allows
00:13:35.840 you to monitor all of the students in your pod. And you will be able to check yes on certain tests
00:13:42.000 for them because while a lot of the tests are multiple choice, some are written answered and
00:13:45.440 some are picture based like the food ones or the making friends ones. And so they need you to
00:13:51.360 approve those particular answers. One of the first updates that we're going to be dropping
00:13:56.560 after, after launch. And we're working on this right now is to have an AI that can handle grading of
00:14:02.240 the written answers and the picture based answers. So if a student wants to approach this without
00:14:07.520 any form of teacher, they can. Right now, when they get to those types of questions,
00:14:12.000 if they don't have a teacher assigned to their account, they can just create one,
00:14:14.960 basically reach out to a parent, use their email. Or if you're an adult learner, use your own email
00:14:19.360 to create your own basically teacher account. But we want to move away from that and have the whole
00:14:23.200 system be automated as soon as possible. But that doesn't mean that the system isn't going to work
00:14:28.000 with outside instructors or that you as an outside instructor aren't going to be able to see
00:14:31.600 what students are doing. Another thing that we move to is it's much more self-generating than we
00:14:39.920 had originally planned. Whenever a student takes a test, they get an opportunity to suggest a new
00:14:44.960 question for that test, which also sort of tests their knowledge on something. Eventually, we'll be
00:14:50.000 giving students a secondary grade based on the questions they're suggesting, as well as the edits they
00:14:56.160 submit. So unlike Wikipedia, what you do is if you want to make an edit or you are uncomfortable with
00:15:02.480 something there, there is a flagging feature and there is a editing feature and then it just needs
00:15:07.040 to be approved by an admin. So we don't do the whole, you have to have a source for everything here.
00:15:12.640 We try to keep a competent enough pool of admins within every subject domain and feel free to reach
00:15:18.400 out if that is something you're interested in participating in, that you will be able to know
00:15:22.880 if the students' changes are good or not. Originally, I was going to hire PhDs to fill out all of these
00:15:29.920 trees and all of these questions, but we moved to doing it with AI because I just feel that the user
00:15:34.880 generated feedback that we're going to get is going to so quickly get to a quality of content that we
00:15:43.760 would have gotten from PhDs, but at a much lower cost and I think much higher readability content.
00:15:50.560 Because the problem with traditional PhDs, and I know this because I was going down the academic
00:15:54.960 path, is you learn to write descriptions and everything like that, like an academic, like
00:15:59.360 you're writing research papers, and that can be very hard for a layperson to read. And I wanted to
00:16:04.240 ensure that we didn't have that style of writing on this stuff, which is why we moved to a public
00:16:09.920 collection mindset. Another thing that we moved away from was the banding thing. We initially wanted to
00:16:16.240 have a system in here that prevented students from becoming non-well-rounded. So basically you
00:16:21.440 needed to get a certain number of points to stay in the school every quarter, and those points were
00:16:26.160 determined by how far behind you were, was in a branch of the tree. As to why we threw this out,
00:16:31.120 basically the only people who wanted this were the people who didn't want to use our system.
00:16:36.480 So it was like existing public school educators who were like, well, how are you going to deal with
00:16:42.000 students not being well-rounded? When we actually went towards the type of people who were like,
00:16:46.160 public school isn't working anymore, I want something else for my kid, and we mentioned
00:16:49.200 this system. They're like, I hope you can turn that off because I don't want that for my kid.
00:16:53.520 Like, I don't want somebody telling my kid what they should and shouldn't be learning. I just want
00:16:57.280 a system that's going to completely allow them to go with their passions because we're moving into a
00:17:01.680 world where well-rounded is the first thing that gets automated. Well-rounded is the first thing that
00:17:07.040 gets outsourced. It's the lumpy skill sets that ensure employability in a post-AI, post-
00:17:11.920 outsourcing future. And I was like, yeah, you got a really good point there. What was I thinking?
00:17:16.880 Another thing that we have thrown out is at first we were really afraid of bias. And I was like, well,
00:17:22.480 you know, we want to have true impeachability on bias. So we'll have students compete in
00:17:27.200 prediction marketplaces or do real world tests, like post a story, a scary story on Reddit or something
00:17:34.080 like that as a judgment of their English skills to see how many upvotes they got. And then use that to
00:17:38.960 modify which questions we were keeping in and which questions we were keeping out. And then
00:17:43.760 I had this realization of one day we might build a system like that, but there's so many negative
00:17:50.640 externalities that could accidentally work their way into the system due to this. And it's just
00:17:56.480 incredibly heavier to build a system like this than to just have editors in a version history like you
00:18:01.920 have with Wikipedia. And so I was like, why am I adding all of this complication when it's really
00:18:06.320 mostly an aesthetic difference? And so I decided I was like, yeah, that's, that's a waste of time
00:18:12.480 for everyone. So, and also you can't automate it as much by that. What I mean is, you know,
00:18:17.680 you would need some sort of person to go and check how they were doing on Reddit to review the topics
00:18:23.200 that they were suggesting to review the authentic tests they were suggesting to create the prediction
00:18:28.560 markets. And I was just like, well, it's so much better if you don't have that because then it's so much
00:18:35.040 lower cost to operate. Like if we can get to cost zero operation per student, you have created such
00:18:41.040 an astronomically better system because one it's operating on demand for every student. They don't
00:18:46.560 need to wait for like a cohort of students to reach a particular stage in a tree to then access a
00:18:51.520 prediction market to gate them moving forwards. Well, and we're also, you know, as parents of four
00:18:56.480 kids, hopefully soon more, we understand that the simpler something is the better. And that while it
00:19:04.560 can be nice to think of all these elaborate, lovely things you can do for your children in the end,
00:19:09.440 what will get done is what is simplest. And sometimes you can go above and beyond. And sometimes you can
00:19:14.080 make authentic assessments for your kids. Sometimes you can encourage that, but that is the icing on the
00:19:19.280 cake and something that you can do on your own really well and very easily. Whereas what we want to make
00:19:23.520 sure we deliver is something that works without extra intervention and that doesn't give you or
00:19:29.280 students a lot of unnecessary busy work. So that is important. And then somebody is like, well, what if,
00:19:35.040 what if I, as a parent think that something in the tree is really biased, right? Fortunately,
00:19:39.120 still have a solution for that. You can just censor those skills. You know, if you're going through and
00:19:43.120 you're like super progressive and you don't like that at the end of the civil rights chain on the tree,
00:19:48.320 there's one on the scams that were run using the black lives matter movement. And you're like,
00:19:52.000 I don't want my student to see that. You can just check that off. Or, you know,
00:19:55.920 if you don't want to see like the evolution of human sexuality, you could just check that off.
00:19:59.680 Right. And then, and then you'll know, based on what you've checked off,
00:20:02.800 what you need to personally teach the student that could have introduced some bias to the student.
00:20:08.240 And so there's, there's just not the same risk of bias from the perspective of a parent,
00:20:12.640 because there's only so many subjects that can really introduce sort of systemic bias to
00:20:18.400 a student and those you can search for yourself and, and remove.
00:20:23.120 Perfect. The, the final thing is if you're like, oh, I, I hate all this AI text here and stuff like
00:20:29.120 that, then work on it yourself. And as AI gets better, we'll replace chunks of it. We could really
00:20:33.920 use admins for this. So anyone who's like, I really am dedicated to doing lots of edits for this.
00:20:39.280 We'll look over your edit suggestion history. And if it's good, then we can make you an admin,
00:20:44.480 because I do want this to grow on its own. I want this to be fully operational for my kids by the
00:20:51.280 time that they're fully reading and can engage with a platform like this. And if you're using it
00:20:55.760 and you're like, oh, this node should be here, or you should have a lot more nodes on this subject.
00:21:00.560 Every year, we're going to be doing a thing where we add a lot of nodes and prune some nodes. We'll call
00:21:05.360 it like the pruning, the yearly pruning that will be made. It's sort of an executive level decision
00:21:10.880 based on feedback that we're getting. We may come up with a more organized way to do feedback
00:21:15.440 or even vote on this stuff. We just don't know what that's going to look like right now.
00:21:20.560 If you are an existing educational institution and you want to use this system, we are happy to
00:21:25.840 work with you to white label it or anything like that. You let us know. And if you're somebody who
00:21:30.480 wants to donate to keep this free, because right now we're trying to launch it free. At first,
00:21:36.160 we had this goal where we can turn this into an alternate source of income and then we're not
00:21:41.840 going to have to worry about being fired from our jobs for saying things that are against the
00:21:50.480 current political or social establishment or too controversial. And we basically got to a point
00:21:55.840 where I was like, yeah, but I really want this to be free and I'll find another way to support myself
00:22:01.040 if things go wrong. But if you want to donate to ensure that this stays free to people,
00:22:06.320 you can always reach out to us about how to do that. Or if you want to put the foundation in your
00:22:10.960 will, that's always a nice thing. I'm going to start suggesting that more because that's
00:22:15.440 a good way to without burdening you today, help keep this free for future generations.
00:22:21.840 Anyway, I absolutely appreciate all of you so much who have made this happen. I have been working on this
00:22:27.680 for years. I wanted to ensure that the product that got to you was something that is high quality
00:22:34.240 and usable in its alpha form. But keep in mind, you know, this is an alpha. It is an active development.
00:22:40.640 Parts of the UI are going to be changing. Parts of the system are going to be changing. Features are
00:22:46.560 going to get added pretty much constantly over the course of the next few years. And so, you know,
00:22:52.800 you're seeing now it took us about two years of development time to put together. Imagine where
00:22:58.720 it's going to be in two more years of development time. This isn't like a work on it and finish it
00:23:04.080 game. This is something I want to constantly, constantly, constantly be developing and making
00:23:09.440 better because I think it's the only way we can really protect the next generation is to protect
00:23:13.360 them from the types of brainwashing that goes on in the existing school system. Were there any other
00:23:18.800 major features that we dropped, Simone? I think the most important thing is if anyone has feedback
00:23:25.600 that they would like to provide to us or they want to discuss this with other people, there is a
00:23:32.400 part of our Discord server associated with our podcast where you can talk about the Collins Institute,
00:23:39.200 specifically this platform, Parisia, on the Discord server. We'll include a link to the Discord server in
00:23:44.000 the description of this video. So if you want to chat with people about it as you're exploring it,
00:23:47.600 that's a good place to go. If you want to provide feedback directly to us, you can email us at
00:23:52.320 partners at Collins Institute dot org. Yeah. Oh, and I was going to add the one other thing that we
00:23:57.760 removed. One other thing that we had dreamed about putting together was when we had actually put it
00:24:02.320 together, a democratized nepotism, which is to say a network of high profile individuals that you can pitch
00:24:08.000 to and work with in various fields if you got really far within any branch of the tree. That network still
00:24:13.440 exists. I just couldn't find an easy and automated way to work it into the system. The way the network
00:24:19.920 works now is if you have a student who has worked their way to the end of any of the branches of the
00:24:25.040 tree, just reach out to us. We'll be able to see their user history. We'll be able to talk with them,
00:24:30.960 get a brief vetting of them. And then if they are extremely precocious, connect them with people in
00:24:37.920 most fields that can be useful to them. So that feature is still there. It's now just part of the
00:24:42.880 free version of the software until we get so big that I need to find some other system for doing that.
00:24:49.040 As a bonus, because the launch has taken so long, we actually got a few new features working. One of them
00:24:55.520 that I'm very excited about is now we have an AI that will grade written answers, as well as picture
00:25:01.520 based answers, which are common on things like the food test, like make eggs in four different ways,
00:25:06.560 or learn to make fried rice, or things like that, as well as the friendship based ones, which is like,
00:25:11.840 you know, show a picture of you and a new friend that you made. So very excited that we got that
00:25:16.000 working before it goes live, because that means that anyone can pass through the entire skill
00:25:19.120 seat without an instructor to be attached to their account, if they want to. They can, of course, attach structures.
00:25:24.720 write this out.