Season 5 Episode 1: Byte: Back on the Mic, Updates, Ai and more

0:01Ravi, we're back.

0:03We are.

0:05It's been some time, hasn't it?

0:07When when was the last time we we did this?

0:112021, according to Apple Podcasts, which is what I had to go and check.

0:15Um, I've got a screenshot of Apple Podcasts here.

0:18It says 28th of April, 2021.

0:21Um that's when it was released, but that's not when we last spoke.

0:24It would have been two weeks before that.

0:25So probably let's say mid-April, right?

0:28Mid-April, yeah.

0:29I mean th this makes it sound like we've not spoken now for three years, and the only the only means of communication we have is

0:35It's via this podcast.

0:36It's a podcast, absolutely.

0:37Yeah, no, but um it it's good to be back, honestly.

0:40I think um uh it's probably worth starting with introductions because we're doing things slightly differently this time around.

0:46So I'll let you go first.

0:47Um why don't you introduce yourself?

0:48Um yeah, sure.

0:49I'm I'm Ravi, um Ravi Mystery um for work.

0:53I work at a company called City Football Group.

0:55I'm head of Football Insights there.

0:57I've used Tableau as my primary data visualization tool since 20 I think it was end of January, early of February 2015.

1:08Since then

1:09Tim and I work together at a company called the Information Lab.

1:13We are both data and tech um professionals, geeks, nerds, whatever you can call it.

1:19I think.

1:21in in terms of yeah, that that's that's what we're about.

1:24Uh my my passive interests lie within like, yeah, using using tech to augment our lives, uh, our work and also like

1:32What is the future of this this space within the industry, right?

1:35And I think, you know, we've we've talked in the past around Tableau specifically, because that's sort of

1:40the crux of of what we what we know and what we work with but there's obviously all these um technologies that surround it um we've talked about databases I think and

1:50another one.

1:51So but yeah, no.

1:53Great to be back and great to be talking about this stuff.

1:55Data literacy is another thing, data strategy, data literacy.

1:58I'm always sort of like two huge topics.

2:01I'd like to think everyone watching this on my channel knows who I am

2:04am but I'll do I'll do an intro anyway.

2:07Um I guess uh the thing I'll start with actually I'm a consultant so you are very much uh

2:13uh uh you know on the other side of the fence from across the bridge across the bridge yeah oh you cross the bridge okay fine fine so yeah when we were at information lab both consultants you crossed the bridge became uh

2:26uh let's just say a member of staff for now uh rather than like a consultant right um and uh for me what I do for a living is I help companies work with data.

2:36Um I work at endpoint digital

2:37I think a lot of people know that.

2:39But if you don't, check them out.

2:40Uh I'll put a link in the show notes, as I would say, for podcasts, right?

2:44That nearly caught me out.

2:46I was about to say the description or comments, but that's not the right thing.

2:48Yeah.

2:49Um there's there's another big thing that's changed, right?

2:51Like I think it's quite a few.

2:54Yeah, it's quite a few.

2:56But the first and primary one is you you've gone big on YouTube.

3:00I have, I have.

3:01Uh by if by big you mean like a drop in the ocean in YouTube terms, then yes, um, you know we've we've we've crossed the threshold, the what the threshold of a hundred thousand

3:11subscribers, which is great.

3:12Um but I mean that's the p the plan was always to try and do the podcast alongside the YouTube channel

3:19because I think um you videos videos tend to help people do very specific things but what the podcast allowed us to do is to talk about sort of the peripheral subjects around the the many things that encompass not just tableau but the analytics sphere in general

3:32And so um this was very much a an opportunity to scratch that itch in a in a more, let's say, structured format.

3:39Uh we tried to put some structure to it and um

3:42I guess one of the only bits of structure we let down is uh consistency.

3:46We we stopped for two years.

3:49So maybe not on that one, but uh

3:52Um, we're back.

3:54We're pushing through.

3:54It's never too late to keep trying.

3:56So um yeah, we're back.

3:58So yeah, no, really, really glad to be back.

4:00Um obviously a lot's changed for us in terms of life as well, you know.

4:04Um

4:05I think our jobs have become vastly, vastly different since uh big split, hasn't it?

4:11Yeah, exactly, exactly.

4:13Um I think you've seen more of the, let's say, applied side of

4:18uh of of of analytics whereas I've seen more of the delivery and solution side of it.

4:23So I think we have an interesting um let's say

4:27We stand at different sides of the same room, right?

4:30And we kind of see problems very differently.

4:32And in the same time, lots of new tools have come in to fill the space between, right?

4:38Because analytics is still one of these areas where

4:40the solutions of yesterday aren't solving the problems they have to.

4:45I think the other the other thing is the difference between someone who's working as a member of staff is

4:49you say and the consultant you work very slowly within as a member of staff right like um right you know it's

4:59Talking about Tableau, we we don't upgrade maybe once or twice a year, if that.

5:04And usually it's either a security patch or a yeah, we want that beach or probably upgrade sort of moment.

5:11rather than a no we want the latest and greatest because we need to be using the full feature set.

5:17Same goes for problems, right?

5:18Like you're solving more specific business problems with a toolkit.

5:23Rather than the opposite way around where you're bringing the toolkit to the problem as a specialist to accelerate and then you almost dip out.

5:31Um whereas it is almost like the um the fallout of of the work done um is really what what we end up dealing with, I guess, inside rather than outside as a

5:43Yeah, exactly.

5:44Um and you know, in in the in our in our podcast, we kind of had these three types of uh of formats, right?

5:51And I think it actually speaks to this

5:53this challenge.

5:53So I'll briefly touch on them.

5:54We have bytes, which are essentially just a normal show, us talking about specific topics.

5:59We have bits, which are um essentially snippets from shows.

6:03So maybe us taking us an extract from the show or talking about a very small topic.

6:07For a very short amount of time.

6:09And we had analogues, which were sort of discussions with guests.

6:12Um, we have a famous uh, let's say, curse on this show that everyone we seem to talk to ends up leaving the company they work for.

6:20So

6:20Anna Casey, who I think who was our last guest, left has left Tableau since.

6:27Yeah, yeah.

6:28Before that, uh Kent.

6:30I'm not sure we can say it's a curse day because it's only happened twice.

6:32If we had Borg

6:34more uh guests and we actually got on top of that maybe we'd we would sort of turn the tide a little bit but nonetheless we'd average out in the middle somewhere yeah yeah exactly exactly so that that's sort of the broad format of the

6:44the show and that's sort of how we kind of break this up.

6:47The other new thing here is obviously video.

6:49Um uh yeah obviously I know this but video has become a big thing in the last three years so

6:55Um it's really important for us to sort of add this element to um the podcast, but actually the podcast is the main thing.

7:02The conversation is the main thing, and if this is another avenue where we can share that conversation.

7:06Then that's great.

7:07So we're gonna host the videos uh for this, the podcast actually, um, on Apple Podcasts, Spotify.

7:14You can find it in the usual places.

7:16But we'll also have the video podcasts.

7:18uh here on the Tableau Tim channel as a podcast.

7:21So uh YouTube does a pretty good job of separating out my usual videos with the podcast.

7:25So if you just want to follow the podcast uh there's a way to do that through YouTube.

7:29and through through Google, but if you want to catch everything else, then of course just use Apple Podcasts and Spotify.

7:35That will get you going

7:37So, um, what were our last two shows?

7:42We've also spoken about Anna, um, and then the previous show to that was about NFTs, Snowflake, and agriculture.

7:49Gators.

7:51NFTs are not.

7:52I think he had a hot take on NFTs.

7:54Yeah, they're still rocking.

7:57They're still rocking.

7:57I think um and NFTs I think will still play a big big part in

8:03the work life in in society.

8:05They just won't be monkeys or sorry, bored apes or like JPEGs that people are using, right?

8:12Um I think the the authenticity part that we talked about in that podcast

8:17is gonna become even more important like that, especially in the world of AI, right?

8:21When we will talk about that today, I think.

8:23Um but being able to trace the route of where something came from and who owns it is going to be really, really important.

8:31Yeah.

8:31I mean blockchain's a great example of something that everyone uses but doesn't really realize they use.

8:36Like if you've ever done a DocuSign document that is using encrypted blockchain technology.

8:41Um to in order to make sure it's you and there's a digital fingerprint on who you were, where you signed, what device you signed from, um, rather than you know just typing your name right.

8:52Even more necessary now in the world of AI.

8:55Um, we'll get to AI shortly, because I think that has pretty much taken over every f every technology company on the planet seems to need to or needs to be showing that it's working.

9:04working with AI.

9:05Um but it it's interesting because I think maybe NFTs were the start of that journey.

9:11There were NFTs you had blockchain, you also had uh crypto sort of have its big crash and

9:16and rise, it seems to be on the rise again.

9:19Um and I I I I'd sort of take one step back and say that's also when analytics products I think started to feel like they were becoming stale.

9:28Is that fair to say?

9:29I'd say I'd say so.

9:31But I I'd argue against whether NFTs and blockchain almost push towards AI.

9:36Like yes and no.

9:37I think the the biggest push for AI in general is really the the computational resource.

9:42Um, I think compute is cheap these days.

9:45Um the the like network like network computing is a big thing as well.

9:50Like you're able to harness multiple computers around the world, even desktops in some cases.

9:56just to create this like bootstrapped supercomputer that is able to do multiple levels of processing.

10:02We saw this during COVID.

10:03Um I was part of um I found this app Vodafone we're partnering with we could leave your phone on overnight

10:10And it will be contributing to like your internet usage and your phone usage would be contributing to the algorithms cracking through what hopefully or could have been one of the vaccines.

10:21And this was the first idea of like network computing that I saw.

10:24But I think that you know AI and large language models for sure are really the basis of just computing rather than

10:31um anything to do with blockchain.

10:33I think both of those things will definitely be enhanced.

10:37But I don't think it will be the be all on end or

10:41I will argue that I think NVIDIA was at the heart of both of those trends.

10:45And I wonder if they would have been able to capitalize on AI.

10:49had they not benefited from the boom of crypto that they also benefited from?

10:55And they were sort of there for two big trends.

10:57And I wonder if the first one was a necessary

11:01Was a necess there was a necessary upside in that for to allow them to invest in AI to make them become the de facto currently they're basically the only show in town.

11:11Yeah, that's interesting.

11:12But to answer your question, um how did is almost AI coming or coinciding with analytics products becoming stale?

11:20I don't think they became stale.

11:22I think they stopped moving.

11:24Um, which might be two sides of the same coin.

11:27Two sides of the same coin.

11:28Yeah.

11:29Um and by that I mean I is you almost forget

11:34Why you existed in the first place.

11:37Like the problem you're trying to solve, is it still the same problem?

11:39Have you solved that problem in some cases?

11:42And do you need to think of what the new problem is?

11:44And I think.

11:46Not not having the ability to stop and take stock and be like, you know what?

11:50And this is this is the problem of any company in general, right?

11:53You can't just stop and be like, we're gonna uh just

11:56Think for a moment.

11:58What are we doing?

11:59Why are we doing it?

11:59How are we doing it?

12:01Um Yeah, you don't you don't you're not able to do that.

12:04So dunno.

12:05I think st st going stale is

12:08Is because they they probably backed the wrong horse with the platforms.

12:11Right.

12:11When when they should have when they went to become a singular platform that served everything.

12:17They should have been a disaggregated way.

12:19We're like, you know what?

12:20You can plug us into anything

12:21And we'll do our job is this part of whatever you want it to be.

12:26Yeah.

12:28Yeah, it's an interesting one.

12:29I I I still think like

12:32The the moment you realize uh an industry needs innovating is always a year or two later than it should have been doing that change.

12:40Right.

12:40So when I say 2021 is when things were going stale, I don't think I could have gone to the 2021 version of Tim and said things were getting stale, because I think at that time I was still excited.

12:50Um, but I think looking back at it now, um, and you can see the the the major pivot that a lot of analytics companies have taken

12:59I think it's clear to see that actually they lost they lost their focus on the mission, which was just helping people answer questions in a most in the most quickest and efficient way.

13:07And instead they almost productized and focused on

13:11on how to not how to do that for you, but it's your it's your point you keep making about dashboards are gonna die, right?

13:19It's

13:20It is the thing that people need dashboards or is the thing that people need answers?

13:24And what's the fastest way to enable people to do that?

13:27And I I feel like with AI they've had the second wind of realizing that it's the answers that matter more than the dashboards, right?

13:34Right.

13:34Yeah, yeah, exactly.

13:36And and I think it's it's this is where the rethink will happen.

13:40I think there's there's a popular um visualization company that would say, yeah, dashboards are dead as their big slogan as a big sort of shock factor.

13:55Yeah, I mean I've never said dashboards are dead.

13:57I think that the future is dashboardless.

13:59Right, like almost like serverless is a thing where you have an era like compute compute that has um no um no sort of head effect.

14:10It's just it turns on, does its thing, turns off.

14:13And this is what I this is almost effectively what I was trying to say like a couple of years ago when I wrote

14:18about that.

14:19But I think the yeah the the the the ability to ask a question and get a rich answer is nice.

14:27I think LLMs will be able to solve some of the problems.

14:31I right now I see them as assistance, right?

14:34Like they're then they're not a knowledge hub yet.

14:38You know, you've got all these like

14:40Shadow AI stuff like if you let's take a really simple example where you give it some SQL code and you say this doesn't work.

14:47Why?

14:48And it'll be like, oh, you need to build this CTE and then you drop it in.

14:51And it's like, well, that still doesn't work.

14:52It's now doing what I didn't want it to.

14:54Oh, sorry, I misunderstood your question.

14:56Here this one should work.

14:58And really that that that's become the consultant, right?

15:03The intern consultant, yeah.

15:05It's the the the experience isn't quite there.

15:08You still have to guide it a little bit.

15:09But but yeah, it's it's actually capable of doing the things with a very prescribed set of directions.

15:14Yeah.

15:15And this with this is why like almost on LinkedIn, on Twitter, uh you you see those viral trends that say like

15:23Here's your the top hundred AI prompts you need.

15:27So well you don't really need an AI prompt.

15:29You need to work out how you use GP chat GPT to answer your questions.

15:34um what the next thing so yeah I think analytics is changing I think the flow is changing is probably the more important pertinent question here what is the analysts flow in 2024

15:46And who's the analyst as well?

15:48I I think there's a there's a bit of a another refocus on redefining analysts to be a much broader set of people.

15:55Not just the people who build these forms of analysis, but also the people who consume them are being brought into that description in a more deliberate way.

16:04Anyone can be the analyst because

16:06at its heart, the the metrics and the things we all track, uh, that they should be common within the business.

16:12And therefore, why are we gatekeeping that metadata and that information?

16:17To just the people who build those solutions.

16:20They should be more broadly readily available to more people.

16:24And that's sort of, again, that's where the thought spot mentality comes in.

16:27Right.

16:27Um and and to be fair, many many analytics tools are having to add this element of capability to their products.

16:36So you've had DBT come in with

16:39um the whole uh data modeling capabilities and metrics being uh an important part of that.

16:45And then from the other end you've had Power BI has been doing metrics for a while.

16:48You have Tableau enter the game with Tableau Pulse essentially metrics, you know

16:52with the with a big boost with AI at the heart of it.

16:55Um the key disappointment I have with this though is that it turns out

17:01st the the big s the the big the fang as it were Facebook Apple Amazon Google are still the only companies able to train their own models right yeah and fundamentally everyone else is just a customer to

17:15those models at this stage where we're at.

17:17That the stuff is still too expensive.

17:19The NVIDIA GPUs are basically not available unless you're one of these big companies with big checks.

17:26And so really every model we're looking at, and you know, if you take Tableau Pulse and we look at Tableau's use of AI as an example, it's using Chat GPT behind the scenes, which means it's limited.

17:37by the scope of what chat chat GPT can do.

17:40It's not a model trained on Tableau's knowledge bases.

17:43It's not a model trained on all the calculations.

17:47But then the key thing is.

17:49How far are we away from that?

17:51That's that's actually been my big disappointment.

17:53I thought when, you know, ChatGPT came out and GPT 3.

17:575 was open source, I thought, wow, that's going to show the way to all these companies to get

18:02Even if you had GPT 3.

18:045 levels of capability and you trained your own model, I assumed that would be attainable within the next year.

18:11Instead, it just turns out the existing models are moving faster and faster and faster and no one else can get close to the hardware that you need to build the more basic things in the first place.

18:21Yeah.

18:23I think the other the other interesting area um a around GPT, especially.

18:30Um

18:32Is people didn't really know what to do with it first.

18:34Right.

18:35And then suddenly you had to do something.

18:37Like I think um I can't remember exactly what maybe 2020, 2021.

18:42I think I shared with you like an early version of GPT.

18:46Uh Oh, you did, and I completely dismissed it, yeah.

18:51Uh you did it on our work chat.

18:53uh convo at the time, right?

18:54I think it's probably still up.

18:55If you're at the information lab, go Google Ravi's post uh about GPT.

19:00I think it was two point two point five.

19:03It was two it was it was early.

19:04It was it was two it was early.

19:05Yeah, very early GPT, yeah.

19:08Yeah, and I and I've seen that.

19:10Gotta give a shout out to Joe Mulberry who shared it onto onto my feed and he was almost like, this could be interesting.

19:16And I was sort of thinking, like, hang on, you've just given it a natural language prompt.

19:20And this thing has just gone away and done it for you or assisted you to do this based on everything that it's learned and known in the past.

19:28I think the the the problem you've got is

19:30the speed at which this is just s splashed, right?

19:33Like it's come as an asteroid into the sea and there's a wave came and everyone's r trying to ride that wave.

19:39ensure they're doing something with it because every webinar, every person, every tech company, every even employee that I work with, like, yeah, but how is AI gonna change this?

19:50It's like

19:51you know, this is my smart question I can ask you.

19:53Um and and the real answer is well, dunno.

19:58Like in so in s in some cases we've already using AI.

20:02In other cases

20:04it doesn't make sense to use AI.

20:05And you have to be really purposeful in my opinion to start using the compute of this.

20:10The problem is, as I as I alluded to, you can't just stop as a tech company.

20:15You can't just be like, we're gonna kinda like work out

20:18what we're doing in this space.

20:21That said, Google and Apple absolutely have stopped.

20:26Like the like Bard came Bard and Gemini came out.

20:29Apple still haven't really shown or done anything

20:32with with these.

20:33Yeah.

20:34Yeah.

20:34And they're almost like sitting there by a sign and being like if we still got Sarah, you can crack on with that.

20:39If we've got something good we'll tell you.

20:41And it's almost they've got ability to to be solid there, I think

20:46I think Apple and Google actually represent the challenge many big companies have with AI, which is what do you do with it across your product suite?

20:56Because

20:57Chat Chat GPT and OpenAI don't have existing products they need to integrate OpenAI with.

21:03So they actually built the most

21:05common interface that you would want to something like this.

21:09And they did that well.

21:10And that's why it took off.

21:12Everyone else like Google and Apple have the burden of having to need to integrate it with their technologies.

21:17And so that actually makes it immediately possibly more useful, but also immediately easier and harder to to to do well because the low the bar you have to cross when you have a billion people on Gmail, the bar you have to cross when you have a billion iPhones.

21:31It's just so high.

21:32You can't just rock up to, I don't know, anthropic and say, hey, can you run your model on an iPhone?

21:37Because as soon as you do that, yeah, it's a it's

21:40It's it's definitely expensive.

21:42Uh you'd need a lot more.

21:44NVIDIA would need to be five times bigger to keep up with the demand the iPhone alone would generate.

21:48So um

21:49The really interesting challenge is that, and I think companies need to watch that space a little bit more closely because it's exactly how Google and Apple integrate AI into their business.

21:59that are going to set the the standard for how you should do it within an analytics team or within a business uh that that's using analytics as a capability and want to mobilize that data.

22:09For the purpose of AI.

22:11On the other hand, I think they've both realized they were both caught sleeping.

22:16Funny, Google was caught sleeping, GPT.

22:18They invented the T in GPT, Transformers.

22:21Yep.

22:21Um nonetheless, I think Google came from the side of caution and realized that there was actually an angle where if you're Google, yes, you have to be cautious, but if you're open AI and no one knows about you,

22:32you don't and so you can just make mistakes and have absolute howlers because there's no damage to your reputation but when it when you google and that happens and you're

22:42you know, pride yourself on cataloguing the world.

22:44You can't make mistakes.

22:45They had a recent issue with um uh their model generating non-accurate uh racial representations of

22:53Of historical um scenarios in society.

22:55I think it was generating uh Nazi soul soldiers of color as an example, which was a pretty bad mistake.

23:02But you can see how

23:04You can see how some sort of dial on um, you know, being equitable was dialed too far up.

23:09And that's how that mistake happens, right?

23:11Um and and Apple, on the other hand, well, they haven't done anything and they they they seem to be just publishing papers at the moment.

23:17to kind of demonstrate that they're catching up.

23:19But as they say, you've not done anything until you've shipped.

23:22And the the challenge they have is actually

23:26You know, out of the gate, there's probably only one model that can deal with the with the scale of of the iPhone, and that is again Chat GPT.

23:33Many of the other models probably couldn't deal with that.

23:35Um, and so

23:37You have lots of different challenges all coming out at the same time.

23:40But yeah, uh, Apple and Google definitely caught sleeping.

23:43Um they're good case studies for Harvard Business Review, maybe in a couple of years' time when this is all shaken out and

23:51Yeah, I'm really excited for Google to really step into this game though.

23:56Like genuinely, because the moment that Google Photos said, you know what, we're not going to be free anymore.

24:02was like, okay, they've got everything they need from all of your pictures, guys.

24:05Like they've scanned everyone's photos and understood them at a high level for a long, long, long, long time.

24:12And now they're like, yeah, we're gonna charge for this.

24:14Or photos have become that that's the cynical answer.

24:18The real answer could be photos have become and videos become so big.

24:21Hosting them doesn't make sense for for Google anymore.

24:24Yeah.

24:24But like Google have that as their cache of sort of image recognition and image development.

24:30And then you've also got on top of that the entirety of YouTube.

24:35Which just opens up an entire massive broad knowledge base of everything humankind have come up with.

24:45Um right, wrong, biased, unbiased.

24:49What like reviews, opinions, voices, all of this stuff is is real.

24:56Nishes as well.

24:57Like uh uh I was listening to another uh YouTuber and he was saying

25:01how like every niche has a video with like a million views or something like that.

25:05Like and so if you have a niche and you don't find that video, you should just go make that video because it

25:10like YouTube will find a million people for you.

25:12And it's yeah it's it's a very it's a very interesting problem.

25:15I I will say that Google have have done a few things.

25:19It's just it's not at the same scale and impact.

25:21Chat GPT, I think because they're not the first mover advantage here.

25:24Gemini Pro, for example, I actually use this quite often and I I like it.

25:29It's actually quite good at very specific

25:31types of tasks compared to Chat GPT.

25:34Its context window is a lot larger.

25:36That's a huge thing if you're trying to assimilate a lot of information.

25:39That said, the thing I keep hearing a lot about is you know things like

25:43Plexus B, Anthropic, all of them are reaching GPT-4 levels of capability today.

25:50And people assume that Chat GPT is just gonna stand still for a while.

25:54And I know for a fact what will happen in two, three months is uh

25:57you know, Sam Altman who at the moment feels untouchable given that he might manage to survive his own ousting and and come back to the to the company as CEO.

26:06Um is going to announce GPT 5.

26:09It's going to blow the world out.

26:10They announced Sora just a uh you know a month ago and they're already partnering with people in hot

26:15Hollywood.

26:15Um OpenAI Whisper is a great transcription model.

26:18We'll transcribe this podcast absolutely flawless.

26:21And you think actually

26:23the the the group of models openai is building, DALI, Chat GPT-4, Sora, like actually, no, like.

26:31One model from Google, one model from Apple is not going to compare to the the the momentum that Chat GPT has.

26:38And if they put all of that into one generalized model

26:40That is going to be sensational.

26:42The hard part, as you say though, is stitching it all together, right?

26:46Yeah.

26:47Jarvis, right?

26:48Like from Iron Man.

26:49That's effectively what

26:51Almost it this this edge could end up being, you know, with the Apple Vision Pro also coming along around the same time.

26:56You've got this like

26:58May maybe we start suddenly at a race or on augmented reality again.

27:02Like Google Glass that failed project might come back around again.

27:06Um

27:06Because again, computationally, it's now these things are now possible.

27:10Like we understand computers better than we ever have done, and this will continue to grow and develop.

27:15Where like

27:16Consumer electronics might be at their peaks, right?

27:19But it it's now beyond consumer.

27:22What what's possible?

27:24And that's that's quite exciting.

27:26I feel like hardware has you're definitely right there.

27:29Hardware has peaked and it's now software that's kind of it all the frontiers are to do with software or processing power.

27:36But the hardware itself, everything else around the hardware is solved.

27:39Like, you know, we all know what a phone should look like and feel like.

27:42We know what a laptop should look like and feel like.

27:44All we're really debating now is what goes on in them.

27:47And that's that's exciting on one hand because these are these are sort of well-known form factors.

27:54Um and AI has a potential with with the you know augmented reality example, I think AI is the absolute missing piece that augmented reality lacks.

28:04You remember like five years ago when we all had these AI

28:06apps on our phone, we thought, hey, look at me.

28:08I'm playing cards on the table.

28:10Like very, very basic stuff.

28:11Uh then you take the meta glasses, the Ray bands, and you look at the video quality that's coming out of that.

28:16And then you hear that Facebook is going to put llama inside of those glasses so you can talk to it.

28:21Then you're like, oh, I get it now.

28:24So you can ask the thing, what am I looking at?

28:26And it's going to tell you, give you the context.

28:28Nope.

28:29Now you're talking.

28:30Yeah.

28:30Now, antitrust, privacy, all of these things are really gonna come out and play now.

28:37Like you're gonna have to get some proper

28:40legislation and people who understand tech to an degree and ethical technology.

28:45Because like even like let's take the Raybounds for example.

28:48If someone's filming everyone that they see, every conversation they have every day and then reviewing it, that's creepy.

28:56It's not sorry, it's not creepy.

28:57It's

28:59It's a violation of privacy and trust.

29:01It is creepy.

29:03It is creepy.

29:04Doesn't mean it's wrong.

29:05There are lots of creepy things that are wrong, right?

29:07And um and the lots of creepy things that we we all do that we

29:11you know, we get on with like in in the UK today, it's absolutely illegal for you to go out and take a photo in public of anyone, no permission required.

29:18Uh

29:19Is it the same if you're wearing glasses at a constant recording all day, right?

29:24And uh is it the same if that that interaction is extended?

29:27If that interaction then picks up a conversation that you are not a part of

29:31Correct.

29:32But it is occurring in a public or private place.

29:35Like there's there's so many stipulations you have to unwind.

29:38Like I said, this is for lawyers and smarter people than me, or more clout than me to jump into.

29:45Um antitrust huge.

29:48Um there is the ongoing antitrust stuff with Apple and uh you know uh

29:53uh the DMA in Europe and uh we're not experts on that so uh we won't go into that.

29:58I absolutely love those debates by the way because as an Apple user who sits here getting for

30:02frustrated with their iPhone and the lack of integration of basic things that have been in Android for years uh just don't appear on the on on the iPhone.

30:11I'm like, this is why this company is too comfortable.

30:14They need a little bit of stress.

30:15Um, but you know what will happen?

30:17Um we'll go through all of this uh you know debates and and all that.

30:22Same with AI, I think, and we'll end up in a place that's worse than where we are today for both sides.

30:28both for the companies and for consumers.

30:30That's all that ever happens, right?

30:31You have the EU with their cookie policy things.

30:36Yeah, exactly.

30:37And then the EU thought, oh, websites will stop putting cookies on their websites if we just tell them to ask the user.

30:42Nope.

30:43What did every company do?

30:44Ask you every single time.

30:47It's the exact opposite.

30:48So GDPR, like I still get tons of spam, and I'm you know, you can do erasure requests all you want, but another data breaker will just sell your data on and back in back into the thing.

31:01This is the this the way that

31:03Um telemarketers have found a way around this.

31:06They just use middlemen.

31:07Right?

31:07Like you you get a spam phone call, you tell them, don't use my details.

31:11Yeah.

31:12Oh, sorry, you got this from this company, you have to contact them and then who then tell you to another one and another one and another one.

31:17And you just get stuck in this web and that what they're relying on is you not wanting to do that.

31:22Data breakers, yeah.

31:23I've got I've got a friend who like

31:25just in the spirit of nudges that are barriers.

31:28I've got a friend who um had a beer fifty two which is this beer subscription service.

31:34paying £30, £35 a month.

31:37Um and the way to cancel your subscription after the free trial is to ring them and I'll and speak to someone and say, this is what I want to cancel.

31:45Now being the millennial that we are, no one wants to pick up the phone and speak to a human being.

31:51So just kept rolling out of like, ah, I don't really want to spring someone.

31:56Um and then yes,

31:59Yeah, yeah, exactly.

32:00What what is that worth to you?

32:01And and it's creating exact it's exactly that, creating that barrier.

32:05But on on on untrust like

32:07We we've got an election coming up here in the UK this year, probably, um, maybe early next.

32:13And I I just know disinformation videos and audio clips and all of these things are just gonna be rife.

32:22Right, like it's not it's not gonna be quite like Cassette Boy, that YouTuber that did like split things of um of politicians and and Alan Sugar off the apprentice.

32:32It's gonna be to another level.

32:34Like you think if you even look at the small segments around um Kate Middleton.

32:40And the the the um Easter picture.

32:42Is it Easter Picture?

32:43Photoshopping.

32:44Yeah.

32:44Yeah, Mother's Day, it was Mother's Day, wasn't it?

32:47Photoshop plus, that's not her plus.

32:50That video of her um giving her news is also AI.

32:55It's like, come on guys.

32:56Wow.

32:56Like, and this is where going all the way back to start, NFTs

33:02And blockchain are gonna have to become really come to the fore here.

33:06How do you have that trust of where that source came from?

33:09Um

33:09Um, you know, when you take a picture on your phone, you have all that that information stored within the metadata.

33:15Like how how do you actually utilize that?

33:18You know, in a way we're talking about data lineage in the real world, right?

33:21Like it's kind of to bring it back full circle, it's it's funny how the things we we dealt with uh for you at work, for me as like a consultant

33:30Lineage, understanding where something has come from, even at its most basic s like um form

33:38That is not solved yet.

33:39Like we have copyright, which is very much a data thing.

33:42You stamp an image, you you apply for a trademark, and that's where you go.

33:45We've now reached a point where that is no longer good enough in today's world.

33:49Um, you know, um I when I edit videos, there's a lot of course by the way.

33:54It's really hard to enforce now.

33:55I think it's very hard to enforce it.

33:57Hard to enforce, yeah.

33:58Yeah.

33:59It's not only hard to enforce.

34:01I don't think the laws that allow you to do anything with it, right?

34:04Um, you know, in in a very basic sense.

34:06So here's a here's a very basic example that

34:09Um I kind of I kind of use, let's say uh I make a video and I post it on YouTube and it defames someone.

34:17So they come after me and say, hey, you've defamed me.

34:20Here's the lawsuit.

34:21And I'll go, actually, that wasn't me.

34:23And I and they'll go, what do you mean?

34:25Uh it wasn't me.

34:26Um it was an AI.

34:27The script was actually automatically generated and it was

34:30It was compiled from a bunch of sources which were pulled from public opinion.

34:35So if you want to go sue someone, sue public opinion.

34:39But I'm just using an AI to aggregate all these thoughts.

34:42Um the only thing you can sort of probably go and do me for as a lawyer there is the fact that I have it on my channel on YouTube.

34:49But then enters TikTok.

34:51Okay.

34:52Then enters shorts, then enters all these platforms that have very disposable profiles and everything else.

34:58And then you've got nothing to stand on.

35:00And what do you do in the instance where someone's taken a screen cap of your video uploaded to TikTok?

35:06Someone on Instagram has taken it off TikTok and put it on Instagram of the YouTube source, right?

35:13And you just get this all the time.

35:15You see this all the time where like you're just watching a recording of someone else's video on someone else's account.

35:20Yeah.

35:20And that's how you're getting the numbers uptick.

35:23Uh so you need you need something that has to work across platforms.

35:27It has to be multi-company, it has to work across if if you take social media example, you have to be YouTube, Instagram, Rails and TikTok, all of them.

35:35Not that these companies will ever collaborate with each other.

35:37TikTok's about to get banned in America.

35:39So um that that it that in itself is interesting.

35:42And then at a very sort of root level.

35:45Like you and I were brought up in a world where we were we were taught the skills on how to discern whether information was trustworthy or not.

35:53Um I'm not saying that those skills aren't still being taught today, but they're definitely harder to teach because

35:59the the pace at which information is being generated at far outweighs the uh the pace at which any one person could um you know sift through them.

36:08So even if you had the right skills, you and me today.

36:11Uh Chat GPT could be creating tons of tons of tons of pieces of content.

36:16The election you talked about, probably the first election where I'd be surprised if 50% of election material isn't created

36:23using some sort of AI assistance because it's dead easy to just put in a prompt and say, hi I need a paragraph about this, so I can put it in the leaflet.

36:31Out it comes.

36:31And it's

36:32just a couple of edits and off you go, right?

36:34Before you hired interns to do this, you hired copywriters to do this.

36:38And you're still going to hire those people, but they might be editing instead of crafting, which is which is in itself interesting.

36:45I I'm wondering whether Yeah, I mean this this is this is this is why we brought the podcast back, right?

36:52Like these conversations where we can just keep going and going and going.

36:55Going and going, endless.

36:57And endless.

36:59But like as I said, I wonder if there's you know, is there any anyone or group of people who are really experts in this?

37:07space though.

37:08Right.

37:08Like I really you're looking for an academic that's thought about AI dystopian in a dystopian manner and research and written papers on it or group of researchers.

37:18You really want academics to have this conversation, be leading this.

37:22But the problem is who's listening to these academics?

37:25Who's who's finding those papers and understanding the right people and also then the decision makers?

37:32that like it's only been till recently that you know the BBC have got uh any sort of uh understanding of where people are watching iPlayer, right?

37:42Like there was legislation that meant that if you unplugged your laptop

37:45I think someone had told me or uh I read somewhere.

37:49If you're using some a wife someone else's Wi-Fi but your laptop or mobile device is unplugged, you're technically covered by the TV license you pay at home

37:58So therefore for a university student it would be that that would be like I I don't know what the ins and outs are, but like legislation I think was then had then had to catch up a few years ago

38:11to then block those loopholes and those areas.

38:15Now famously legislation is really quick to to be designed, developed, and passed through any sort of government.

38:21Yeah, right.

38:22So

38:23Yeah, it's it's it's I think it's such an interesting space that you know I'm I'm definitely not an expert on, but it's I w it's gonna have to be

38:32Good.

38:34I'll pull one out for the UK because we left the EU and all I'm gonna say is that I'd hate to be a tiny little island.

38:41out in the middle of the sea trying to form its own opinion on this when at the moment all the movements being done by what I can only describe are continent level groups.

38:50The EU, America

38:52We're just gonna have to conform to all of this stuff rather than feed into it.

38:55And even if you do even if you do decide to come with a different policy of all these countries, they're just not gonna listen.

39:01Um it's just too small a market now compared to the rest of the the block.

39:05And the only place you can enforce it is your own country.

39:08Because everyone else is like, well, no, we use this one.

39:11As do two hundred other countries.

39:14Yeah.

39:14Yeah.

39:14Yeah.

39:15Yeah.

39:15Yeah.

39:15Yeah.

39:16I'll keep I'll keep uh keep pouring one out for that.

39:19But um I think when we were doing our podcast, the debate had just settled, right?

39:23It just happened.

39:24So it's funny.

39:26We probably started our podcast round about when the debate was up.

39:29actually going on and here we are um three years later back to talk about the effects but hey um we should probably uh probably call it there and start talking about what we're gonna be doing next, right?

39:42So we've had a little bit of a chat today about AI.

39:43I think that was like

39:44the the topic that was top of our top of our mind.

39:46We talked a little bit of analytics as well.

39:49Going forward, we're going to try and do something every fortnight.

39:52So we're going to try and

39:53Make sure we keep this uh going.

39:55So um two podcasts a month uh on Apple Podcasts, Spotify, and of course YouTube and any other video platform that takes

40:02podcasts we'll be doing that.

40:04We'll we'll get the website back up.

40:05We'll do a little bit of a rebrand as well just to get uh you know back in the day we called ourselves business intelligence professional

40:11professionals.

40:12I think um they're just called data analysts uh today.

40:15So we can we can we can rephrase a few things and uh

40:20uh change things up and you've also you've also left um uh the consulting space so I think data analysts or something else of that kind is a little bit more fitting as well.

40:29Um

40:30But yeah, uh love love this debate.

40:32Love chatting to you, mate.

40:33Um great to have this great to have this back.

40:36Um yeah, I can't wait.

40:38I can look forward to many, many topics to be discussed.

40:41I think we've we've scratched ourselves some really interesting ones here, so

40:44No, yeah.

40:45I start again in two weeks.

40:48Cheers everyone