Abonniere den Podcast

LINKS

Episode 12 - If You Are Not ALL-In on AI, You Are Probably Not Agile."

Transcript

(disclaimer: may contain unintentionally confusing, inaccurate and/or amusing transcription errors)

hello and welcome to the scrum sessions podcast here in Dallas I am here on a very sunny day and my friend in Berlin it’s you hey everyone I’m Gary on SQ said I’m based in Berlin Germany and today we want to talk about AI right since at least since 3gpt 3 came out it’s in everybody’s mouth and we hear a lot of stuff about you know what’s going to happen with our jobs what’s going to happen to knowledge word how does it impact scrum and agile in general and we want to just talk a little bit about it and see where this is taking us because it’s still fairly new and I think a lot of people will find it helpful to get some orientation so I think one of the big issues that we have when we talk about AI is that unless you’re an expert in that field then congratulations you are probably in very high demand and can renegotiate your salary package right about now um but unless you’re an expert you probably have some some problems orienting yourself to what is actually going on right because there’s so many terms being thrown around and you actually need to know on a base level what’s going on and so what I would like to start with is this categorization it’s really not important once you have understood it but you kind of have to know what it is to see like what level of AI do you expect and so there’s actually three categories and the first one is narrow AI right it’s some very specific AI let’s say it’s an analysis AI that looks at pictures of cancer scans in a medical context right and it can detect cancers right it’s very narrow it’s not very threatening unless you’re that person that is doing that job then there is General AI meaning it’s almost like on a human level so now we’re talking about an AI that can actually do a lot of stuff not necessarily as well like every task as well as a normal General human but it’s pretty darn Advanced right and then there’s super AI which means it’s a it’s a god-like being right it can do like basically everything we can do on a in a on a high level and in a better way and this is where um things get really threatening


okay so if I had hair would be on fire right now I am gonna lose my job so is that what AI is going to do for me get me fired you might it depends on it right so if we if we build up that super intelligence I think we’re all out of a job because uh I think some Terminator stuff might be happening pretty soon but this is actually quite far away if we ever reach it right so I think people are really fixated on this super intelligence um but that’s still far away I get that we need to start thinking about it right now and put it in security protocols um but this is not going to happen very soon right so we don’t need to worry about that much and then it kind of comes down to it so if you do a job that can be done better by a narrow AI so for example looking at cancer skins and then yes you might be out of a job pretty soon um because the AI can do it usually can do it much quicker than you can can probably do it cheaper than you can and it can actually look at more scans over time right and then be able to detect um cases that might not be as frequent right you might not see enough cases to ever detect them but the AI might so um you know there’s a lot of different opinions on it but I think um yes you might very well lose your job but it kind of depends on what you’re doing you know actually what worries me is not losing my job I don’t worry that the world is going to end in two years like somebody just said last week or we’re gonna start say I’ll be back all the time right I’m not worried about this I’m out I’m really about to worry this about the ludites because when the Industrial Revolution is started people go there and break the machines because they’re gonna take my job right and I think people are very myopic when they don’t realize there are several tasks that are very boring they’re very time consuming and they are downright dangers right and we should actually get ourselves out of that business right and I think when I think about this and forget the three categories of AI I I think about it as something that augments my capacity my capability my abilities of Jew things so in your case for the cancer scans right okay yeah you can see that being scanned you can see that’s gonna do much faster and Etc but still everything has to be peer reviewed so it’s a lot better for me to get some automated tank to scan 500 things and alert me hey this one may be suspicious right and then I can make a decision based on that I cannot accept the AI view of it or I can say no based on my experience this is something that you don’t know couldn’t identifier Etc so I I was kidding about losing my job and all that stuff I’m less concerned about this I think I am very concerned it’s about the people who think I mean there are people setting fires on 5G masts right they’re people breaking stuff I think people are gonna get in the same thing oh this thing’s gonna take my job so let’s set fire on computers or or that sort of stuff how do you feel about so I think the best metaphor that I’ve um heard about this is that it’s actually kind of like a tractor right so if we think about a tractor it’s really unsexy nobody ever thinks about a tractor unless you’re like working in that field or you have some fetish and play computer games around it but I don’t ever think about tractors don’t care right but when tractors were introduced they made a lot of manual labor um Superfluous right A lot of people were working in the fields and the tractor was a revolution and it made those jobs Obsolete and so I think it kind of depends on what perspective you take if you take like a more economic broader strategic view then this is probably one of the best things that has happened to humanity straight up and it’s one of the biggest opportunities for us being alive maybe bigger than the internet maybe bigger than mobile phones right it’s a real big opportunity however if you’re actually one of the people that is working in the fields it doesn’t really help you that somebody’s telling you yeah well you’ll find something new in time and you don’t have to work in the dirt anymore because this is actually what you want to do or what you’ve been doing and you need to find an income pretty soon right and where you’re gonna find that new job if there are not a lot of them around right so I think it really depends on um on the perspective as far as the latter lights go yes I see that um and I’m not actually sure if this is going to be a traditional Luddite problem or if it’s not the same um vaccination misinformation idiot crowd that we’ve been seeing jumping from one topic to the next topic just like always trying to get riled up on on something right and so this is going to be the new new thing um for them um that they’re going to jump on this anti AI bandwagon but you know you always have Village idiots and they’re going to find the next topic and I don’t think they’re gonna really change anything yeah and uh you always have to consider that people overall don’t like change right the majority of the people are not comfortable with change there are people like you and I I think even I I’m even worse than you I like I buy different cell phones every three months I’m looking for a new laptop I’m always looking for technology and not only technology other things I like change right but the majority of the people are uncomfortable and on top of that there are the The Click baits that we see all over the news and Etc that try to Foster this fear of change even further because then they sell clicks right and then people can go after that now I always taught as a professional in any area and I may even inject the world’s business agility here because it’s a topic that’s very near and near to me right you should be always observing the trends right otherwise you always be caught by surprise when you’re caught by surprise well your reaction may be super optimal you can get in a lot of trouble right so I think we always should be observing the trends in the industry right okay so in your case maybe I like to drive the tractor right and now the tractor can pretty much you program the GPS and the tractor goes there right but we are still not to the point that you can get rid of every single operator because there’s some decisions that the system cannot make by themselves right and again that’s why I always keep in my mind the augmentation of things right if for instance I say I need to write a business letter or business proposition or something for some reason right instead of spending all the time doing that I can generate one really easy and then I just go there and make the changes that are necessary for me so instead of spending half an hour typing and you know how bad I am at typing right I can generate something and just make some some minute changes or whatever change I want to make and I did that in five minutes so I can use the rest of the hour I would be typing that stuff for half an hour to do something else right so I am very keen on the idea of augmentation so now that you’ve brought our business agility I actually think that AI is the case in point to see how good a business’s agility is because frankly I think too many people have been taken by surprise by AI um it should have been on a lot more people’s radar right because it didn’t come out of nowhere right it feels like it’s um but it didn’t really so I think a lot of people should have been more aware there um radar system their detection system should have been in place but okay that happens but what’s really striking to me is how quickly did businesses pivot to make use of AI in their internal tools right so for example use copilot use strategpt internally like you just described but also how are they implementing AI into their products for example and so I think it’s been I don’t know I have a bad time feeling let’s say four to five months since AI really blew up and I think if you are an a so-called agile organization right if you if you think you’re agile and you weren’t aware of what’s happening in Ai and then you suddenly found out and you still haven’t adapted to it you still haven’t made drastic changes to it I think you might not have as much business agility as your as you’re thinking because this is a big topic on the development side on the delivery side as well as on the product side and I think this is really the litmus test where you can really check how agile is your organization how much have you Incorporated this change because this is really urgent this has a lot of power behind it a lot of productivity gains and if you’re too slow to adapt to it you might run into trouble down the line no you’re absolutely right but I always like to make this a distinction between the urgency of something in order to be agile to respond to it and the hype right you and I work in I.T let’s be serious I.T is driven by hype more than anything else right we saw that with the crypto thing and now we see the Fiasco the whole crypto thing turned out to be right because there was too much hype I don’t think that’s the case I think the hallway I think is very interesting very important and will bring some really good benefits but we have to Discount some of the hype that’s associated with it and I think hype is always damaging because it may turn down to disappoint a lot of people but I I think that’s the biggest problem I don’t know if I agree with that because yes of course there’s always hype um and you can actually see it in some of the tools that you’re being shown on social media and then you check them out and it’s like ah well this isn’t that great so hype is always there but I mean um Kevin Scott the CTO of Microsoft he said that 40 of the code is actually being generated internally like at Microsoft is being generated by copilot and I know of teams that have immediately pivoted to start using copilot and the results are enormous right the productivity went through the roof and um so you can see I mean you know this is not talk anymore um you can actually see the productivity gains and you just describe how you’re doing it for writing letters and so on right and think about marketing teams how much they can profit from um GPT alone but also more fine-tuned models that help in marketing and so I think on the delivery side the writing is already on the wall I mean you can if you spend the time to investigate you can actually see the productivity gains that you can make and is just think about let’s say something like copilot giving you five more percentage points of developer productivity think about how much money that is if you don’t just have one developer right if you have a bunch of them maybe hundreds of them and you use a tool like that how much money how much productivity can be gained and so yes there’s hot yeah but um we can actually see we can you know do a gamble walk and see what’s going on and it’s crazy


no I I don’t disagree with for instance the code stuff I am big into home automation and about two weeks ago I was looking to do something in Powershell to automate or routine and I I could not remember for my life what Power Circle when they needed to do that right so I I jumped in there say hey show me code to do this Bingo pop me some code in there I didn’t like I tweaked the code slightly and I use the code otherwise right so again augmentation on the other hand you have to be careful with the hype because I think it was last week and the week before it was all over the news this guy I think was Avianca versus somebody right and the attorney for the defendant uh went there and uh generated a bunch of stuff and there were four references in there and he submitted like that and the judge look and said none of these things exist right so again that’s the hype and that’s the misuse of the thing just because you got something in writing sort of speak it doesn’t mean that’s true right which holds true for a lot I mean 98 of you see in the web is garbage right this misinformation and all that stuff and I know that’s 98 I just made that up right but you cannot get an output or something and assume that’s that’s good because a machine gave it to you absolutely that’s where I think that’s where the hype comes to play yeah and they are the models are hallucinating um sometimes um they sometimes give you really crazy results and it’s funny if you’re like a little bit more public if you have podcasts for example and then you like ask it something about yourself and then it gives you like completely made up stuff that’s actually kind of scary right because some people will take it at face value and like where’s your recourse right if a Wikipedia article on you is wrong you can write to Wikipedia right but if the model is hallucinating about you like how do you correct that if people then go on and take it at face value


yeah well even even if the model refers to Wikipedia right there is also the problem of validation what what somebody put in Wikipedia right which there has been several people to demonstrate that you can put a fake Link in there and through a bunch of stuff make the link look like legitimate right so there is always that now I I that’s why I make such a careful consideration about have to be careful with hype right of course the technology is going to be great I I already benefited from that in multiple times you know especially because I keep saying I suck at typing right and all that stuff but we have to be careful not to assume that models that are relatively recent there is still evolving and Etc are like the end of all and they’re gonna resolve everything now because I’m very concerned about security right and you know that I I lock everything down and all that stuff I’m also concerned about privacy insecurity because I believe Samsung just made a No-No don’t put corporate on this stuff right and I think a lot of companies are not paying attention to this people are just typing stuff in there that may be and very likely is intellectual property and they’re throwing stuff in there right and on bleeping computer last week there was actually a bunch of research that somebody said and there were some astonishing amounts of very sensitive data being thrown on those models right so you have to be careful with that so again we have to educate people we have to do a lot of leg work before we get too much into the high mode that’s ISO fear I’m I’m not disagreeing but I’m I actually think that my lens of action is the choice here because if you can get these productivity gains and you were describing them before you have to take them if you don’t tank them and you don’t take them soon then your competitors are going to do that right and so I actually think that on the delivery side if you’re still mucking around if you still haven’t invested the time to get into the tools to look at those that are available to you and start learning them and start to integrate them you’re really missing out and you’re making a mistake right of course be mindful that you don’t put in proprietary data and so on be mindful of the results that you get that the model isn’t hallucinating but if you for example if you’re a developer and you haven’t tried out copilot or another tool like it actually I think you’re not doing your job you have to do it it’s an incredible chance to increase your productivity and if you want to be agile and you’re not looking at a game changer like this you are doing something wrong in my opinion no I I totally agree with you I just have some concerns about privacy data security and etc etc etc and I think a lot of those things are not being considered because of the hype behind it now the usefulness of the tool yeah I’ve been using it I’ve been playing with that for the past two months right and there are lots of good things that can come up from it but I also seeing some really dumb answers and all that stuff you call hallucination just say you know it’s just something evolve but I fear stuff that’s hosted somewhere I don’t know who is manipulating the data and all that stuff right I am quite sure that Microsoft on their models they have internally I doubt that code is getting out of the door right so we have to be very very careful with that now you know me I am against regulation I don’t think regulation solves anything there’s already people saying we have to regulate Ai and all that stuff I said great believe me the people misusing it they wouldn’t care less if you’re regulating or not right so I think what we have to do is instead of regulate we have to educate because regulation if you are a bad guy you want to use that thing for some nefarious usage or some sort you know hey you don’t care it’s regulated or not do you I don’t right but I think it’s more about educate I think we have to take this opportunity to educate everybody on how we can use it how we can properly use it how we can benefit from it and how we can filter out some of the hype instead of just trying to put some heavy-handed regulations something because it’s not going to benefit anybody and it’s going to have absolutely zero effect just look at crypto okay they did whatever they wanted billions of dollars went into vapor and all that stuff oh well you know regulation no it doesn’t matter it’s regulation at all people were up alone good they’re gonna do it anyway so I’m not sure where I stand on regulation as far as AI is concerned however I think the cat is almost is most likely already out of the back right because the there’s open source models out there and they’re getting smaller you can actually run them on on your phone right and so um the cookie is kind of crumbled however the other thing that I want just want to touch on real quick is that you know open AI is pushing for regulations and old men the CEO is and this is from a business perspective this is actually a super smart move right because in him pushing for regulation they being ahead Microsoft being affiliated with them and Microsoft having all those lobbyist Connections in in DC from a business perspective this is the gangster move to do right because if there’s going to be regulation guess who’s going to be sitting at the table it’s going to be Microsoft it’s going to be open AI they’re going to shape the regulation in a way that’s gonna raise the mode behind Razer mode doesn’t make sense it’s going to create a mode behind them right and then um this is a really good business move to keep the competition from creeping up so yeah um we might like it or might not like the regulation but I actually think it’s going to come pretty soon and it’s a good business move I’m not sure if it’s good for the public um benefit but I think we’re going to see it


I don’t question that they’re going to see it you know I just get scared when corporations get very cozy when corporations defend anything there has there’s there has to be and usually there is something behind it right so I fear when people get really cautious let’s regulate this on companies that are always fighting against regulation otherwise so I I question the motives yeah I definitely question the more tips you know and uh there are some data Hogs and we we could say company names here but I rather not that are probable Readiness using this quite a bit you know I’m quite sure they are you know and uh that yes that I think we should regulate you know it’s like where’s this data going and what you’re doing with it all right but uh prevent people from doing bad things doesn’t matter if you are regulated or not it’s just going to happen so like the other thing that I think is really important for organizations if they try to capitalize on it is to look more on the product management product owner side and here I’m getting a feeling that a lot of people don’t understand the urgency the Strategic urgency of actually getting into a defensible position with AI because if you think about it right there’s going to be this generalized models like um GPT or Lambda or some of the other models right so these are these broad big models and then there’s going to be some training that’s going to happen some fine tuning right and there’s going to be specific functions that are going to be created out of this so for example the cancer scan that we talked about earlier and now my question is to all the CEOs out there that are thinking about you know maybe we should be doing something about AI if you’re in the cancer scanning business how many companies worldwide do you think will be left that will have a fine-tuned fine-trained cancer scanning model because I actually think it’s not going to be 5000. it’s going to be three or four because the investment to train models is depending on the use case quite substantial right and um it also like benefits from seeing a lot of cases right so the training to get it to a certain level is one thing but then if you just feed it with more and more data it gets better over time it’s like a Benjamin Button Pro product right it gets younger and better over time it doesn’t age it ages in reverse and so I think um in the end and a lot of these industries where AI is going to have an effect there’s going to be an um there’s just going to be a couple of companies that are going to be left standing and if we assume this to be true there’s actually a race on to develop the AI application in your field and if I look at a lot of companies I think they are not recognizing it right they are not in a race or they’re not behaving like they’re in a race even though they actually are and um it’s not only that that there’s not going to be 5000 specific models for everything there’s also another effect taking place in that um people are waking up to the fact how valuable their data is right so we have always said this term that data is the new oil or data is the oil of the 21st century and so on but if you’re not like Google and have like these masses of data or Facebook and so on it didn’t really apply but now we’re in the situation that if you wanted to create that that specific AI to scan for cancer for example you will have to feed the model with data right and so where you’re going to get that do you already have that do you have the permission to to use it or do you have to buy it how do you have to get it from somewhere for free and now people are actually waking up um how to to the fact how valuable this data is so for example Reddit has blocked um GPT and the other models from using the data on Reddit to train the models right because they’re saying well um these models are very valuable and you are actually training them on our data so we want a piece of of the cake right and so I I see this happening all around that people are suddenly realizing like a year ago they would say here have this medical data you know right they would make sure that everything is legit but they wouldn’t mind giving you the data and now suddenly they’re saying whoa whoa whoa whoa let’s hold on for uh hold up for a second and let’s look at what are you actually going to do are you actually going to build like a specific AI for these cases and then let’s talk about compensation and as people wake up it gets harder and harder to get that data for free or for a good price and so you actually have an incentive to get to get started right now to get the cheap data to train your model and to be one of the three four companies globally that actually have that trained model those are those are great points you know and one one mistake you can never make in product management is not to monitor the market not to be up to date with the situations right but your point is it’s it’s great you know we are seeing this already say wait a minute you’re getting my data for free and you’re making money out of it so where’s my cut on this stuff right which is perfectly fine because it it’s not your data it’s their data right but I also seen a couple weeks ago that that people been used to tag pictures and overall help train the AI they are being paid close to nothing which is abusive right so in that say in that scenario I see regulation being beneficial because yeah we have to make sure that people compensated properly right and we have to make sure that our intellectual property is not stolen by somebody else right with the pretense that they are doing good for the humanity no they’re not they’re lining up their pockets right which every business should but that is also very important for us to understand what people want from our data and how we’re gonna have more benefits from that our financial reasons to let you have access to that data right so I I think there is a lot of stuff that has to be taught and uh I think because for most people it got them by surprise and now it’s like oh we forgot about this oh we forgot about that we forgot about that we didn’t think about that right which now you have to run after that but I’m there with you a lot of people won’t survive because they don’t have the and there’s a broad term they don’t have the business agility to be resilienced business agility is all about resiliency right and in cases like that is when you put to test whether or not your organization is receiving it because if you’re not you’re gonna collapse right and the same token I think about UI and other professionals right if we don’t incorporate a AI in our daily things we are gonna go by the way the dodo bird right we’re not gonna our resilience is minimal or close to not so you always have to adapt to the situation that’s why I said you always have to be monitoring those situations in the market the demand what people are looking for and etc etc anybody who just say well nothing is ever going to change on the you know that’s all I want to do you’re gonna go the way the dodos the dinosaurs and the the fourth pin to you know uh the DeLorean all the other stuff I can have a huge list here right yeah all right any other thoughts on the topic no I think it was a great conversation right I think uh I I initially told you we should talk about this because in the in the latest scrum sessions we talked about this for like 40 minutes out of the hour right and uh it’s a great topic for conversation and I hope when people listen to the podcast say hey uh if I’m not paying attention to this perhaps I should start paying attention right now right it’s about resiliency from Dallas I am Kyo and as always my friend in Berlin has a last word and today he had also had the first one and actually I have something prepared so I would like to end with a quote by William Gibson who is a quite famous um science fiction author and he said the future is already here it’s just not evenly distributed and that means if everyone else is already doing Ai and you’re not you gotta get your act together and go for it all right thanks everyone uh we hope you keep watching us um and talk to you the next time bye foreign