the point is if if someone's younger or someone's changing careers jump on this wave right there's no downside as far as I can tell right because normally you might imagine there's a downside where like what if this technology goes cutting with AI I think we can kind of rule out the idea that it's going away yeah right because it just keeps getting better and also it solves so many problems even if it didn't get better it's still super useful for solving problems as part of other pipelines and things like this an understanding of AI that you get in 2024 is going to last your a lifetime because I just can't see how it would go away it will look very different in a few years but you'll be able to build on that knowledge that you've already got so I think it's a there there's no drawbacks outside of the core of the of the hype stuff that you see in the media there are thousands of examples of smaller Ai and deep learning applications that are solving problems across the world and they just quietly happen behind the scenes they don't you know get all the big hype but they're doing a huge transformative job and so even if you don't go and work for the big tech company training the next chat GPT you might still be doing something with massive impact right and really really worthwhile doing I I think that the the road map right for for AI is hey everyone it's David Bumble back with the amazing Dr Mike pound Mike welcome thanks for having me back it's great to have you you're the person I always talk to about AI how it's changing our lives the hype the reality you know you what I really love about what you do is you tell us not a full for the hype and you tell us what's actually going on so what's the state of AI or you know where do you see it where it is in 2024 yeah I mean it continues to move really really quickly I think that there's still a huge amount of hype so we can talk about about that I'm not going to change my tune on that front I think the thing that we've seen change you know really quickly is is more image Generation stuff so yes we can still talk to AI chat Bots and we can still have good conversations but now we're linking in images so now we can produce images it can describe images we can say remove this thing from the image and put in something else instead and and these images are just getting better and better so you know we we we're fully into the area now where you know you can't absolutely be sure that an image is real or fake now you know not not always anyway I've seen a lot of YouTubers use it really well for generating thumbnails and I've been using it as well it's amazing what it can do exactly like you said you know just tell it like with a quick sentence my daughter's been playing around with it like create a hacker cat or whatever a cute hacker cat and then it creates this thing but I mean like you said some of it's so realistic you have to look really carefully to see what what's real and what's fake yeah and you know there's going to be obviously a lot of um regulatory problems with that because you know we're going to have a situation maybe not tomorrow but sometime where there's an image in a court and someone's trying to decide if it's real because it's really important for some case right and this hasn't happened yet but it's not far off there's the um the famous example of Tom Cruz been um actor pretending to be him and putting out Tik toks or something and you know you have to look carefully you do and and and actually um I made a joke on one of my videos that you know I don't look enough like Tom Cruz for this to work and then the next day on Twitter there was a video of me with the face of Tom Cruz just talking in my voice you know and it was pretty good right you know it the speed at which you can produce these things now is quite incredible I think the um I want to talk about like opportunities um because that's a big thing that people are interested in and also like the worries because I think people are worried that the jobs are going to be taken away there's this concern that you know in the in the old days technology was taking away perhaps factory jobs or basic jobs and AI is going to take away a lot of like knowledge worker jobs I mean at the moment I don't think it is in a sense that I think these AI tools are really impressive but of course when you actually really start putting pressure on them in terms of something really difficult they start to fall down a little bit so I think that you know your job safer now but you know inevitably eventually we are going to have you know image generation that's so good that you know it is comparable to an artist we're going to have text generation that's so good that it's comparable to a writer I think as a society we have to think about what we're going to do when when that is the case now there are huge other concerns like you know who owns the writing that an AI has produced is it what maybe it's the people that trained it right or maybe it's the text that went in to train it so there's a lot of unanswered questions I think in the short term there's not a huge threat to sort of office work and things like this you know if you're doing if your day-to-day job is you know coding or your day-to-day job is data entry or you know Financial you know work with let's say spreadsheets or something like this the ability of an AI to do to to use these tools is still limited right it's it's impressive sometimes but it is limited I've I've had feedback from people that it's great for instance to help with boilerplate and I think you said exactly the same thing boiler plate code it's great for doing certain things but in our previous video you mentioned the fact or the problem that it could write code that has vulnerabilities in it yeah so you know we don't know because we haven't been told what all this data has been trained on what the networks have been trained on so you know there's going to be correct code that it's trained on there's also going to be bugs in the code that it's trained on and we don't know to what extent it will overcome these bugs or they'll go come out in the wash or whether it will just repeat these bugs back to you when you use them I know a lot of people who work for companies who have policies in place about the use of these tools just from a safety point of view more than anything else right you know if you've not got a sort of a set of human eyes on something at the moment we're not in a position where we can fully trust what we what we're getting out and so you know if you're in code for a living by all means use these tools to help you speed up if that works for you that's great but I think that the idea that you're going to be able to go home and just let it let it churn away on its own we're not we're not quite there but it won't and that won't be coming for a while I've heard people say that artificial general intelligence is also a long way out we might not even see it in our lifetimes yeah I think that that would be broadly what I I believe AGI is a term that's used far too much in the in the modern conversation I think because it's exciting sounding and it has you know sci-fi you know implications and stuff like this but I think realistically the majority of researchers think that you know in the next 5 or 10 years we're just going to see better versions of what we have now but nothing's going to substantially completely change in terms of you know we're not going to jump from where we are now to something that can you know think and act for itself um you know we'll see if I'm proved wrong but I'm pretty firmly a believer in that so I think that we're just going to see much better chat Bots and much better generation one thing that hopefully we'll see soon is more grounding in real data so you know when you you know you can see this on Bing already if you use Bing it does a kind of web search and then injects that into its own prompt to try and give you better informed information now that sometimes works it sometimes doesn't but I think that grounding the output is a good idea because at the moment you're just getting a kind of amalgamation of training data which might be out of date it could be incorrect I see I mean correct me if I'm wrong but I see that the industry's kind of moved on a bit since we spoke as well I think we last time we we we spoke about 2023 Outlook and it's good it's interesting to see what's happening now in 2024 in the old days you had to get like a PhD I had to like be like you if I wanted to use AI uh but these days I can use apis so it's like we got the users using like the front end on chat GPT that's one opportunity perhaps to create thumbnails and then you've got like the developers who could use the API and then you've got like the hardcore stats guys math guys people with phds like yourself it's like there's like three groups now and there's a lot of opportunities in each of those there's there's opportunities everywhere yes you're absolutely right so if you take so I mean one one of the terms that gets used a lot is these kind of foundation models which is a model that is incredibly large it's been trained basically by resources that most people don't have access to so big tech company and you can deploy these models on subtasks in you know without too much expertise right so for a good example is a segment anything model is a model where you point it at an image and it will just start picking out objects now actually segmenting objects for many years has been a really difficult problem but you can use this as part of your pipeline and you don't really have to know how the AI Works to be able to use this tool which just run it in Python it produces you some some some objects and then you can decide which of those objects you want let's suppose you're writing a web app that you know let's say picks people out of an image you can use a segment anything model and then just tell it to find people and it'll go off and do that and you don't know how you don't have to know how the network actually segments the people you know you're just using the polygons that come out so in other words there's a huge opportunity now where it's it's open to a wider community of people who perhaps are not hardcore AI guys but they can Leverage The Power of a of AI just by interfacing with apis yeah so yeah these companies are exposing these models at Via apis that you can then you know you can make use of and you know we see a lot of companies doing this many of these Tech startups are essentially a wrapper around let's say chat gpt's API and trying to use it in a clever way you know t ex there are risks doing this because of course you don't know exactly how you don't control trap GPT you don't know how long it'll be online for you know these kind of things but on the other hand it gives you access to models that you couldn't otherwise have trained and you couldn't otherwise have have actually done yourself because maybe you didn't have the expertise you didn't have the resources and so I think that it is a low it is lowering the barrier to entry for a lot of people yeah so rather than me trying to get like crazy data sets CU that seems to be the biggest problem right getting the data and then like the gpus and all the power to run this and obviously the expertise like yourself you know I can just leverage an API and I've got it yeah I mean you know it's it's difficult to overstate the number of resources needed to to train chat GPT right first of all you need massive amounts of internet data in the order of trillions probably of tokens right at the very least High numbers of billions and then so that that in itself is is a dayto nightmare but then you've also got you know the hundreds and or thousands of of gpus across hundreds of machines that train the model in a distributed fashion all the structure to do with this no need to do all that you know just sign up get your API token and off you go right and then you can talk to this thing and and you can tell it what to do I'm interested in this do you have recommendations for instance of languages I need to learn perhaps books you know how do I get started because AI is so much hype in the news but how do I practically use this to for instance have a business or just change my life you know in some ways I think yeah I think that actually one of the things is you what you want to do is educate yourself to an extent where you can get around this hype right because that will allow you not to be taken in by the most flashy tool but the one that's going to work best for your use case right and so I think to do that you need a little bit of expertise in certain areas you know you don't need to do a PhD you know by all means come and do a PhD with me that would be fun but you know if you haven't got the time to do that and and you want to get in on AI more quickly then sure there's lots of things we can do the first thing is you know you don't need a tremendous amount of mathematical knowledge to run AI you definitely don't need much of mathematical knowledge at all you can even train networks without much mathematical knowledge if you want to read the papers and understand networks you might need to know a little bit of linear out and things like this right but that you know that's about the extent of it you can get away with with with a little bit less I don't mind maths I use it when I have to um I avoid it also quite a lot of the time so you know it's not so bad the the primary entry point is going to be learning python that's the main thing you need to do for whatever reason I mean I you know I've said this before I have a LoveHate relationship with python some days it's my best friend some days I can't stand it but ultimately this is what the AI Community has settled on if you're going to use machine learning you know python is going to be the thing that runs your networks python is going to be the things that handles the input and the output of those networks now of course they're all actually implemented in C for Speed and Cuda on the graphics card but you you interface with them using python so that's the first place to get started and do you have any courses or do I need to is it's just basic python just grab some like how to get started with python actually it is really I think um you know any appropriate introduction to python course is going to be is going to be fine because what we're doing in Python is not actually the most complicated things that python can do right if you're writing Enterprise level software you know you might not be using python but if you were you'd be using a lot of perhaps more sort of advanced stuff around the edges that you don't need to use to do U machine learning at all the main thing you need to know how the data structures work so lists dictionaries and then you know pretty quickly once you've got a grasp of the fundamentals you can actually start training up smaller networks to get your head around the the quirks of each particular Library so for example pytorch has a data structure that it uses called a tensor in fact actually most AI libraries do but the tensor is kind of fundamental to how Python and pych works so the sooner you get going on that and the sooner you play around with these things you know you pick them up pretty quickly so start with python just like a generic python course understand lists dictionar basic python knowledge then look at pytorch do I need to understand Concepts like supervised unsupervised learning you know all this kind of AI terminology what I would say so you know to begin with not necessarily right but what I would say is probably there is a nice progression actually supervised learning is perhaps in some sense the simplest and the easiest to get your head around mostly because actually still the majority of AI that we see in both in industry and in research is supervised it's supervised because that's the easiest one right you know you've got some data you want it to perform that task so you just throw the data in and you train it and then we go from there if you download code off let's say GitHub that has that has a new network in it and it trains up something with some with on some top ask the chances are you'll be able to plug your data in and go you know and you'll have to fidd around with the data loaders a little bit in a bit of python there but really that's the first thing you can do once you've done something like supervised learning then you can move on to more complicated topics like unsupervised learning weekly supervisor there there the you know hundreds of subsets of of different configurations that you could train with and then you know you could also start to familiarize yourself with these you know large language models and these big networks that are trained using a hybrid of different approaches is where do I learn pie torch is there GitHub places to where I can go yeah so there'll be loads of there's loads of courses but actually the pytorch GitHub the examples and the P torch tutorials those are actually really good good resources there there's a couple of tutorials at the beginning of the py torch tutorial set which is sort of things like tensors Auto differentiation so it's Auto grad framework it's worth working through these because you'll just get a feel for what's going on under the hood so actually pytorch does a lot of stuff behind the scenes that you don't see and if you kind of skip over that you'll get it running fine but you might not necessarily understand what's going on but I think if you have in your back of your mind what it's doing it actually does make it quite a bit easier to use so just for everyone who's watching I've linked the video below Mike did a great example showing us how to use code that he's used or created and how to like work with images and do some interesting stuff so I'll link that video below if you want to learn I'll also put links to P torch below so if you want to learn that what about last time you recommended like Andrew has got a course on corser that teaches the basics of should I do that yeah so I think that um so corsera I think costs more than it used to right so that's a choice that people have to make I think you can do it for freest can't you I think you can at least briefly right if you're quick yeah I I think I still do recommend that course I think that course is there's two courses really there's the machine learning one and there's the and there's the Deep learning specialization now actually I'm a big proponent of you know I recommend often doing the the initial introduction to machine learning course because I think it teaches more of the fundamentals that you can't pick up through just having a go you know things like how what does a learning rate do and how what effect does it have on the training of a network and what happens if your network doesn't train what do you do in that situation going from machine learning to deep learning is actually not that big a leap at all right they they're very similar Concepts they're trained in much the same way it's those initial machine learning Concepts that are going to take a little bit of time to get your head around so that's a good course there are plenty of courses on all of the online learning platforms you know in s in similar things but I certainly think if you're taking a course that's talking about learning rates how to train the training process how you prepare data those are the kind of things that are really worth learning I think we need to convince you to create a course yeah I mean when when I find some time I have a go I sometimes do think about that yeah I think it will be quite good cuz I spend a lot of time teaching these topics to a lot of people it would be easy if I'd have just pre-recorded myself I could just say go over and watch that and then come and talk to me afterwards that'd be great yeah we got to well so everyone please vote below put your comments below if you want Mike to create a like a udemy course or some kind of online course where we can all learn from him now I really appreciate you sharing this Mike and it's you know it's it's overwhelming and you've been doing this for a long time and like I said you you separate like all the hype from actually what's reality so it's for what it's worth it's overwhelming to people working in now as well right because it's there's so many papers I mean we've we just submitted a paper to a conference where we're I think submission about 12,000 I mean that is that you know 10 years ago there were not that many papers in computer vision almost across the world in a whole year and we're talking now about one conference the the the amount of and the speed at which all this stuff is happening and sometimes you know sometimes papers are small improvements and incremental improvements and not changing any you know much but sometimes you know there big big things coming around you know in graphics we've got these new uh splattered gaans we've got New Radiance Fields these things didn't exist a couple of years ago and now we have to learn what they are as well you know and so it actually it is overwhelming but I think also it can be exciting right if you're if you're willing to delve in and read some of these things there's load of resources to explain all of these topics you don't have to necessarily read the maths in the original paper if you don't want to I think it's it's like when there's a shift in technology it opens up a lot of opportunities for anyone who's willing to jump in right yeah absolutely I mean I think it's like you know when we 2 came along right it was the same kind of idea I think there is a lot of opportunity and actually I know plenty of people who are working in AI who don't have you know a PhD in AI they just got sufficient amount of base level knowledge that they could get in on the ground floor and work their way up and then you know you learn these things as you go as well most of what I've learned about artificial intelligence I've learned by doing it you know in my research rather than just reading tutorials and things like that so I think you you get yourself that base level of knowledge and then pretty quickly you're ready to go and you can start training things and you will learn quickly if you just you know keep training these models do you think there's going to be a shift towards like domain specific AIS where it's like an AI for cyber security an AI for networking an AI for XYZ technology I think that um so in a way I hope there will be actually because I think that the the things like chat GPT and these these large language models they're very interesting and they're very fun but actually I find them quite hard to apply to my own research because my own research is kind of specific right I'm often looking at a medical image and trying to work out what shape that thing is and chat GPT doesn't know because it's never looked at that kind of thing before and so actually it's much easier for me simply to just train my own thing to do it right and ignore the large language model I think that foundational model models or big models trained in specific areas such as segment anything they're more tailored to a specific task they're slightly less Catal and that means you might have a chance of being able to control them in a way that's helpful to you so I hope it goes that way at the moment I I have fun playing with these tools but for for the majority of the time I don't use them because I've got my own models that I've trained up Mike what's what segment anything segment anything is a is a really interesting um should we say Foundation or a large model that's been released by by it's kind of transformative in the segmentation space segmentation prior to that has always been a supervised task that is you basically well I mean pretty much so you give it a bunch of labeled regions in a load of images and it LO learns to label those regions back so you might use it for example for labeling you know products as you're walking around a shop or something like this now segment anything is a tool that just segments all the things it can see in a scene and gives them labels as to what it thinks they are but you can also prime it with points or boxes or text to say I want you to find all the footballs or I want you to find all of the uh laptops and it will go off and segment them for you now it's not perfect because it's a very general model it may underperform a very very specific model trained on a very specific task but of course you didn't have to train it did you you you just run the thing and and and it's done it all for you so I've seen it I've already started seeing papers that are using segment anything as part of their pipeline like they build in just segment everything first get rid of the things we don't want and then we've got some useful data already and we haven't trained anything at all so it's with images right it can pull stuff out of an image yeah yeah you can think of it yes it's a bit like a kind of reverse you know stable diffusion or darly right instead of producing an image it's taking an image and finding interesting information from it so one of the tasks that people do is is image to text so they try and describe what's in an image this is image to object so it's trying to find everything in an image for us but like in your specific use case with like Medical Data you at the end of the day it's better to use your own AI because it's that specific domain of knowledge if you like right yeah I think at the moment it probably is um also you know we have the expertise to run it so you know we might as well use our own our own technique I think you know in the long term for certainly for outdoor scenes the sort of scenes you see typically in day-to-day life outdoor and indoor normal photographs I think that it will work pretty well and it might not find everything but it will find everything pretty quickly for you for medical images or super specific specific data sets you might find find you have to train your own specific AI but you know you could use what you could do for example is you could use something like segment anything or a less powerful you know a network that gets you the gist of what's going on to start that annotation process for you so actually it can be used sort of in with you in the loop to get data really really quickly so from a job's point of view or like I'm just trying to think how can I use AI to put me in a different League like job-wise or you know just in general use loads and loads of companies are using AI right now some of them are going to be working on financial data some of them are going to be working on networking and security data some of them are going to be working on image data it will depend on the company you know going to be dozens of job types under that umbrella term so there'll be people actually engineering the networks themselves they may they may not be anyone in a company that does that because they may be using offer shelf networks and then they're going to be people who are you know interested in the data and the storing of the data and and moving the data around you know your site and then there are people controlling the cluster and G us and training the data and there are people who are providing a web interface on the front of the data so you know front of the network so I think whatever your expertise is there's there's a place in that pipeline for you right and you know if you have some understanding of AI that's going to make it a lot easier because ultimately that's that's what's actually running under the hood so Andrew SC I think is just is is an absolute recommendation for anyone who's technical in any capacity right because it gives you good understanding and then you can take that knowledge yeah no that's abolutely right if you want to learn how to run a model then what you do is You Learn Python and then you can just basically start running some tutorials and you can run a model but you won't necessarily understand exactly what it does now it might be that your job doesn't necessarily require that you understand exactly what it does but I suppose what I would say is at some point it's not going to work right at some point you're going to train it up and it's not going to work and you're going to need to know why and so some expertise in AI That's where that will help you know and it will go okay so it's doing this that means that means that we need to increase the learning rate or that means we need to change increase the amount of data we've got Andrew Ang's machine learning course and similar courses are that starting point that gets you that that AI knowledge right so knowledge on things like when you actually train what actually happens right what what happens to your loss as it goes down and is it good that it's doing that and and you know how can we monitor this and make sure it's working correctly there is a bit of maths right so if you're absolutely not okay with doing maths then I would suggest a more practical kind of pie torch course that just teaches you sort of the fundamental of running the network Without Really obsessing too much about exactly how they're trained bearing in mind you can always pick that up later right you know it might be easier in some sense to learn the mathematics behind a deep Network after you've been training them for a year and you understand exactly how they work so you know there is that angle you could take as well yeah just think I mean even if you just dip into the course it gives you kind of like terminology you understand like if we talk about supervised unsupervised we talk about Transformers at least you understand what that means because if you totally NE to it means nothing right sorry go on no no that's absolutely right I think that learning the terminology and understanding a bit about what they are lets you work out what the best tool is for the job as well right so for example Transformers which you mentioned are very common all over the research and all over research and Industry actually most big models use a transformer in some way mostly to do the text and you know the text to image and image to text um but actually we've had tasks where we haven't the Transformer doesn't work right we run a Transformer doesn't work not enough data the problem is different it doesn't work but we have other things we can use as well right because we know what they are so I think understanding a bit about these different technologies will mean you apply the best one for the job cuz sometimes you know these big models are are overkill for what you need actually you need to just detect something in some object some image your your image is very consistent it you know the same picture gets taken every day nothing changes that's a few hundred images in a small Network and you've got and you and you solved uh so you know sometimes you don't want to overthink these things you can you more basic is better I just think we've got to the point I've seen this shift in the last year where you know chat GPT was like this Catalyst where it became mainstream but since that time I've SE I've seen Tik toks I've seen uh Instagram videos I've seen YouTube videos where it's totally AI generated you can see it's an AI voice it's an AI image so this stuff's getting applied more and more in a lot of places I think the days of like ignoring AI are over you have to get involved at some point would you agree with that yeah no I I mean yeah I work in AI so I'd love everyone to do that too right but I think it's it's so General now right there you've got it is so I mean you know most people that are using AI that we see in the media and you know on Tik Tok and social media and so on most of them are just running it right they're running it without too much of a concern for how it works I think you put yourself in a really strong position if you just know a little bit about how it works as well and particularly because at some point someone's going to present you with an image and say this was AI or this wasn't Ai and isn't this amazing and hype stuff is going to start to come up and you can go hang on a minute I'm not sure about that I'm actually I don't think that was done in in that way and you know I don't think that's quite as impressive as you think or maybe it is but I think you you know that understanding really really helps and so I think it does you don't need to totally be able to you don't need to be able to write your own transformer paper I think just understanding a little bit about how they work and why they've been designed that way is a super useful thing to do and then you know you can get on the AR you can get with the AR and you can start using it right you can use these things as well I think for like someone who's who's like not technical perhaps and I I ha to break it up but I see like again like I mentioned in the beginning like three groups people who perhaps not technical but can inter interact with an AI and learn how to leverage an AI just by talking to chat GPT but that's like the beginning right and then you want to go to the apis and you want to use Python and and understand it and then the third is like going and actually understanding more and more and the opportunities are in all three places but you're going to be in a much better place I think if you're like in the second group where you're a programmer or someone interfacing with an API or like in your position where you're actually doing much more with AI so there's three groups right so the first group is just people who've had a go at generating smes online so they go on on on the website there's a text box you type in your you know your prompt and an image comes out and you know you are controlling AI in some sense but only in a loose sense right and I suppose the the interesting thing about that is there's a lot of stuff you know you hear this term prompt engineering which I think is a really interesting term next job sorry what did you say about it sorry yeah I said it's a really interesting term because in a way you're trying to control something that we don't the people that train the network don't really understand it right in terms of what it will and do with certain prompts which is so interesting it's like trying to you've got someone working for you but they don't speak your language and you don't speak their language you're just shouting words at them to hope that they do something that kind of works I think that's really interesting but of course you run the risk that you know if you build any kind of product based on something like that you run the risk that the underlying network will be changed or the API will change in some way and suddenly it doesn't work anymore because when these networks are retrained we can't guarantee that they will do the same thing the second time you know some understanding of of how these networks work lets you you know control for that a little bit and you can be a bit more sure that what you do is going to be predictable so perhaps the next level up would be someone who's you know using the API and is perhaps doing something so for example something that that companies often do is they will prime a chatbot with some text that allows it to control allows you to control exactly what it says so let's let's imagine a hypothetical situation where I want to write a tech support but I don't want to pay anyone to do tech support so what I'm going to do is I'm going to write all the answers for tech support into the chat bot so it has that information and then let it do my tech support for me now that might work quite well for at first right because as long as people ask simple questions that you've just given it the answers to it will produce lovely answers for those things but if they are what are your opinions on you know you know this conflict that's going on across the world or what's your opinions on you know really political Hot Topics or what are the instructions for bomb making it might tell you those things because those are still in the original training set right so you have to be quite careful if you're going to create a business that's based on hoping that one of these large language models will act exactly how you think so that's a really interesting problem that's not been solved yet and we'll be seeing people hopefully addressing over the next few years is and then perhaps a little bit above that is okay so you've got some tasks you want to solve you know a large language model or segment anything network will kind of work but it's not good enough right for your task that's when you probably need to train the network yourself and then you know that's going to be a case of you download the repository you you write you write your data loader so that your data is going in you write your objective so that it knows what to learn and then off you go and wait so I I suppose there is there's different levels but you can always start with one and move to the next one as you as you as you build up more expertise I've seen people the the example for the second group I've seen is someone who and I'll link the video below who uses Ai and python to for instance he says go get this video off YouTube and then give me like a 10 bullet point summary of the video and tell me if it's how how important it is that I watch it based on my preferences and he's like put that all into a python script it's ultimately that's not training a network or changing the configuration of a network or something like that and you may find a time where your script doesn't work because some something's changed in the video or the network that means it that no longer no longer functions in a in a way it's a bit like using any programming Li using if they deprecate a feature and change it you're going to have to change your code base to to reflect those changes otherwise your code won't work and so it's kind of a similar deal except perhaps that the pace of change is very very rapid so you might find that your prompts from from this month don't work next month and so you know there's a bit of a risk there but if I want to go to like the third group do I have to like Buy CPUs or do I have to rent stuff in the cloud or is there like a way to you know actually do that just like to learn no I think if you want to if you if you're a company and you're looking to do you know proper training of AI then Cloud resources like Aur or AWS will will have you covered right if you install your own local stuff you have to obviously support that you have to pay people to look after the machines and make sure they run but on the other hand ultimately it's probably marginally cheaper depending on how much training you're doing the cost computations get quite comp licated depending on you know what what you're up to but of course if you're just playing around with these things trying to learn them then something like Google collab is perhaps the best place to go to so Google collab is a is what we call a jupyter notebook style interface so essentially you have lots of text boxes where you can put in Python code but it also has access to gpus it has all of the libraries basically installed already so you don't have to do any of that you can just say use pytorch or import pytorch and off it goes and you can and actually a lot of the big models like detectron 2 which is a really good model for object detection they come with a link to collab on the git repo so that you can click on it and just run it straight away and see how it works and actually that's how I learned how stable diffusion works as well when stable diffusion came out I went on the GitHub I went to the collab and I started playing around with the networks to see what they would would do and it's a really great way of learning and that's free or low cost right it's very low cost it's free for the sort of entry level you'll just find that if if you get over excited and you use it a bit you might have to wait for a GPU right so you know you know you can pay I want to say it's about eight or N9 a month you know British pounds for a a monthly access which gets you pretty much you know all the access to a GP you could need unless you're some sort of ridiculous power user but you know there are lots of tiered pricing models but I think it's a good place to get started because you may end up not using collab in the long term you may have your own systems or you might use proper Cloud compute but I think for just running things and trying them out it's a it's a great place to go I mean that's why where I wrote my demo and I've actually written other demos as well in collab because I know that when a student clicks on that they're going to get access to the code they can run which is you know is is reassuring so for everyone who's watching once again I've linked a video below where Mike actually demon demonstrates the stuff I'm asking all the questions that I'm hoping a lot of you are thinking about so I know the ANS is already but that video is linked below great demo so Mike again thanks for sharing that another question that always comes up is books do you have any recommended books or or um study resources apart from like going to the pie torch and and GitHub and I'm a big fan of any not too long introduction to Python and pytorch right yes you the problem is if you if you buy so Yoshua Benjo and his colleagues wrote this incredible book called Deep learning Now by all means get that book I own a copy but it's a it's a big read right you know and there's a whole section on reinforcement learning there's a whole section on superv you know and you'll be there for they'll be there for weeks so if you need the a really good reference for one specific area that you're looking to go into then sure right if you're looking to pick up pie torch and run some stuff then really what you need is a book that will get you there quicker and I think that's something that teaches you python teaches you the fundamentals of how pytorch Works things like data loaders the training Loop right which is takes a bit of getting a head around the first time and how it how it trains the network things like this that's the place to go so there are a few books that we can link to that have you know like not too high a barrier to end I think and that's where I would start best I'll link those books below so if anyone wants to get them I'll just be aware I'll use Amazon affiliate link so thanks if you want to support me but I'll I'll put those links below yeah Mike thanks for sharing that because it's you know some people learn by doing some people like watching videos some people like reading so this gives us a whole bunch of of resources now another question you train or teach a lot of people at University and a lot of them are perhaps beginners so talk to people who are starting out or people who are changing careers um and it might be a bit of a nocy question but like even advising yourself what would you advise yourself to do in 2024 is like jump into AI as soon as you can do computer science or what would you advise if you're like talk to you young yeah I mean I I spend my entire life telling people to do computer science and I will continue telling them that until either I retire or I die one of the two one of the two things um I love telling people about computers and AI is just one of the cool things you can do it doesn't you don't have to do AI right you know um but for the sake of this video you absolutely do have to do Ai and I think what I would say is you know I um so actually as an example I recently took up the drums just for a bit of a laugh I wanted to be good at the drums and I was absolutely hopeless at them but I kept doing it about an hour a day and now I'm not not bad right and I think actually it's it's similar with everything right like with AI I have students that come to me for let's say Project work so that's in in the final year of University where you basically do a sustained period of work and then you write up a large dissertation right or you know big essay essentially but you have to produce some software or something that works and obviously I mostly supervise AI projects and so these are students who are decent you know computer science but maybe haven't done any AI before at all and within a few weeks their training networks and within a few months they've got proper tools going and it's really really impressive it isn't quite as much work as you think right yes becoming a sort of PhD expert will take you some number of years but I I think to get a baseline level of AI knowledge is only a few weeks and months really I also a big fan of learning by doing but I do think that just clicking run on a collab notebook isn't quite enough you need to have a little bit of video Intuition or a little bit of reading just to make sure you understand what's going on under the hood but it's not you know it's not as much work as you think I love that it's really motivational because you know people look at this I think a lot of people would look at this and think it's like a huge mountain how many years of experience have you got I mean all these you know degrees and stuff PhD I can't do that um and I'm glad that you you're saying it's not as difficult as that no I don't think I think doing a PhD is a very different set of skills so for example one of the things I can do quite well because I've had a lot of training is learn new things very very quickly so if a new if a new paper comes along with a new kind of network I can pick that up within an hour or so and know what's going on and what's good about it and what's not good about it and that's useful for my job because that is basically what my job is right but actually if you're if you're working for industry that's less of a concern perhaps you want to be keeping an eye on what's coming along but actually a lot of it is you've got some product to deliver and it's going to use AI how do we go about doing that and so there's a sort of set set of steps that everyone's going to do in that situation you know learning those steps doesn't require that many years it you know it's just something that you just you you start off you get that Baseline knowledge and then you can learn on the job so like you've been saying and a lot of people say start with python if you haven't learned python already you need to learn python in 2024 learn P torch go and look at Andrew's course on corser if you want to get like an understanding of AI but if you just want to jump straight into it P torch is a great way to get started but you need python as your as your beginner you need you you absolutely need python um pytorch so there there are other libraries you could use tensorflow for example but my experience at the moment is that the the sort of the momentum of the research community and industries is behind pie torch as of today now of course we don't know what will happen a year from now some new fancy product comes along we will get over excited but at the moment pie torch has got a huge number of repositories a huge number of tutorials huge amounts of help if you read a paper that does something cool like detects objects that you wanted to detect there's chances are there's a GitHub repo that implements that paper and you can kind of run it and and in doing so then learn how it works right in and put yourself in that code and Fiddle about with it and see what happens when you change some of the configuration yeah well I mean what I would say is I think there at the moment because there's so much hype around AI there's perhaps a tendency for people to LEAP into the shallow end of AI really really quickly so that is things like prompt engineer ing just running models look all this AI I can do and that's a great place to start but I would quickly advise people just try and get a little bit more of a feeling of what's going on underneath because then you're better skilled when those things change and and you know adapting to those things right and also for what it's worth I think it's a really fun place to work there's a great Community around AI it moves very fast but the Core Concepts don't move as quickly right so you know supervised versus unsupervised learning is still the same as it was before for pretty much and so you know you can learn these topics and they'll keep you going for a good number of years while you can work on the other faster moving topics but was think it's just money right I mean a lot of people obviously want to get paid well for doing this and jobs pay really well yeah they do I mean it's it's a it's a constant complaint in Academia because you know all the best students off the go right they pay really really well and obviously it depends on the job like you know if you're if you've if you've sort of self-trained you aren't going to be able to jump into a you know a top AI job straight away I mean maybe I guess if you if you can really prove yourself don't don't rule it out but I suppose but you can move up quickly right if you show your potential and you and you learn these techniques if you obviously have a PhD you've got a set of skills already there written down on paper so it's perhaps a bit easier but I don't think it's out of the question for anyone to do this right I think you just have to it's just like any job you know you don't start at the top managing you start you start somewhere at the bottom but you you pick up the topics and you get better at it and you know and they all they all pay pretty well I just I always advise people to ride the waves I mean you and I have been around the block a few times and I've in in Tech I've ridden quite a few waves like Voiceover IP going back many many years opened a lot of doors and then there's like in networking there's automation etc etc the point is if if someone's younger or someone's changing careers jump on this wave right there's no downside as far as I can tell right because normally you might imagine there's a downside we like what if this technology goes Kut in you know I'm learning some specific Library let's I'm I'm going to pick a bad example let's say you're a JavaScript programmer and you want to learn react right now reacts great but it might not exist in 5 years time right it might not let's not um this is not a prediction now I've got a really good example I learned open Flow and that was hot for like a while and then it just died sorry go on you know exactly you can't rule it out absolutely with AI I think we can kind of rule out the idea that it's going away yeah right because it just keeps getting better and also it solves so many problems even if it didn't get better it's still super useful for solving problems as part of other pipelines and things like this an understanding of AI that you get in 2024 is going to last you a lifetime because I just can't see how it would go away it will look very different in a few years but you'll be able to build on that knowledge that you've already got so I think it's a there there's no drawbacks I love that because it's you know like some technologies it's a solution trying to find a problem but like you've said here there's so many problems already being solved by AI outside of the core of the of the hype stuff that you see in the media there are thousands of examples of smaller Ai and deep learning applications that are solving problems across the world right and so and they just quietly happen behind the scenes they don't you know get all the big hype but they're doing a huge transformative job and so even if you don't go and work for the big tech company training the next chat GPT you might still be doing something with massive impact right and really really worthwhile doing I I think that the the road map right for for AI I is is a subidon of maths dotted throughout right but no but I'm not going to mention it is is is Learn Python learn P torch you could use another Library if you want but I recommend P torch get a supervised model going so supervise some some learning right so get your data set it doesn't matter what it is it can be a public one that already exists it can be something you went out and shot on your phone it doesn't matter at all and train a supervised model that solves that problem and you're already a good chunk of the way there now right and then you can start making the problems a little bit more complicated or you can move from a simple classification problem to maybe a segmentation problem or something like this that's I think the right the Right Way Forward I wouldn't start with anything other than supervised learning because it's the most intuitive right you give examples you learn from those examples um and once you've done that you also have seen all the pie torch you really need for all the more complicated examples they just have more code right that's that's the difference it's still the same stuff Mike I really want to thank you for sharing you know as always you you separate the hype from like the reality and give people hope cuz that's always the one of the worries I I I hear from a lot of people is like if I'm 18 why would I bother even learning Tech going to do computer science cuz AI is going to eat all my jobs I really appreciate you you know giving us hope with that it won't right and you know particularly if you know about AI right there's always going to be a need for this I think that whether or not something will take someone's job is something for a government to worry about I'm not particularly worried about it at the moment on behalf of everyone else you know at all I trade I'm hoping that I still have a job soon we will we will see if you keep inviting me back as a as an associate Prof we we see but I think that it's an exciting time it is overwhelming but actually the there's a huge opportunity and it is really fun so you know you start training there's nothing actually I find more satisfying than when you train that Network and actually does what you asked it to do that's super cool when that when that happens even though it's ultimately just python code you're running when it when it actually works and you and you see that output it is good and you don't and you don't understand how it's got there right cuz you just giving it data and you're teaching it no you don't understand the it's essentially a really complicated function that has done something clever we sort of have a gist of an we have an intuition as to what it's doing but you know ultimately we don't really know what it does but actually I've stopped worrying about that a little bit if it works well I'm kind of okay with it I love it
2024-01-08