How To Solve It With Code—background on the course; a discussion with Jeremy and Hamel

okay so Jeremy if you don't mind reminding us why why did you start fast AI what was the initial Beyond like the Grand Vision of like how you wanted to help people I understand some parts about wanting to teach people AI but like what what is beyond that like what is the end result on people's lives we saw that AI we thought it was likely to be the most significant technology for hundreds of years and will impact everybody in society and we felt that could go really badly if very few people understood it and were part of it it could lead to all kinds of inequality and um la loss of opportunity for people or it could go really well if lots of people were able to use that technology and benefit from it and and be part of Designing a society that uses it so we started answer AI to help as many people as possible take advantage of AI and so we did that by creating courses to teach people how to use AI which at that time basically meant like training your own models um we uh did academic research to figure out how to make it much easier to use Ai and cheaper um faster and um so out of that came stuff like ULM fit which creed the foundation of the modern language model movement and we built software to make it much easier to to to use AI um so that was that was the foundation and that worked really well like most of the top practitioners I talk to today tell me that vast AI has been part of their Journey um um uh but I guess what's really exciting is we're now at a different stage where you don't have to train your own models anymore uh and AI can do a whole lot more stuff than it could when we started fast AI uh so that's why Eric Reese and I started answer AI to like basically go to the next level basically to say well we can be more bold now we can do more with this uh um get funding to build a team to to go even further than fast AI could and what do you think it like compared to fast a what do you think is going to happen with the audience or like the the target sort of persona that you're able to reach in terms of being able to use AI Now versus the way it was before so so now you don't need to train a model uh and in fact just like there was a point at which you could suddenly start using the internet without knowing anything about tcpip and SMTP and whatever you just like you know kind of it that was possible a long time before the iPhone but in some ways the iPhone was a thing that brought that to most people's fingertips um I think chat GPT was a thing that did the same for AI it meant that now lots of people can use AI without knowing how to code without knowing how to train a model without having lots of computers um and that's that meant that suddenly we've gone from the fast AI the original fast AI target market which was anybody who's already pretty good at coding we'll teach you how to use AI um and now our goal is to say um you know we we actually want to basically teach everybody to use AI um having said that there's still a lot of benefit in knowing how to code and so a lot of people now are using AI that don't really don't know how how a code at all and so we've decided so so two things um fast AI has become is is is now part of answer AI uh and fast AI is so it's has access to all that additional resources funding people working on the company um and through that we can now and and have now built a new kind of course it's not just a new course it's a new kind of course uh this kind of course has never existed before it's it's like uh a course that is only possible with AI and it's going to teach people who to to to solve problems that they previously didn't know how to solve using code and AI together so it's called how to solve it with code that's the name of the course um and um the when I said to new kind of course it's not it's not just going to be me on YouTube anymore we've literally created a whole new tool with language models built into it which people will better use to work through these the the material and how to solve it with code so we we built a whole new platform um for this new kind of course tell me a little bit about um the new this kind of new kind of course from a more like okay let's zoom out from the tool let's talk about like the new way that you can teach something or the education the way it's delivered like how how is this actually different like the learning process so um so Rachel Thomas and I created the first version of The Fast AI course in 2016 and we did it every year or two the last one was in 2022 it didn't change you know I was proud of that like it it it didn't change and like it did the the the the structure of it the method of it it didn't change and in fact a lot of the content didn't change much either because it was very you know um it was not ephemeral kind of stuff it was stuff that the basics of SGD and so forth that asked forever um but I haven't done another course since 2022 and the reason why is I've felt it it doesn't make sense to create the same kind of course when the technology has now shifted and the kinds of people that can use the technology is now shifted it's so much more accessible and you know with language models the world's different so I didn't feel comfortable creating yet another course that was basically the same because things aren't the same now which is great um so this is like the first beta version of this new kind of course I think things are going to go a long way from here but the basic idea is like let's make AI part of the learning process not a little add-on chat box in the corner or something right but but a whole program built in and around AI as part of what is taught and and and how it's taught so let's let's uh talk about that a little bit more a lot of people when they think of AI they think of as something they can use to do things for them I think a lot of people don't uh consider them in ways they can use for learning can you talk a little bit more about what that means in terms of yeah learning with AI yeah okay so why I mean so I think the question I start with in is well why are we teaching a course how to solve it with code when we're saying like people don't need code as much anymore like why is that the first course that we're teaching in this new way and um you don't really need a course to say like how do you use chat GPT like you can get help with prompt engineering and stuff like that but the kind of thing you can like make a start on in a half an hour of reading a blog post like but we've seen again and again uh so many of our friends hit a wall with that approach where they're trying to do stuff that's you know they're trying to solve problems and build things that go beyond what chat GPT can like solve in a single step um and it's really impressive what stuff like chat GPT and Claude can solve in a single step but you know as anybody who's used these tools for a while has seen there's lots and lots of stuff they can't um just not a criticism at all of them that's just that's just how it is you know um so you actually do need to to go beyond that you do actually need to learn to code but there's that we've we've developed a way of combining coding and before we go further can we Linger on that a little bit Yeah of why is like chat GPT plus coding so exciting like the combination of them oh yeah how does it expand the problems you can solve or yeah so if you know a bit of coding right and if you have a good way and we've built this way a good way of combining coding and AI you can take the stuff that that the language model is giving you and tweak it modify it pull it apart recombine it and then interactively do that with the AI to form a dialogue in which you're constantly feeding back oh this you know this was a bug or this actually doesn't quite solve the problem I had in mind or whatever but not just with more and more prompting right because that just as we've all seen you get into this really annoying Loop where you're trying to get the AI to write a bit of code and you know what it needs to do and you just so hard to like sometimes get it to do what you want just write the damn code you know you you like you actually it's just so great when you can just dive in and add a couple of lines of code yourself here and then have the AI add a couple of lines of code there and now it's man and machine combining together and um at least at at at this stage of History um that's vastly more powerful than machine alone so so that's what this tool we've built and this course we've built is all about which is and and a lot of people think that coding with AI is just about building SAS apps or building software but do you think it goes beyond that into just solving everyday problems that might have it's that's why this is called how to how to solve it with code it's about how to solve problems and actually like when you're creating an app or a dashboard or whatever there's just that consists of solving lots of problems but lots of things do as well like if you're doing a data journalism exploration of you know organized crime company links based on thousands of PDFs or if you're um you know trying to segment pathology slides or if you're trying to understand the uh change of um uh the word literally over time in in usage through through studying texts online or whatever like stuff you're doing or if you have an office job and you have to make that dreaded weekly report exactly you've got to do your TPS reports whatever like and they have to be done by Friday at 11: you know otherwise there'll be questions being asked so you know these are all problems to be solved um and they generally consist of sub problems to be solved and so if you know how if you know how to solve problems with code plus AI then it's this big unlock basically um and you don't hit this wall where you're just prompting chat GPT and getting into this bigger and bigger hole of technical debt as it writes more and more code that you don't understand and it's increasingly complicated and it's like you know full of weird bugs that you don't even know are there because you don't really know what what's building for you when when you and the AI are working hand in hand then you create things which you understand exactly what's there you understand exactly why it's there it's combining your knowledge your understanding the constraints your understanding the opportunities your creativity together with the ai's massive you know knowledge base and understanding of syntax and and all that so it sounds like combining AI with code just gives you superpowers in terms of using the AI for using AI for a lot of different purposes and we've seen a lot of different tools being becoming popular things that promise to allow you to write code without without knowing anything about code but I think people getting a bit stuck in those and they can you talk a little bit about what you've observed and sort of explored yeah yeah the the thing I've observed is that there's this like huge excitement as folks that don't know how to code at all or barely know how to code jump on Claude or chat GPT or whatever and within 15 or rep it whatever you know within 15 minutes they've built an app for their kid which they've had in their mind for years and it's like keano Reeves and the Matrix like I know code yeah it's very exciting and they've gone from zero to mark five and then the kids like oh thanks Dad that's nice but when I click this button doesn't actually do anything so oh all right how do I fix this prompt prompt prompt prompt prompt like okay the next bit takes twice as long just to fix that one damn button and then the kids like oh yeah my friend haml wants to do it as well we want to have a leaderboard it's like and it it doesn't work great to build and build and build and you know um this is how what it like you look at the real world of like who's building the apps that we're using every day they're not built by people using who can't code using just chat GPT they're just not you know like 99.99% of them out um because we're not there that's not where the technology is they're not using AI at all they just they still know how to code and they might be using AI but none as exclusively using AI right nobody's really built the tool specifically designed to really harness the best of humans and the best of AI so that's what we're showing people as a process and a tool for doing that so like Google CEO said that most recent the most recent few months 25% of their code is generated by AI um and so yeah it's a good example right like they know quite a lot about AI at Google and it's not that they've replaced all of their coders with AI but they're also not using it like it's not they're avoiding it either like this is where we are with the technology today is it's great to know both but like most of the labs big Labs have been focused on building AGI you know they're really focused on building AI that stands alone um and we're one of the very few groups that are entirely focused on building stuff that combines humans and AI together in the best possible way so that's why you know we've created this very new kind of course and very new kind of tool to power the course and tell me a little bit more about the tool and the Gap that it fills specifically um you know like you have all of these tools like cursor Claud and they write code for you and you can imagine someone saying okay I'm G to try to learn about The Code by reading what it's producing or asking questions but I suspect there something there's still a gap so like you made this tool and this platform for a reason can you talk about sure so the platform for now at least we're calling solv it okay so there so when as part of this course you'll get access to solve it uh and Sol it supports a very particular kind of workflow or a variant of a very particular kind of workflow which already nearly all of the best developers I know use I find most folks who have been programming for more than 20 years or so when are at the very top of their game they they have a workflow that very heavily uses a very highly interactive approach um and some you know so for example if you look at three blue one Brown videos um which are very popular very good uh Grant Sanderson has demonstrated that the way he builds those is very interactively he's got these Cod here he types in a bit and immediately he sees the results you know um Peter norig who ran research at Google you know very heavily used this kind of you know similar kind of style of trying things and getting mediate feedback um it's extremely different to what something like cursor compose or or clawed artifacts or whatever use where it spits out like 200 300 lines of code like there's no exploration there there's no learning going on there as you gradually build up your solution So Sol it is designed to kind of be something that creates one or two lines of code for you and then you write you know maybe one or two lines of code yourself and then like you gradually ping pong back and forth you know um and so you end up with something that's generally a lot less lines of code than you would have had if AI it's spat out this big thing for you why is that because to be honest like uh these these uh language models have not been trained to create they can't yet create particularly high quality code uh particularly concise code and also like the nature of language models is that they spit out the kind of code they've seen before so often like it'll have like all kinds of unnecessary stuff that like you would need in some different environment um and so it just ends up being overly complicated and not really customized for your particular problem that you're solving um so yeah you know this is really all about harnessing this much more interactive approach that that really pretty much all of the world's top practitioners use so like if you watch Andre capa's videos for example or you know uh or watch how George Hots buils um you know tiny grad or like with all of them you see it's like type type type type type try check this see what happens like two lines of code two lines of code you know it that's that's the process that we're supporting but it's one where AI is part of you know helping to generate that code part part of helping to check that code and teaching you also like what for the bits where the AI is written some code for you is like make sure you really understand that that code and why it's written that way and what it does you know some people might think that hey writing one or two lines of code at a time and understanding these one or two lines of code that sounds a lot slower than you know having 200 lines of code written at once it does doesn't it I I happen to have the privilege of knowing that you've used this system to write real software and so what is your observation is it like a lot slower because done so yeah so you've seen it so yeah so yeah it it it it it honestly for anybody who's been doing the like cursor compose you know Claude project thing of writing hundreds of lines of code I have a feeling this is going to feel terrible at first you're going to feel like did I just create two lines of code when I could have created 200 um but like um to me it feels like the difference between telling a kid in grade one hey here's a calculator here's how you add and subtract okay that's it school's over you know go start your career versus saying like you know what we're actually going to teach you to add yourself and we're actually going to teach you to subtract yourself and and actually you it turns out that that's going to build and build and build and you know um so you'll still use the Cal you just use it in a way more powerful way yeah yeah yeah no so and then it is slower for a bit right but then the kid that never actually learned me math than being taught to use a calculator has a huge limit on what they can do you know or else the kid who actually learned math can can create things you know um so yeah it it it it and and they end up much faster right because they're the ones who are like oh you don't need to go and spend six months doing all those calculations M you know summing manually all the numbers from one to a million because there a short hand which is this closed form you know solution which I can do in five seconds so this happens all the time so when you build the code up slowly you end up much faster than the person who does it 200 at a time and you never hit the wall of like okay that's I'm now stuck I can't go any further because I don't understand the codes that there and I can't get chat GPT to solve this issue and so okay that's the end and I've seen like people when they go from like zero to mark five so quickly at the start of their coding journey and then they hit these walls they become really despondent you know this to spare hits and they also feel embarrassed about it because they were out there on social media and telling their friends I can code you know the days of software developers are behind us you know like we can just use AI now and then they hit these walls and it's embarrassing almost you know it shouldn't be right but it can feel that way because it feel like oh everybody else seems to be able to use these tools to do whatever they like I guess it's not for me I guess coding is not for me you know and so I understand understand like the best software developers they use this iterative approach where they write one or two lines of code and they they understand you know kind of they're able to interact with that code and kind of Build It Up over time and it's a lot faster yeah one um I think even like so I I buy into that as the best approach I think something that's really interesting is it happens that that approach gives AI further superpowers because you have built up a conversation right that I think that part is not intuitive and no one knows about that right right because there AR there aren't tools there aren't tools that do that right so so yes so Sol it knows about all of those interactions and the interactions are extremely Dynamic and we actually have built a new way of thinking about working with language models which we call dialogue engineering as opposed to prompt engineering which we'll we'll have a lot more to say about over the coming months and years this is solvit is the first time we've publicly released a tool that's built around the principles of dialogue engineering although we've been using it internally at anai for quite a long time now and everything we built in this course for example example has been been built using these tools so there been a bit of a bootstrapping process here um yeah when um when language models can see examples of questions and answers that have gone well you know in a process that's working well it it has this amazing self-reinforcement Loop effectively that happens um uh there something it's based on something in the literature called in context learning but with dialogue engineering at takes this idea of in context learning a lot further where there's actually this feedback loop that happening so yeah solit is explicitly built to support this this really strong positive feedback loop where the the user is helping the AI for its next step and the AI is helping the user for the their next step and I think that's the that's the part that is exciting is is the AI is actually making the AI do things that are not possible with traditional approaches oh absolutely like just using cursor or whatever CH or really anything else like the level of difficulty of problem that you're able to solve and the complexity that the AI can solve is much higher and almost somewhat unbounded in a yeah yeah yeah yeah I mean I've been cing about that yeah I've been coding every day for over 30 years um and coding a bit for over 40 years um and in like and I constantly try to become better at it always reading and learning and practicing so I become a good coder but with this dialogue engineering approach that we're that we're building into solve it over the last few months I've been able to build things that I could not build before um and and it's definitely stuff that uh chat GPT GPT for uh1 preview Sonet none of them can do on their own because I try um it's only by combining the the the human plus the AI that I've been able to build this and so it's been interesting for me like I would say even for a quite experienced coder if you're prepared to be a bit patient and humble and learn some new things you might find this course interesting as well because for me it's showing a process which I found has given me superpowers and so so it sounds like we're going to be teaching students this uh of unique technique called dialogue engineering where they can really you know let the AI do a lot more than any other approach that I've seen what are some examples of problems you've tackled with dialogue engineering that you suspect would be very difficult otherwise um I'm trying to think I'm not very good at remembering specific examples of things um the the the the kinds of things that I find very often um well so um well here's a good example actually um the our dialogues contain messages this is just something that you know um happened somewhat recently and um you know I realized it would be nice to not like in a chat bot not always add messages to the bottom like normally happens with chat GPT but sometimes I'd actually like to I wish I'd asked a question earlier you know um and then that becomes kind of part of the dialogue um and so I wanted the ability to insert a message between two messages and I it's like how do I do that because my messages have like a primary key which is integer you know and like you can't insert something between two integers and I was like oh maybe I should make the primary key float gee that would be a weird you know but then I could insert it between two messages by taking the average of the two float keys and and so I started this conversation you know where I was like hey here's some code I've got you can see the issue with the primary key I kind of insert messages between them I thinking of making it a float instead is that like possible and and the AI is like oh yeah that's not actually unheard of at all you know he you know he you could use something like this I was like okay well is this a known technique you know does it have a name and it's like oh it does actually it's called fractional indexing I like never heard of it and I looked it up and I so then I came back and I was like oh okay I've here's some lines of code to impl this fractional indexing idea or here's a library I found how does that look and you know so I ended up uh you know with this better understanding of a whole new kind of data structure and algorithm that I'd never heard of before and I ended up building you know something that I didn't really know existed and it also gave me the confidence to kind of follow up this idea because to kind of be like hey can you make a float a primary key and use the average of keys to add things between like you can't really search stack Overflow for that it's very conceptual so yeah um and to try and do that in kind of like a classic chat GPT kind of environment you're not you don't have that ability to to build the code as you go and get the kind of the feedback as you go it's really you have to beg uh you know the model to create the code for you um and also then might you know the the the way we've built this is then the code is is that you're working on together is working in your environment you know that that has everything set up the way you want it so yeah that would be an example of something I created recently which I guess I could have created anyway but my version would have been honestly a lot less good than the one we created together so it sounds like dialogue engineering not only makes you smarter it makes you faster it also makes the AI smarter what about other people you might be working with if you're using a dialogue engineering approach does it change the way that other people may be able to read your code or understand what you're doing it's been great it's been great yeah so we we share this a lot at answer AI now we share our dialogues with each other so here's here's the thing I built and here's how I built it and we give each other our our dialogues um and then we can build on on top of each other's dialogues by adding more code and more prompts and you know sometimes in the middle and um the you know the the version that we'll be using of of Sol it for the course is quite simplified from the internal tool we've developed you know to give people this kind of like Get get them going with it and not be overwhelming um but over the coming months and years you know this particularly this group of people in this first po are going to form this first community that we're going to build out together so they're going to be the first group of people that understand dialogue engineering and just like I think you were part of our first fast AI cohort H all the folks from that first cohort as we see I still know and hang out with and um you know I think that Community is going to be exciting to see how together we we build this new way of thinking about human plus AI together I do want to add though that this is not designed to be a replacement for something like clawed artifacts or cursor or whatever like there there will be times where you do just want to say to Claude like just just oneshot it just make this and it does it correctly first time and you don't really need to add new things to it and a leaderboard and blah blah blah right it's like um I have a specific thing I need right now it's not going to be too complicated it just builds it it works first time and I use it cool you know it's the kind of thing that like repet works pretty well for as well let's uh let's linger a little bit on the dialogue engineering and how it makes it easier for us other people to understand your code can you talk about why like uh versus like normal like traditional ways of writing code because you know when you um yeah so here's the thing reading a finished piece of code is intimidating because it tells you nothing about how it's created you know it's like the difference between watching Bob Ross paint versus saying here's the Mona Lisa it's like hey here's the monisa are you inspired okay go make a beautiful portrait like I don't know how to do that you know it's like okay here's Bob Russ he's like okay well I put a bit of paint here and I wash this over there and then we can use the edge of this brush here so with dialogue engineering you end up with a dialogue so you see the whole process of building it and you you understand like why was this done this way what does this thing here mean why was this Library selected um it's um yeah it's a really nice thing then to not just share with your team but to also to have available for yourself 6 months later when you've totally forgotten um and so yeah it's uh it sounds like you understand your own code better the AI understands the code better and your co-workers and colleagues understand the code better yeah students or whatever yeah yeah exactly so we're very we're really at on AI as you know Hammer we're all in on this you know we been using it ourselves for months and we're like holy crap it's changed our life so we want to start bringing everybody else into our into our world that we're so excited about um you know because I know you and work with you think I have some appreciation of like you know this is only the tip of the iceberg um you know the these methods of dialogue engineering are really powerful um and this course is just a tip of the iceberg too right there's going to be a lot more courses so you know we're planning for example a kind of full stack startup engineering course you know and where we'll be building on these ideas to be like here's how we can use dialogue engineering to create a company you know to to to do all the system administration to handle or the email to build the marketing materials to to write the web application so on and so forth you know um and you can say that with confidence because you've been building answer AI with this exact system it's not you're just this is not a this is not a prediction this is actually happening right now correct exactly so uh yeah so we'll be giving the rest of the world a bit of a insight into how answer AI is able to do such an extraordinary amount with such a small number of people so quickly which I know a lot of people are like how the hell do you do this you know you guys must be like Geniuses sure okay we could be Geniuses but I don't think that's the main reason I think we've we' we're building some powerful processes um which we're which came to share all questions I have thank you very much H that's great I will see you over in chat okay all right see

Discover how Fast.ai evolved into Answer.ai and is revolutionizing how we code with AI. Join Hamel Husain and Jeremy Howard as they reveal a groundbreaking "Dialogue Engineering" approach. https://solveit.fast.ai/