Hacking AI, Startups, and the Future of Software: AMA with Jason Mars

I've been on a while on 13, 14 years. Um and really, you know, I was always very curious as a kid. Um I always wanted to tinker, I wanted to build things. I got super inspired um to, you know, watching movies like Pirates of Silicon Valley and Antitrust when I was a little kid. You guys probably haven't seen these. Matrix. Um you know, and I wanted to engage technology. Um so, you know, as a student, I was kind of a, you know, I was doing many things as a student, right? I wasn't very focused on academia or focused on research. Uh you know, I was a musician. I used to uh DJ all across the country, open up for electronic music stars, and all kinds of craziness while in college. Um and then a professor discovered a number of my toy projects uh and was impressed and invited me to her lab, and that's where I discovered this pathway where you can tinker, innovate, and create. Uh and there was this almost institutional means for celebrating that. Uh and the real clincher for academia for me was the fact that I can still pursue my dreams. Um you know, one of the things you have to do when you join a larger organization is you have to essentially sign away all of your ideas and get permission. This actually is covered in Pirates of Silicon Valley when I was I don't know, 12 watching that where uh Steve Jobs had to beg HP for the rights to go build the personal computer. Um that's kind of how industry is, and academia lets you and celebrates going out and having impact even in the commercial realm, even in the industrial realm. And so, uh that's how I chose this particular pathway, this particular journey. That's I think you want to put the mic a bit up. Okay.

>> Yeah, no. Well, you know, honestly, there's a number of things So, there's a some massive tectonic events that happened recently. Number one, COVID. Number two, AI. And you know, one of the things that really I spent a lot of time thinking about that really bugs me, honestly, is the fact that students are not getting exposed to reality of how the world works, right? And and information as to what the pathways look like and what the tradeoffs are with those pathways. I think there's a number of interesting reasons why that's happening, but you know, I'm always incredibly excited to see like the work you're doing. And you know, you're kind of an enigmatic character that is actually connecting ideas to people that matter. And so, I think this is a like this is a really nice opportunity we have to share some of those uh those secrets. Happy to talk about those if you're Yeah. Yeah, yeah, absolutely. Yeah.

>> Yeah, so, you know, honestly, it's a mindset. Um and so, when I started academia, my number one objective function really is linked to, you know, my core perspective and principles and values, and it was to have impact. Um and so, you know, I I almost had this mindset that if I can't really have impact in the way that I believe impact should matter as a professor, then I shouldn't be a professor. And so, when I started the job, um you know, I focused on the bean counting type work that you have to do for a tenure track, right? Publish papers, sit on certain committees, etc. And I knocked that out very quickly to the point where I felt there was enough of a momentum in that category of work that I can quickly turn the impact we were doing as research uh into impact on the world and impact in people's lives. And so, I I essentially sped run. So, after maybe 2 years uh of of being a professor, getting acclimated to what that means, uh I decided to take some of the things we were innovating and turn that into real world impact. Uh and and that's what started this entrepreneurial journey that led to multiple companies.

I think it starts, honestly, with every student, every single course I create cuz I'm I'm creating a number of courses. The number one first step is to be genuinely curious. And if you let yourself be curious, let yourself play, let yourself meander into ideas and into topics, you're going to build enough of a perspective to start a conversation with any professor, right? And so, you start by being curious, getting to understand what is the landscape of things being worked on outside of coursework in the academic realm cuz that's really what professors do. We have to invent the future technologies as engineering professors. And then, whatever you're curious about, engage. Professors have meetings all the time where they they welcome minds with ideas. Uh that's the ethos of being an academician. And so, engage, communicate, talk to professors, mention the ideas that they're working on that you're interested in. Face-to-face is is huge. Uh and then join their group meetings. Go understand who the PhD students are. Ask questions. And very quickly you'll be snapped up, right? Curiosity is the number one attractive feature when it comes to a professor investing time and energy in encouraging someone to to participate in the research. And so, you demonstrate that, and through that you you know, folks will see how clever you are. Folks will see how you think. Uh and and that's how you'll get noticed, you know?

Yeah. Well, you know, honestly, one of the things we have to do as leaders and especially CEOs or founders, one of our biggest jobs is predicting the future, really. Uh and then you have to execute on those visions and insights. And it's actually not rocket science. It's not anything special. What you have to do is just observe observe the possibilities in the market and understand the inevitability of economic incentive. And so, you know, I always tell folks there's really two disciplines that's critical for being any kind of creator or um you know, innovator, and that is understanding the core of philosophy, how to know information, what is knowledge and what is logic and what is true, and then understanding economics, which is the incentive structures that create behaviors in the human species. And in both the micro level and the macro level, understanding the dynamics of those of those systems. And if you understand that, whatever is possible that can create that can create markets, whatever the market is thirsty for, is going to have implications on what people work on and what people figure out. So, in 2018 that TEDx talk, you know, I talk about being able to talk to AI like a human in a room uh and using messy language and it understands you. Uh understanding that intelligence to the human species is really uh other humans. That's how we understand it is through communication is how we access intelligence. There's no other way we access intelligence. And so, understanding that dynamic, it was clear that the problem to work on is getting artificial intelligence to that level. Uh and we were working on that problem feverishly well before 2022, but of course without the means of um you know, of a Google, which really invented the technology, and then of an OpenAI, which really harnessed the technology to create ChatGPT. Um so, we were doing it with small means, working with um you know, transformer neural networks right well before there was the generative pre- uh pre-trained transformer observation. Um uh and you know, we had some interesting successes, and now uh it's clear what the future will be. Uh and the future is going to be outside of the innovations in the model and in the innovations around the uh the technology and the engineering around the model. And so, these are things like the harness in in uh you know, the harness engineering that folks are starting to talk about, in-loop and out-of-loop agentic AI systems, really getting these models to perform specialized tasks well, moving from in-loop toward out-of-loop, uh which is kind of manifest with Open Claw and uh Hermes, which people aren't really talking about enough right now. Um you know, and um, and those kinds of uh systems. And so, um, yeah. I mean, so that's the, you know, that's the secret.

Well, there's a couple trends going on right now that is folks are not talking about enough, especially in this country. The number one trend, and this is a massive shift this year, is 2 years ago I used to talk about, oh, these open weight models and these small language models, etc., will catch up one day. Well, it they've already caught up um, months ago. And it's through models like Kimmy K 2.5, um, MiniMax, uh, M 2.5 and uh, GLM 5. Like these models essentially are performing like Opus 4.6, uh, you know, ChatGPT uh, 5.4, etc. They're competing and they're open. So, that's one important trend. There's two more quick ones. No, the second trend is Apple is orienting their business 100% around out of out of uh, data center AI, right? So, what uh, Tim Cook is saying is he wants to move AI out of the data center. And they're executing on it. There was an M5 processor that just came out uh, with 4X tokens per per per per second on your MacBook. But, what hasn't come out yet that's going to be a doozy is going to be the Mac Studio. And it's smart because then you can put an Opus grade model on your desk, control it, fine tune it, do anything you want with it. And there's clearly demand when you look at the the Mac Minis, the sharp uptick in Mac Minis when Open Claw came out, which isn't even a model, but just prompts around a model. That's two. Uh, and the third the third killer uh, observation is the generality of Open Claw itself, right? So, that's hardware, that's open models, and then the generality. In China, there's grandmas lining up, thousands of grandmas literally lining up. Uh, I think it's Alibaba is doing a program where they're going to help you set up your Open Claw. And everyone in China is trying to create their own Open Claw that can run at home as a buddy that can do things on the computer for you or as you. And so, that that shift and we we recently saw Nvidia's now launching Nemo Claw, and there's already Null Claw, and there's like six other claws. Um, and then there's Hermes. Uh, so, so, that trend it changes everything we knew about agents, right? We're outside of the land of Claude code now. Claude code is code is boring, open code is boring. What's interesting are the skills, which are really the facets of, say, your Open Claw. And Open Claw level architectures, which is beyond the previous notion of agentic AI. So, this is going to change the landscape fundamental. Um, yeah. Yeah. Basically, Nvidia [clears throat] is at risk. Uh, data centers are at risk in a big way. Um, so, we're going to see some really disruptive movements in 2027 from from what I'm observing.

I think the most is like the most concerning thing that's happening right now in academia is the fact that professors and the academic institution is moving so slowly to to re-understand what skills need to be held by students in the future. Um, you know, I I I teach primarily seniors and grad students, and almost all of them are wildly behind the realities of the market. Most of them don't understand that models can call tools. Like seniors graduating right now, um, and we are actually actively there's at least half of the uh, the academics that I I I colleagues I I work with, um, put a lot of energy into keeping models away from the work that students do and agentic AI away from the work that students do. So, students are coming out pretty ill prepared. I mean, I think the number one thing a a student should do is they need to learn on their own for now how to integrate these tools into their workflow. Um, you know, the Y Combinator CEO, what's his name again? Garry Tan. Yeah, yeah. Just released just released G stack. Uh, you got to play with G stack and understand why it ticks and why it works. Um, it really is um, a disruptive emergence of systems like G stack and then the way people are using skills now. And if you're not using skills, and if you if you don't have experience with models, uh, building agentic systems and wielding agentic systems, you're not ready. Uh, and so, that's what I would try to really brush up on right now.

Yes. What one of the So, Jack and Jaseki really spawned out of my observations around 2018 uh, and beyond. Uh, around we've got a massive, you know, Oh. the way we build software Oh. Uh, Uh, your mic is mute. Um, I don't know how to unmute it. Same here. Okay. Is it good now? I don't know. Yeah, yeah, it's good. Oh, my gosh, the mic was mute the whole time. Oh, no. >> But, he could hear us from Oh, you can hear me from your mic cuz I'm loud. Got it. Yeah, anyway. So, so, Jack and Jaseki uh, the programming languages haven't evolved since uh, 1989 is when Python was invented and 1994 is when JavaScript was invented. But, 90 like most of the most of the code we run today live as Python, JavaScript, or native code. Uh, and you know, w- there's like we currently have to make these massive decisions as to what kind of developer we want to be. Do you want to be a front-end developer? Do you want to be a back-end developer? What languages should you learn? What tools should you use? Uh, and I think that that's a fallacy. I think that we haven't done our work as uh, academicians and intellectuals to create the next level of abstractions to simplify uh, the way we build software. Back in the '80s, one person can create Mario, launch it to the world, and have great success. Today, one person cannot create the next Twitter, launch it to the world, and have big success. And so, Jack tries to tries to realize that vision. Um, and so, you know, what would take thousands of lines of code could be articulated cleanly in just 50, 60 lines of code in Jack. Uh, and Jack doesn't like Jack is interoperable with Python, JavaScript, and C ABI. So, you can use the same ecosystems, the same packages, any node package uh, can be imported into Jack and Jaseki. The reason I created is I created it for myself, right? Um, I don't want to learn Rust. I don't want to have no Python and Rust and JavaScript to to create a thing. I don't want to have to learn, oh, Kubernetes YAML files. I've got 80 YAML files. There's a break in like YAML files 57. And now I have to go into my Rust code base. Now I have to go into another polyglot problem. It It turns out that the technology exists and the insights exist to create one language that can span all of those spaces where every line of code relates to the actual thing you're building and not what's underneath the thing you're building. Uh, and that's what Jack uh, has achieved. And so, we're seeing incredible results. There's number of um, companies that are building around Jack. Um, and there's a number of open source software and products that uh, we're observing it's going into the enterprise and so forth. It's a young language. It's a young ecosystem. But, the beautiful thing is developers today don't have to give up any package they love, any node package, any Python package, any design workflow is interoperable. Um, and so, I think that's why we're seeing such such great traction.

Yeah, well, so, you know, there was a time, and so, the computer science folks uh, would understand this, where developers had to decide what data would be in registers and what data would live in memory, and how do you move your your memory from the processor to memory so that your function call knows where to get the data. Developers had to do that. But, then that moved into the runtime system. That moved into the compiler. And so, the insight of Jack and Jaseki is all of that hard stuff. Why do you need to know what put, get, patch uh, you know, uh, delete, post like is in order to articulate a to-do list, right? Like the compiler can actually handle the complexity. That looks a lot to me like moving registers into locations for a function call. Uh, and so, um, we move that complexity to be the realm of the runtime and compiler that automates it. And then we also give you full access where you can leverage any wisdom that you would want to bring to the table. So, you can tool it yourself.

So, Michigan Michigan Michigan is a very unique university, right? Cuz Michigan is one of the most let's say, I've worked at a number of universities in my career, right? Um, Michigan is one of the the best environments to lead and create new ways of doing things, right? And so, I think there is a burgeoning evolution towards uh, focusing the department and the school towards the translation of innovation into companies, into startups, into you know, industry. And so, there's a lot of energy to cultivate students to become founders. That's burgeoning. There's There's a lot of celebration and encouragement in changing the way we teach, changing the way we we we really kind of participate with the unique role universities have in the world for the next generation of leaders. And so, and so, there's a lot of activity going on there. And you know, you know, I'm involved in a number of efforts like the Poise Grant. And then I'm working with colleagues that are involved with a number of efforts like Chris Blattner and Ling Jia Tang. And so, you know, I think that's a great place to start when it comes to understanding the exciting possibilities given how you're oriented towards research, towards entrepreneurial pursuits. But there's tons more.

Yeah, no, I mean, I think that's a That's that that one is one of the most stark things I'm observing year over year is we're definitely getting to a place where we are a community of strangers more and more every year. And I think, you know, COVID really is really created a step function in that. The concept of remote became a thing. Like all of the students today probably experienced the notion of remote work or remote watching your classes through the screen. They probably experienced that as normal, but that was new. That was that was unprecedented at that time. So, that kicked things off. And now we're observing this interesting trend where folks folks would rather engage with a GPT to ask questions than with each other or with a professor. And that's becoming a culture. It's becoming a culture of expecting to communicate to learn, but communicating with this new tool. And so, that changes the the ethos as to how students engage with each other and with professors. And I think that you know, that that actually diminishes one of the most important things and opportunities about university. I actually think university is more than 50% even more so now, more than 50% education through connection with each other, with I professors. And if you don't if you don't capture that, you like you're not getting your money's worth, right? A GPT professor is pretty good for the stuff that is in a textbook. You know what I'm saying? But it's pretty bad like at things like having this conversation that we're having right now. You know what I'm saying?

Yeah, so I'll tell you guys the truth, right? Because I guess I guess as a professor I'm supposed to give you a certain answer. But let me just give you the real my real perspective. I honestly believe that if you can find a community of people that challenge your ideas and you grow from that and you you develop perspectives around understanding truth, perspectives around study, I think you should pursue that in the way that enriches you most. A four-year The four-year institution is in when it comes to technology and innovation in technology, it's not as relevant as it used to be. I mean, if you can go to Y Combinator, then and and get an and and get the same kind of challenge and growth, that could be incredibly healthy. I think I think four years also is a long time, you know? If you don't engage it properly and get the enrichment in the right way. There are options for which your personal growth could be more optimized and optimal beyond just the four-year academic institution. And so, there's a number of right answers to that. But I think the key tenant is to be challenged and to have to adapt both your thinking and also have to adapt your skills for survival in a community, right? Like those aspects is what really turns you turns you into a leader or turns you into a thinker or turns you into gives you gives you the skills relevant to have impact in this world, you know what I mean? So, yeah, I think there's a number of right answers.

Yeah, that's a that's a really you know, a very few I think there's not a lot of folks that can kind of um it psychologically that that have had that experience, right? Psychologically it's a lot like I think the it's most akin to the to the concept of constantly being on tour as a musician. Um and also the concept of going to war. I think the psychology of an individual that's fighting a war, battle after battle, being deployed, and then mixed with the psychology of someone that's tour after tour performing, it was my experience. It was continual performances that were battles. And indeed, that is that's kind of one of the important roles of a CEO. But I was a really you know, at that point I was very forward deployed CEO. So, that means that I'm engaging with customers, solving problems, and also creating trust, constantly going to Europe every weekend, going to Singapore every other weekend, you know, like you really had to be where um where folks needed you to execute the mission of the company. And so, it's um it's unsettling. You always have this feeling in the back of your head as I can't wait to go home. I can't wait to have normal regularity and like brush my teeth with the same pattern and see the same stuff when I drink my coffee, you know, as opposed to flight, airport, location, meeting, talk, meeting, talk, talk, meeting, location. It's incredibly um it's fraying. It's fraying to the psyche. Um but sometimes it's necessary and you have to understand if you're doing that all the time, you're building your company wrong. And in in a lot of ways um if you are the critical failure point of success or growth in your company, you haven't built it right at a certain scale. You know, after about 15 to 20 people, if you're still in that spot, you're having trouble building your company. And so, you can have great success, but you're having trouble building a company. I mean, I'll invoke Elon Musk like for a second. Elon Musk's talent, I think, is the opposite of building your company wrong in this way. Because if you look at Elon Musk, he's got Tesla, SpaceX, um more Starlight, others X, etc. And I know his talent is building his company in such a way that he is not the failure point. You know what I mean? Because those company you that's impossible to be the failure point and do that many at the same time. And so, and so like these are the kind of like observations you can make. If you have a CEO like Steve Jobs, Steve Jobs probably wasn't like he was a failure point to Apple. Now, he grew into the point where he was able to put into place the right thing so it can sustain, but but Apple is not the same Apple today. We have the same product line as we had when he was around. And Apple was known for creating new concepts of products. So, so, so it's a very interesting dichotomy, right? Because we celebrate everyone. And there's different kind of talents. But it you know, the the observations you get when you're you're in this war gives you the the kind of perspective to understand the difference between say a Steve Jobs and a Elon Musk and what the strengths and weaknesses would be across these these prototypes archetypes.

>> Yeah, exactly. You want to minimize bus factor. So, we call it you know, like a lot of folks call it bus factor. In open source, that's a very common term. Is the open source project dead if the one maintainer is not able to continue? Yeah, I I think that if you want sanity, you have to do that do it that way. Like you you know, um otherwise you're going to be you're not you know, you're going to have to come up with some other mechanism to have the the balance that one needs in in one's life, you know, to to be to have well-being, you know what I mean?

Well, you know, this is um I think this is a very important point, right? And we were just chitchatting about this the other day about the the the fallacy and the hazard of of linking one's identity to external factors, right? Many many CEOs, especially young ones, especially early CEOs with their first company, let's say, are going to see that company as their baby, right? And why do we even use that term, their baby? Implicitly, a baby is a is an externalization of one's identity, right? And that's what we mean when we say this is your baby. Um and I think that that is fundamentally um that's fundamentally not ideal. It's not wise, right? Um and I think that you know, you turn down all the offers. You're like, "No, I'm going to take it to IPO." Oh my gosh, you know, I'll never let go of leading the thing. Um and so, you know, like one might say a prototype of this, and I don't know, right? I would never speak for someone else, but at least symptoms of this may look like a Facebook, right? Where you still have the same uh CEO uh on Facebook and making very top-down decisions. And I think I I mean, I I actually like Mark Zuckerberg a lot. I have a lot of respect for him. But I he might be an an element of the the kind of connection connecting one's identity to to the company. And I think that I turned down everything. I was like, "$50 million, nah. Maybe a bill." And this is why nobody was selling companies for a bill. But that was the mentality because it was like selling one's one's identity. Uh and I think that we do this with many things, right? People We do this all the time, and you have to find it in yourself. Oh, I love my car so much. If it gets a ding, I'm crushed for a week, right? It's almost like the car represents more than it just being a vehicle for value in in the world. And so, we I really think now I see companies as vehicles for value, and I really believe that realizing that value or proving that value, the first thing I'd want to do is to establish the longevity of the company with with more umbrella and more de-risked than say having wrapping a company up into one's identity, right? Yeah. If that makes sense, right?

>> It's a tough balance, but there's a distinction between being crazily obsessed with a vision as to how the world ought to be and with a company. >> Yeah. Right? Cuz a company has a logo and a name. But then there's the vision of what the world ought to be. And I think one of the biggest realizations is that your capacity, an individual's capacity to create the world as it ought to be in their eyes, lives inside them for real. It doesn't live in the company. You know what I mean? And this identity attachment is very very fundamental. Uh it athletes see it. So, you'll have these athletes. There's an athlete uh I you know, it's a kind of a sad story. There's an athlete that there was a story about who recently was at risk of never being able to play the sports again. Uh and she made the worst choice possible, right? Based on that devastation of that. We see athletes get injured, and they're lost. Um and folks not find their way. We actually get attached and see identity in even our

Jobs. And you lose a job, it's a crisis, right? And so, and so this this notion of being attached to what I call a manifestation of the persona. And this is like, you know, like Nietzsche and Jung and folks have discussed these kinds of topics, but when you're attached to the persona and you don't understand there's a distinction, then you're at psychological risk ethically. And to understand that the value of talent or the beautiful nature of a vision doesn't live in the projection of that persona, it lives in the in the in the person themselves is actually incredibly freeing because then it doesn't matter what your job is. It doesn't matter what it doesn't matter if you keep this role. Oh, I'm a ex-Google. I'm a Googler. I did FAANG. So, that I went to Stanford. You know, I'm a Stanford I'm a Stanford grad. Thank you very much. And then what if that was stripped away? What are you then? And I think that's the important question. And if you are talented and if you have a good vision, you'll under you'll be more powerful in executing it if you understand it lives inside of you.

Sure. And and that's why like people sometimes don't even look at you if you don't have that brand.

As concept. As concept, exactly. And this is the this is the key crux of the question. And I'm glad you asked that cuz it's important to to to tease out to understand that these affix the when when we affix ourselves with these titles or concepts or I you have this or you did that or you went to the school, we as the person need to understand that that is a tool. That is not you. So, it is useful, and it does help in affecting your vision. But but you're but it it's not you. So, if there's a death of that like let's say like you know, Stanford. I don't I don't know. I'm trying to find a good analogy here, but whatever. Let's say that that that that that externality that's giving you value goes away, then you have to then understand the landscape and create new externalities that can be used as a similar tool, right? You create something else that's that's important. You you you you you you you add value to another organization in a way that they celebrate you and in a way there's a mutual benefit that that organization also helps you attain more opportunities. So, it's okay to it's okay to participate in that world because we are in that market. It's a human fundamental thing that we have first impressions, that we use heuristics. It's psychologically well-founded. Like I look at this person, uh I have to use heuristics or I might make bad choices, right? If you don't use heuristics. And that's this is a manifestation of that part of the human uh uh uh behavior system. But but the the important notion is understanding that there is a separation between you and that in a way that you can't get wrapped up into that identity because then you make unnatural choices, and then you become more at risk. You you you someone can threaten to take away that from you. "Oh, I'm going to strip you of your board seat on I don't know, McDonald's board." And then you're like, "Okay, I'll do whatever you say because I am that." But to understand that you're not that and that you have the talent to procure Burger King. And it's lives in you gives you freedom to operate in this world and make natural decisions as opposed to unnatural decisions, if that makes sense.

Yeah, honestly, I think the the number one thing that is a common mistake that founders under estimate that they're going to have is you have to understand that you have to speed run failure. And so, a lot of founders will love their idea and want to get it right. And that's actually not the goal. The on the the goal is to create that feedback loop that has failures that creates the best learnings on one's journey. And so, you should you should do things where you know you will fail. Uh you know, you should go before you're ready. You should pitch an idea before you've thought it through. So, you can understand so you can fail to understand the reality of the world. We can't understand reality on our own. It's only through these mirrors, the idea mirrors of other people's reflections, and many of those other people's. And I think that's the number one mistake. Folks will try to build something, and they have a great idea, and they want to get it out, and they'll expect it to succeed, but it's a process of failures to understand what the world really is. And some people might project that into product market fit. It has to be a loop. There there's a lot of ways that this kind of manifests, but but speed running failures, incidents, you want many failures up to success, and then you want them to happen as quickly as possible. Like that's what you're optimizing for. So, if you're not failing enough, you will never succeed. Honestly, like if you're not failing enough, you'll never succeed. It's really the the the key piece of advice I would give to any um any burgeoning like leader, you know?

Yeah. Yeah. You know, so that's a very I love that question, right? That's a very good question. It really kind of gets to the crux of what, you know, So, I think it's really important. I spend a lot of time I I'm I'm really kind of fixated or I spend a lot of time thinking about um the next generation of of of humans and Um, you know, I really think like the ideal version I think of education looks like um, you know, I think it's philosophy, it's economics, and it's understanding the tooling of technology together. And I think the most important aspect is is imparting into students uh an orientation around the pursuit of rigorous truth and what that is. You know, we have all these beliefs, right? We have all these thoughts. And there's this illusion of the of the human mind that we assume they're true. Right? And so and we live we live in a in a world like that and I honestly think college is to diminish that because that is to me, and I'll just put it like this, it it's the childish thing to assume that you're right and valid. That is a lack of maturity. I think ultimate maturity is to assume you have to refine. You have to evolve everything you think at all times. And you're you have to continuously be evolving. And I think that a good education will produce a trans a translation of a human from uh everything is this kind of like everything is I think is valid is real. Right. And then into someone who thinks and pursuits a refined model of truth continuously. Uh and I think that that's an art. That's a that's a that's a something you do. It's not something that you are. It's something you learn uh through you know, the the journey of of a real education. And so like that'll be that'll be it. I think it can be done in 18 months. I don't think you need 4 years.

Check out Mars Institute. It's going to be very different.

Yeah, yeah, yeah. So there's a number of courses that I'm innovating. So I have a conversational AI course 449 that evolves every semester because every semester the entire landscape of AI evolves. Uh it's going to evolve into understanding agentic AI systems and and for next semester, I also have an open source course that I'm creating this semester along with my colleagues Lingjiao Teng and Chris Flattner and we're also working with Anthony Sotille who's open source you know leader in our community and a a member of our university community. And we're innovating this class around understanding open source and how to have impact through true truly creating something valuable. So that's another one and another big one is I'm creating a compilers and runtime systems for artificial intelligence and accelerating AI. And that just deals with how do you get models to like map to GPUs and TPUs and accelerators maximally performant. And that's actually a software compilation discipline for which there isn't much education around. And so those are just a couple of the classes I'm creating here at Michigan. And there's actually more classes I want to create. Uh and I might actually be working on um this kind of philosophy the intersection of philosophy economics and uh AI creation, creating with AI as one course. That'll be a wild course.

Oh yeah, yeah, Jack Hacks. Oh, everybody, you have to be at Jack Hacks. It's going to be insane. There's going to be incredible vibes. We're going to be engaging new technologies. We're going to be still wielding the old technologies that we know and love to create new things. We're going to be doing a lot of AI pair coding, vibe coding, agentic creations. We're going to be creating agents. We're going to be using agents to create agents. It's going to be nuts. So if you haven't signed up, sign up. And just be there. I mean, the the people that'll be talking um and the people that you're going to get exposed to are are folks that really capture the ethos of this conversation and our community here.