Subscribe
Share
Share
Embed
The boys have a whisky-infused talk about AI use cases, share fun stories and have 4 different English whiskies throughout the episode.
Creators and Guests
Host
What is Office 365 Distilled?
This podcast is presented by Marijn and Steve; two experienced Microsoft 365 and SharePoint consultants. In each podcast you will get a smile or two with a chunk of interesting and practical help in Microsoft 365. Then you get introduced you to what we hope is a Whisky you have not tasted before.
Hey. Hey. So let's just go with it. It's kickoff. We got water.
Speaker 1:We got whiskey. Whiskey. Whiskey. Plural. And whiskeys.
Speaker 1:We got you.
Speaker 2:You got... We got you?
Speaker 1:You. I'm here. We got Just a minute. Ouch. Yeah.
Speaker 1:Yeah.
Speaker 2:I'm here. We got our devices, everything set up.
Speaker 1:Plugged in, ready to rock and roll.
Speaker 2:Nice. Nice. Nice. Nice. Nice.
Speaker 1:So we could find something interesting to talk about. You said something about skills. I said something about AI. I said, I actually know. I'm gonna go back a little bit here.
Speaker 1:So today's gonna be a whiskey tasting because
Speaker 2:of Poor
Speaker 1:us. I know. Because of East of England Power Pour Shell conference. Hey, guys. We had a great time in Norwich.
Speaker 1:Thank you very much. And... There were
Speaker 2:a lot of very cool people there.
Speaker 1:There were. Somehow. Yeah. It was a really... You know something?
Speaker 1:It was a community event. If one room could even hear the other room speaking, they were sort of in these these sort of open aired spaces. Yeah. So every time... I actually...
Speaker 1:I have to say, guys, I should apologize. I set our audience up. We had about 60 people in. Yeah. So I said, look, if they clap loud, I'm gonna do this, and we need to cheer louder.
Speaker 1:So it was great fun.
Speaker 2:Nice. Nice.
Speaker 1:So I apologize. I don't know who was speaking in the other room, but if you were in Norwich and you were getting annoyed, that was my fault. There we go.
Speaker 2:Yeah. Let's let's talk about conferences first.
Speaker 1:Let's do that.
Speaker 2:Because it's been quite some time since we've
Speaker 1:been recorded. So... Yeah. Yeah.
Speaker 2:So you went to the Norwich Yes. Or was that East Anglia?
Speaker 1:That's the East Anglia Power Platform
Speaker 2:User Group, community, events, conference.
Speaker 1:It was a community. I mean, it was one of the things I really enjoyed. And Duncan and the crew, yeah, kudos. Because their keynote was about how they got started and how they moved from being just evening events and the support they got from all the people through to where they were now. And the whole day, even though I have to say we had to leave early because I can't go to England without seeing my mom, would have killed me.
Speaker 1:And so we we kind of shut off just after lunchtime. Mhmm. But no, it was a a really friendly cool stuff because we all try and do big ish conferences, don't we? Everybody wants to kind of do them formal unless you're Jamal. Of course, Jamal has a great way of just saying, guys, we're here to teach you and show you and we'll not worry about who introduces what.
Speaker 2:Yeah. But...
Speaker 1:Yeah. Mhmm. So it was good. I really enjoyed it. And we did the Microsoft one.
Speaker 1:The... One that you were away for?
Speaker 2:Oh, yes. The AI CD. The AI Community Day Brussels. Yes.
Speaker 1:Yeah. And that turned out to be the same kind of thing. I mean, that's mainly because I was kind of hanging it together, so everybody had no choice but to get involved and have fun. But it turned out to be
Speaker 2:good. Nice. Okay, okay, good.
Speaker 1:Alright, now I did all the ones where I travelled all of, you know, a couple of hours. Where did you go?
Speaker 2:I had a little longer travel. So first of all, I went to Orlando to the m three sixty five community conference, something like that. Yes. So the the big one. The big
Speaker 1:Just a second. Open your T shirt, and I'll tell you exactly what it was called. It was a Microsoft community conference. There we go. The twenty twenty six Orlando Microsoft three six five community conference.
Speaker 2:Yes. So the one that is once a year in Vegas, and then the other year, it's in somewhere else in Orlando most of the time. So I've been to a few in in Vegas Yeah. Because I like Vegas.
Speaker 1:You've been to a few in Orlando?
Speaker 2:I've been to
Speaker 1:a few in when I was there. Nights in Orlando. Yeah. Yeah. Actually, that is very true.
Speaker 1:It is nice. All the bars close at 02:00, and it's nice.
Speaker 2:Yeah. Well, we... There there was a lot of funs and drinks to to be had. So... Yeah.
Speaker 2:That was good. So I I landed and I immediately went to a Fogo place, like a Portuguese skewers, sorts, lots of meat place organized by a friend from Siskies.
Speaker 1:Alright. Cool. Thank you, guys. That was the best
Speaker 2:food I had in weeks.
Speaker 1:I drinking the whiskey from the podcast we did with Siskies the other day and it's still interesting.
Speaker 2:Yeah. Interesting is a good word for it.
Speaker 1:It is. It's not for everyone. Yeah. No.
Speaker 2:Yeah. But no, that was really good. And I had, on my left, I had Mark Cashman sitting next to me. On my right, I had... Oh, damn.
Speaker 2:What's his name? The Belgian MVP.
Speaker 1:There's somebody listening to Podnask going. Brian, how can you forget you were sitting next to me?
Speaker 2:Yeah, exactly. Kokan. That was it. That was it. Yes.
Speaker 2:So yeah, that was that was a lot of fun. Lots of good food. Excellent night. And then just like with Ignite, they opened up the park. The the the...
Speaker 2:Not the universal one, but...
Speaker 1:Well, they opened
Speaker 2:up the park.
Speaker 1:Which park did you go to?
Speaker 2:Well, this one, this time, they opened up the universal one. So we went to the Jurassic Park area and then we were on this awesome ride called the Veloci... It's not the Velocity Raptor, but it's the Velocity Coaster, which was amazing. And we went to Harry Potter World and I had I had... I went on a ride that was like the motorcycles from Hagrid.
Speaker 2:Yes. So the the motorcycle with sidespan. And maybe this is not the best podcast to talk about that, but I've got a few friends that might have these little gummy bears that contain
Speaker 1:some Alright.
Speaker 2:Material that might or might not be illegal in some places. So I have one. And, of course, Dalby
Speaker 1:This podcast is completely fiction as Steve and Marijn does tell stories just to entertain purposes only.
Speaker 2:I potentially had one.
Speaker 1:And like AI, it may be hallucinogenic. No. That was the... Oh, yes. Give me
Speaker 2:a bit.
Speaker 1:No. No. No. It may be hallucinating in its facts. Yeah.
Speaker 1:Anyway.
Speaker 2:But... So by the time I I got on that ride, it was getting dark, and I was on the first one. Oh, So I was were on with my hands on the steering things while it was flying around me, hi as a guide, flying as Hagrid on the motorcycle on the first place. It was the best ride of my life. That was very,
Speaker 1:very, very cool. And now you found out the secret of how to get the best out of Universal Studios.
Speaker 2:Oh, man. Yeah. Yeah. But this this was great because this was the park that we've never been to. Yeah.
Speaker 2:This was a new a new experience. But also the conference was as always amazing. Yep. Lots of people. Really good show.
Speaker 2:Yeah. And Cool. Cool. So every time... We've we've been to a few MVP Summits
Speaker 1:We have.
Speaker 2:As well. And everyone always says MVP Summit is great for networking. I have to say on this one, in Orlando, I did a few boot duties on the on the SharePoint boot, and I got into contact with so many people from Microsoft. So this was the best networking event for me with Microsoft Nice.
Speaker 1:In per Cool.
Speaker 2:That was excellent.
Speaker 1:You crammed a lot in in three days.
Speaker 2:Oh, yeah. Absolutely. Absolutely.
Speaker 1:That's why you've not been contacting me. I thought you died. I mean, you know? Yeah. Hey.
Speaker 1:Where's Marijn? And somebody said, yeah. It's Orlando. Oh, yeah. Of course.
Speaker 2:Yeah. Got it. Yeah. And then last week, was at ECS, the European Collapse Summit, which was also a lot of fun. Seeing a lot of people, hearing a bunch of really good talks.
Speaker 2:Yeah. So, yeah, that was good.
Speaker 1:Number one takeaway?
Speaker 2:It's all complicated.
Speaker 1:That was
Speaker 2:good. And nobody really knows where we're heading, but it's gonna be fine. It's gonna I learned be a new word. What's An apocalyptimist.
Speaker 1:You're an apocalyptimist.
Speaker 2:The world's gonna go to
Speaker 1:shit. True.
Speaker 2:But it's gonna be fine. Apocalyptimist.
Speaker 1:Yes. I like that. Yeah.
Speaker 2:I like that as well.
Speaker 1:Yeah. When we've gone through these adonyms whiskies, which we have in front of us, I want you to try and say that again. Yeah. And if you can, I'll let you drive home and give you your car keys back. Excellent.
Speaker 1:Excellent. Apocalyptimist.
Speaker 2:Yeah. Thought I thought that was a that was a cool word. Like, the world's gonna go to shit, but I'm fine with it. Okay.
Speaker 1:No. Nice. Nice. Well, Anok and I did 10 steps to governance. Mhmm.
Speaker 1:Which created a huge amount of interest. Mhmm. Because people were all saying, we have no idea where we're going with this. Again, community thing. So you were instantly into chats with people.
Speaker 2:This was governance for AI?
Speaker 1:Yeah. Governance for... Well, of course, it's for both, in fact, but it's governance for Power Platform. Yeah. And with all the...
Speaker 1:We actually... In Brussels, we had a superb guy. And now... God. Now I'm gonna I'm gonna say Ken somebody or whatever, but it doesn't matter.
Speaker 1:And if I'm wrong, forgive me. Mhmm. But he did a brilliant demo on... You're always so
Speaker 2:well prepared on these...
Speaker 1:I know they are. Yes. So we are. But he was talking about how the environments are set up for AI and he went through all of the specific AI bits for the environments that got powered by. It was really very good.
Speaker 2:That's cool.
Speaker 1:If only we were advanced enough to be recording these events. Yeah, yeah. But that was good. So yeah, being able to work out what in that environment and how it was gonna be attached to AI and LMs and stuff. That was pretty cool.
Speaker 1:So so we did that. But yeah, we ended up with a number of people that really are stuck on this, trying to work out where they're at.
Speaker 2:Yeah.
Speaker 1:So it's a good one.
Speaker 2:Because it's so all encompassing. It's over everything. So everything needs to be in order.
Speaker 1:You know something today that struck me, we're gonna get onto this a little bit because I wanna talk about use cases for AI. But as I was having this conversation with some really smart salespeople today around data and MCPs and stuff, it reminded me of electric cars.
Speaker 2:Okay.
Speaker 1:The world was going to change. Yeah. The engine, diesel, and petrol was gonna disappear, and we would all be driving electric cars.
Speaker 2:Are you talking about ten years ago, or are talking about a hundred years ago? Because the first cars were electric.
Speaker 1:And I've seen that. Yes. If you go to the car center in the Hain Mhmm. In Dehan Yeah. Have you been
Speaker 2:there yet? I have not.
Speaker 1:You should put it on your list. It's a, it's classy place. Yeah. But they also have those first electric cars from the late eighteen hundreds. So, no, I'm talking about five years ago.
Speaker 1:Yeah. I mean, you know, mean to Formula One, of course, and we were talking about the new engines and about the fact that the manufacturers really wanted to get into battery, but now they're all doubting it because they think we'll go to biofuels and Yeah.
Speaker 2:Know. That will solve problem. Lotus, the Lotus cars, they went all in on electric. And now this week, they said, you know what? We're gonna bring out...
Speaker 2:In two years, we're gonna bring out a new V eight.
Speaker 1:Yes. Everybody's talking V Formula one looks like it's going back to V eight as well. But that belief that the world had Mhmm. About electric, and the world did have it, and there was a a few apocalypse to pin... To Pinterest.
Speaker 1:Alright? Yes. But now everybody really isn't gonna go electric. Yeah. And there's a real...
Speaker 1:And I don't wanna go there, but I feel there's a parallel in terms of AI that unless we find the real value, we may be in the same game that AI just becomes a little bit of a Viva thing. Yes. But, I... You
Speaker 2:in in my session, I did in ECS, it was user adoption for AI. Yeah. We'll talk about that later. I had a quote saying, if you are not gaining eleven minutes per day for the next eleven weeks, you're not working with AI, you're playing with it.
Speaker 1:I like eleven minutes. Yeah. And like playing with
Speaker 2:it for eleven minutes. Yeah. Okay. Yeah.
Speaker 1:Accelerated interest.
Speaker 2:Yeah. Thought that was a really good quote.
Speaker 1:I was working on talking about stats. I mean, I was talking about, did something... When was it? Oh, it was probably about five or six months ago. But I worked out that if you take standard pricing for how much it pays for somebody to be at a desk, the cost of the desk, the cost of employing them, the wages and everything else, the average figures in Europe, you need eleven hours a year to pay for your copilot lessons.
Speaker 1:So if you can save just eleven hours a year, then you're good to go.
Speaker 2:Which is two days.
Speaker 1:It's two days. Yeah. But of course, what tends to happen is if I only do an hour a month Mhmm. Then I'm good. But you tend to lose interest after two or three months and go back to the old habits unless you get that adoption stuff sort of like Yes.
Speaker 1:I like that. Eleven minutes a day, you're just playing with it.
Speaker 2:Yeah. So that was... Even there was someone on LinkedIn saying, Oh, went to ECS, but the best thing was this quote from Marijn. And I was like,
Speaker 1:Oh, yay, yay, yay. Eleven minutes. So how did you work that out? It was just something that I came to didn't. You stole it.
Speaker 2:Well, my AI stole it. But let's talk about that because you want to talk about use cases.
Speaker 1:Yeah. Today I was just talking about some people around databases and Oracle and stuff. And apparently, the database important people. Piece. Well, if you think about the whole model, yeah, you've got LLMs, MCPs, You've obviously got people using it.
Speaker 1:Mhmm. But all of that is connecting through to data. Yeah. Exactly. Yeah.
Speaker 1:And that's a big question mark at this point in time. Yes. Right? In terms of protection, in terms of everything else. We're already hearing the great stories where AI deleted all the data in the database, you know, and nobody can restore it because it thought you don't need it and it deleted the backups.
Speaker 1:Yeah. Awesome. I'm glad that was not me. Mhmm. Alright.
Speaker 1:So you're kind of hearing those horror stories. But apparently, one of the database companies has put a call out to their sales team to say, go find the use cases. Yeah. Because we're not entirely certain where we're at. I was in a presentation for legitimate business reasons the other day where the team was talking about developing an app that they expected to do in seven months and they did it in five days plus another eleven days worth of testing.
Speaker 2:Yeah. Some vibe coding and...
Speaker 1:Yeah. But nope, wasn't through vibe coding, but it doesn't really matter how they did it, but they they did it. So they managed to do this thing in seven days, which is fine, but that's not a use case. That's just an opportunity. Not all apps will work the same, alright?
Speaker 1:No, it's not
Speaker 2:repeatable. Exactly.
Speaker 1:So, and so it struck me, you know, that's not wrong. And we could narrow it down a little and say, hey, look, just within M365, what is a use case? Or what is the obvious opportunity? Or where can you make this work? Or where can you, you know?
Speaker 1:So I thought that's what we'd talk about.
Speaker 2:Yeah, because... Are you ready for the second zinger?
Speaker 1:The second... Well, what was the first thing? The eleventh?
Speaker 2:The yeah, yeah. The second one is a lot of people are using Copilot as a expensive spell checker and an auto complete spell checker. They write an email and they say make this email look better or something like that or just this is what I need in a Word document. Write this Word document for me. And basically, that's not really where the big value is.
Speaker 1:No. And that is part of the problem around those last few days where where is that value. Yeah. But I do quite like the spell checking feature. Well, I like the translation I've feature texting
Speaker 2:with you for quite some time, for quite some years. And yes, you could do with a good spell checker, yes. Also, have you heard about this thing called Whisper Flow?
Speaker 1:No. Yeah. Maybe. But I can't do an
Speaker 2:app that you put on your phone and then you can just talk to the app and it will just write down everything you say. So instead of typing everything in, you can just talk to it and it will just write down everything you say.
Speaker 1:Okay. Using AI?
Speaker 2:Don't know. It's just the
Speaker 1:app's good.
Speaker 2:Yeah. Apparently, the app's really good. I I haven't used it yet, but instead
Speaker 1:of GPT me then when I
Speaker 2:talk In to in basically, it's like me texting people all the time. I could just open up my WhatsApp or whatever and start talking and it will just write it down. Okay.
Speaker 1:No, I haven't.
Speaker 2:No, okay. No, sure.
Speaker 1:No, but I was gonna say something really exciting until you said whisky flow. Oh, have you tried the new one?
Speaker 2:Not Whiskey
Speaker 1:Flow. Well, I was just undoing our first little taster for the night. Whiskey Flows. Whiskey Flows.
Speaker 2:Yes. Absolutely. It does. Good name for a podcast.
Speaker 1:It is a great name for a podcast. I wish I'd got pen and paper here to remember that I wanna talk about workflow, MS Teams and workflows because Let's I've do been playing with that later, but let's talk about some of the whiskey first. I assume we're gonna do a little taste throughout
Speaker 2:the year.
Speaker 1:It's gonna Just be this rinse out my glasses. So I went to the Norfolk show, as we've talked about it. It's to Norwich. I had a great time. But I also wanted to take a few memories in as well.
Speaker 1:So I went to a really tacky seaside place. And so I haven't done that for years, but got into real childhood stuff of playing in the arcades and winning all the tickets. And and after about four days, I got a Mars bar. Actually, after four days, I couldn't find anything that I would still want to get with my winning tickets, so I made some kids really happy because I said, here you go. There's 10,000 bloody tickets or whatever.
Speaker 1:So they did their thing. But I went to a place called Southwold. Southwold is on the East Coast. If you ever got time in The UK and you wanna go somewhere quaint, old, and and just be in this thing, Southwold is great. I went there many years ago, and they have a just a brewery there
Speaker 2:Yeah.
Speaker 1:Called Adanham's in Southwold. Mhmm. And, you know, you get to a point where even thirty years ago, even less, they were still delivering beer on the back of a dray with two cart horses pulling it around the village, dropping the beer off at all the different pubs. I kind of...
Speaker 2:True electric cars. Yes. Let's bring back horses.
Speaker 1:This is... Yes. Powered by malt and barley. Yes.
Speaker 2:Two horsepower.
Speaker 1:Yeah. And and in fact, I think I might have mentioned it before, but this also has award winning bacon because all of the Okay. All of the used products from the beer goes out to the farms to the pigs, and the pigs started winning all these gold medal awards. So it's a kind of They
Speaker 2:must be doing something good there. Correct.
Speaker 1:Okay. So I went to this... I went to the hotel that the brewery owns, which is in front of the brewery, started looking at what beers I wanted to just revive, you know. And then I saw four bottles of whiskey on the shelf called Adenums, and I went, oh, crap. My plan was to buy one bottle of whiskey on this trip.
Speaker 1:That was the plan. I didn't, you know, I didn't wanna go nuts.
Speaker 2:Plans change.
Speaker 1:Plans did change at that instant because I knew that they were gonna be great, and I knew that I would not be able to turn them. So tonight, we have English whiskies.
Speaker 2:Alright? Celebrate, Adam's. Yes.
Speaker 1:And they're they're all really quite cool, and they're all interestingly done. The guy that actually set it up passed away. He had a heart attack, unfortunately. Mhmm. So when he set the distillery up and and blended these together, it's not there anymore.
Speaker 1:But, yes, that's why we're doing English whiskies today. One from a distillery we've tasted before, but two from Adanum. So I'm gonna start off with... And
Speaker 2:I have to say these bottles look amazing. They're they're pretty standard ish bottles itself.
Speaker 1:Shape and size
Speaker 2:Yeah. But they put like a colored cover over it. So, we got purple. We got red. We are now doing the the blue one.
Speaker 2:Yep. Nice. And the blue one is the single malt. This is the single malt whiskey.
Speaker 1:Yeah. Handcrafted in England. Actually, do want... Can we have a rude joke? I've just sat watching one of the...
Speaker 1:I was doing Instagram while I was kinda waiting for you to turn around, and I forgot the damn comedian's name, but he has a lot of comments coming from the audience, and then he jokes about them. It'll come back to
Speaker 2:me in life?
Speaker 1:No. Anyway, somebody said, so what was your first job? He went hand. So, then somebody from the audience said, And did you make him cum?
Speaker 2:Okay. Yeah.
Speaker 1:Yeah. Alright. Maybe we'll cut that down.
Speaker 2:No, we don't. His name
Speaker 1:will come back to you in a second. Alright. So, yeah. So, this is... It's also matured in French oak.
Speaker 2:Oh, I
Speaker 1:like French oak.
Speaker 2:Nice floral. Okay. Good. See? That does not disappoint on the nose.
Speaker 1:It's not. I think this might have been the first one I tasted. And as soon as I I stuck it under my nose, it was gonna be there.
Speaker 2:Yeah. It's it's light. It's floral as I already imagined it would be. There's a little undertone of
Speaker 1:apricots. That earthy skin of a Maybe. An apricot.
Speaker 2:Maybe. Maybe. Well, it's definitely something earthy that's that's in there. A little bit of a little sour note that I get a lot with French oaks. Yeah.
Speaker 2:Mhmm. Okay. Okay. Okay. Good.
Speaker 2:Interesting.
Speaker 1:He's diving in. Smooth. It is very. That was my first taste.
Speaker 2:That is super smooth. No kicking the back. Mhmm. Does that have an age statement?
Speaker 1:I was just working.
Speaker 2:Not really. No. Oh, that is super smooth. That's almost like Irish.
Speaker 1:I know. I know it is. No age statement.
Speaker 2:Okay. That's fine.
Speaker 1:The interest... Interesting thing which I've just remembered because I've seen the note on the back of the bottle about the brewery is that during the Second World War, they got hit by some incendiary devices, which took out all their yeast. And so they had to... It took out... Sorry.
Speaker 1:No. It took out the London yeast producers producers for the breweries, so they had to dig out their yeast and they created their own strain. So this whisky, which obviously uses the beer mash as part of the distill process, is actually created from 70 year old yeast strain that they created since the second war stories. I know.
Speaker 2:Oh my god.
Speaker 1:I know. Wait till we get to the story of the last one. The last one
Speaker 2:is pretty pretty
Speaker 1:Oh yeah, this is nice.
Speaker 2:This is like how do you say that? In in a... I mean this in a good way. It's not complex. It's a drinker's whiskey.
Speaker 1:It's a drinker's whiskey.
Speaker 2:Yes. It's not not like a campfire thinking about Nope. The apocalyptic. It's it's no fuss. It's...
Speaker 2:I I... And once more, I mean this in a very good way. It's plain and simple. It is what you see is what you get. You smell this.
Speaker 2:You get this. It's just this... It's good.
Speaker 1:I agree entirely. But it's still gotta finish. I mean, even now while you were talking, twenty seconds later, it's still there. The orange is just starting to come through. Not strong by a long, long way, but it's...
Speaker 2:Yeah. Yeah. But I I have to say I do love French oak casks, but they... Because they bring that that... I don't know, like springtime in the air kind
Speaker 1:of Springtime in Paris. Mhmm. Yeah. No. That was still
Speaker 2:Yeah. Bad at That's that's very, very nice. Yeah. It's a very drinkable, like an aperitif whiskey, 04:00 in the afternoon kind of thing. Yep.
Speaker 1:Yeah. Excellent. And it's not super expensive either. Mhmm. I think I paid about $5.50 or €60 for the bottles or something.
Speaker 1:So, yeah, if you want something in there with some honey, vanilla in there, and a nice light, bright, then go for this one.
Speaker 2:It's it's very light. Yeah. Mhmm. In in taste wise. Yeah.
Speaker 1:Yep. Mhmm.
Speaker 2:It's fresh.
Speaker 1:I quite like it. Fresh. Mhmm. The first blue one is fresh. So that is cool.
Speaker 2:Small batch copper pot still.
Speaker 1:Yep. Good. If it feels a bit echoey today, we're not in the studio today. We're actually up in my whiskey collection. Which is lighter and brighter in here.
Speaker 2:The whiskey cave.
Speaker 1:The whiskey cave is not quite a cave. No. But that's cool. So, yeah, so we're just sort of sticking it up here today, so that's good. So MS Teams workflow.
Speaker 1:Workflow. Yes. You played with it yet?
Speaker 2:Yes. And? Because the... So, basically, we're talking about the workflows app, which is powerful. So preview
Speaker 1:at the moment?
Speaker 2:Kind of. Let's get there. Yeah. It has been in Teams for a number of years, but they just revamped it a few months ago. And I played with it in Teams as a Teams app.
Speaker 2:But now they've also added it to document libraries. So in the document library, there's now a new button say workflow. Excellent. Next to new and and upload and editing grid view, there's a button workflow.
Speaker 1:I did not
Speaker 2:know And you can create easy, light, small workflows from there.
Speaker 1:I won't argue with that. Mhmm. Definition of easy is always relative.
Speaker 2:Well, it's definitely easier than doing it in Power Automate. It basically says what should be your trigger, what should be your actions.
Speaker 1:If you take the if you take that.
Speaker 2:I agree. You...
Speaker 1:Kind of, but every time I tried it to do next, next finish, it went,
Speaker 2:Okay.
Speaker 1:Okay. But I was in Teams. Yeah. And I was kind of trying to pull out specific messages from specific users. Okay.
Speaker 1:But that's okay. I get that. But that's Power Automate Workflows. But what about the AI Workflows? Exactly.
Speaker 2:So that is the new thing that they put in the Workflows app is if you have a Copilot license, you can also do Copilot actions.
Speaker 1:Yes. And that's what I think the power is going to be in this. I think I think it's actually going to bring it in front of people for them to actually be able to play and do things. And I got a couple of simple things in there like, you know, go through all the meeting transcripts that I've been involved in and give me my top five priority things just by default.
Speaker 2:But, basically, those are things that you could also just do with your Copilot chat, for example. So that's not really something
Speaker 1:you You couldn't put in a Copilot chat. You... It wouldn't look at your emails. Copilot chat is grounded in the Internet.
Speaker 2:No. No. I mean, the the... CoPilot icon. The CoPilot for work.
Speaker 1:In three six five CoPilot. Yeah. Yeah. Okay. Yeah.
Speaker 1:You you could do that. But the thing is that it's now... And and this is my point really about use cases. It's actually in the place it should be. You know, it's my work is in MS Teams.
Speaker 1:Yeah? My communications, my projects, my planner, my
Speaker 2:Yes, those action do
Speaker 1:there is. But it also means you kinda gotta have some level of knowledge and experience, whereas this really does take people in and actually say, hey, what do you wanna use this for? It gives you suggestions, gives you some predefined templates. It's kinda getting people with one foot on the first step, ready to get moving forward.
Speaker 2:Can I give you a use case for that one? Go. So, the use case for example for the document library workflow with AI could be that as a trigger every time I put in a new file. As an action, create a summary of this file.
Speaker 1:And then put it into some metadata somewhere.
Speaker 2:Yeah, exactly. Yeah. Yeah. Yeah. That's it.
Speaker 1:Is that a use case?
Speaker 2:I believe so.
Speaker 1:It's a user story about something that you could do. Is it gonna change the world? Is it gonna save you eleven minutes?
Speaker 2:Potentially, because you go to the library, you see a lot of documents, and you wonder if this is the right document. And instead of opening it up, scrolling through it, you can actually see the... A little summary and see, oh, this might not be the one I'm looking for.
Speaker 1:Similar idea. I would like it to be able to tag my documents with one of my sensitivity labels. Yes. We can know that has to be an e five scenario, but actually with an e three in using workflow, you could describe for your organization what is a confidence document, what is a sensitive document, and then tell it to actually label it.
Speaker 2:And you don't need an e five for that.
Speaker 1:No. That's what I've just said. I'm glad you was listening.
Speaker 2:Already so much money that you're saving. Yeah. Interesting.
Speaker 1:Yeah. Yeah. And then you can customize then what the label should be and what type it should be created for.
Speaker 2:Oh, I love that. I love that.
Speaker 1:Depending on the I think
Speaker 2:that is a really, really good use case from a governance perspective. Yep. Compliance perspective, I love that.
Speaker 1:No, I like that too. All I gotta do is write the
Speaker 2:prompt. Yes.
Speaker 1:And test it. Yes. It'll take me ten minutes to do the prompt and forty five years to test it to make sure everything is created with the right label on it.
Speaker 2:Yes. But that's fine Yes, yes, exactly, yeah. Yeah, so
Speaker 1:I was thinking along those lines where you're trying to make it to do simple things. But again, we really need to do some work on these eleven minutes. I really like that idea.
Speaker 2:Yeah, yeah.
Speaker 1:So when we finish this, dig down on that a bit because I think that's fun. So again, use cases, and they're not bad use cases, but they're not gonna change the world. This is the electric car. True. True.
Speaker 1:If I continue working the electric car this way, it's not gonna change the world to stop people using, you know, fuels and diesel and everything else.
Speaker 2:And basically what you need to look at is, is the thing that I'm doing, is this just an automation? Or am I really changing the way I work and changing a business process?
Speaker 1:No, I think that's key to this. Mean, I think automation is an interesting word because for me, automation is probably only 20% AI and 80% workflow.
Speaker 2:Yeah. No. Absolutely. That's what I mean. Mean, is this an automation like I can create a workflow or is this really AI thinking, coming up with solutions, creating something and doing something extra.
Speaker 1:No, I agree with that. And I don't know where that fits the moment. I have some ideas and perhaps we'll cover that after the next whiskey because I wanna dig down a little bit into processes. And something I did is I did the keynote for the AI CD stuff. Yep.
Speaker 1:Think the title was AI is a Problem Opportunity. And so that was the title. I love
Speaker 2:that. It's a problem opportunity.
Speaker 1:Yeah. Yeah. Is nice. It was really good. I enjoyed doing that post.
Speaker 1:You know the coolest thing? And then we'll get onto maybe the second whiskey. So here I have a picture of a graphic artist by the name of Dan Kitchener, and I really am becoming to love his work. And in fact, when we do Commsverse, I'm hoping to hit his gallery and I think it might cost me for one of his pieces of work. I've seen some of the other stuff.
Speaker 1:Okay. So I took a photograph of that piece because I just love it and I said, okay, I want to create a set of slides for the keynote that is taking this color. So I told it which colors to take and also the essence of it, you know, the Asian ness. I mean, you guys can't see this. I understand.
Speaker 1:Yeah. Maybe Marijn will take a photograph and we'll include it somehow into the image for the for the podcast. Yeah. So I basically said, okay. I I create me an image with this title on.
Speaker 1:So, hey, AI is a problem. Give it an Asian background using these colors, alright, where it's actually got people in it because, you know, it's a people thing. So it created this amazing slide. Alright? So then I decided to test it because it had Chinese characters all over it, alright?
Speaker 1:Now, the prompt I gave you is basically all I said it to do. So then I basically said, hey, I went to a different chat, I uploaded the image and said, you translate the English, the Chinese characters, please? Collaboration, sharing, all of the characters were all about m three sixty five, all about... Yeah. Was...
Speaker 1:I was blown away. It's... Mean, I hadn't asked it to. It just... And and that identified for me where AI is really gonna...
Speaker 1:Is... He made the choice. Yep. She made the choice. It's Holly, of course.
Speaker 1:She made the choice on what to do with those slides, but I never told her to put... Make sure the Chinese character... The only reason I checked is just to make sure that one of them didn't say fuck off this guy's talking shit. Yeah. You know?
Speaker 1:Yeah.
Speaker 2:Yeah. Yeah.
Speaker 1:Of course. Of course. But actually, he said, no. No. All of these are based around the following subjects.
Speaker 1:Collaboration, sharing, blah blah blah.
Speaker 2:And this was Holly being chatty pity. Yes. Okay. Okay. Cool.
Speaker 1:Yeah. Basically creating the slides.
Speaker 2:I've got a great story as well.
Speaker 1:Go on then. Before I do the whiskey... I tell you what, you start telling the story and I'll would pull you wanna go red or purple?
Speaker 2:Let's go purple first. Purple. Yes. So friends and family and people that are listening to the podcast might have heard me say once or twice that every time I try to do something with AI, it's just The result is just not what I want.
Speaker 1:It's a bit like opening the Christmas present from your mother that...
Speaker 2:And socks....
Speaker 1:The seventeenth year in socks.
Speaker 2:Yeah. Exactly. Yeah. So, I've been hearing a lot of... Oh yeah, let me rinse that.
Speaker 2:Okay, okay. So, I heard a lot of really good things about Claude and Anthropic and it being very good at doing a lots of things. So now we have in the frontier model of Copilot, we can enable Anthropic as the subprocessor. So I did that. Thanks to Peter who helped me with a few bits and bobs to do that.
Speaker 2:And actually, so my session in Orlando was the one for Marie Kondo. So the Japanese cleanup lady that will clean up your your business data. And I've done that a few times before with numerous people and they all... So I had a slide deck. So I had it...
Speaker 2:I had this done. And I actually said to co work, you know what? Make me a slide deck about Marijn Kondo cleaning up your business data. And it made the most awesome presentation I have ever witnessed. It actually...
Speaker 2:It it created a bunch of slides. I said it it must be fifteen minutes, so it made like 25 slides. It was amazing. It had all the good things in there. It had a number of things I haven't thought about.
Speaker 2:And it actually said... So you you can see through the reasoning process and it said, well, it's about Marijn Kondo. She's Japanese. So I'm gonna use like a minimalistic, gray sandy, background, things like that. It just created amazing slides.
Speaker 2:I did the same thing for ECS for last week, which was about user... Atcar using Atcar for AI user adoption. And this time I said make this look like just awesome space themed whatever. And once more the slides were just amazing. It was so good.
Speaker 2:They looked so beautiful. I was so amazed. And when I was giving the presentation, I saw so many people taking pictures of the things. Now I have to say, it came up with 25 slides. I removed a few.
Speaker 2:I added a bunch that I wanted The ones with
Speaker 1:the intelligence you wanna
Speaker 2:talk about.
Speaker 1:The that
Speaker 2:I want to talk about. So it was not it was not 100% AI, but let's say it was 70% of the content was what AI came up with. I saw so many people taking pictures. I was so proud. It was so good.
Speaker 2:It looks so nice. But on the Marijn Somers one. One of the... So basically the slides, they all kind of look the same. They all have like a title and like a subtitle and a few cues that are always the same.
Speaker 2:Also, it's like, a starting slide, an agenda slide, and then it starts with some numbers. And for the Marijn Somers, there were three sets of numbers with... About emails being in an inbox and things like that. Like default, there's 200 emails that are sitting in inbox. And a document that is shared is never read after five days or something like that.
Speaker 2:And I actually said, those are cool numbers. So where did you find those numbers? And it said, well the first one I got that from a McKinsey, study from 2004. The second one I fabricated and the third one I also fabricated. So it's illustrative.
Speaker 2:You can use it, spin it around, have a good story about it, but they're not real numbers. So... Okay. Interesting.
Speaker 1:You're gonna make me a liar.
Speaker 2:That was interesting. But I mean the numbers, they look pretty cool and I actually used them on session as well. And everyone was like nodding like, Yeah, yeah, can see that. Yeah, 200 emails in inbox. That sounds about right.
Speaker 2:Yeah, documents never been read after five days. Yeah, sounds about right. Yeah. And then I said, well, this I had to show them the screenshot like I asked it like,
Speaker 1:where did it come from?
Speaker 2:I just made that up. No. Interesting.
Speaker 1:Okay. But if say to somebody, how many emails do you think people have in their inbox? They will go a number. Yeah. Because they go what they've seen, where they're at.
Speaker 1:And the really frightening thing about this, in a sense of a word, is it is like talking to somebody that knows their shit in this space. Not an expert. Mhmm. You know, your average Joe that understands that I don't I don't have all my emails cleared up. Yeah.
Speaker 1:You know? Mhmm. And, yes, 200 feels about right. Yeah. Unless you look at mine, which is now at eleven hundred, I think.
Speaker 1:Yeah. But that's, you
Speaker 2:You've been busy and traveling and doing all things. Yeah.
Speaker 1:No. I've already cleared a bunch out.
Speaker 2:Okay. Okay. Okay.
Speaker 1:If you know a good presentation, I'm clearing stuff out. Yeah. But... I know it's interesting. I mean, I think if we think where AI was eighteen months ago, it's tremendous.
Speaker 2:And that also kind of brings me back to that eleven minutes thing. So a few weeks ago, I was a guest on Rebecca Isaacson's podcast. Yep. And she gave me some questions beforehand and I just fed them into... Basically, fed them into chat GPT and I fed them into grok.
Speaker 2:And I got some nice polished answers back and I said be bolder. Be contrarian. Be difficult.
Speaker 1:Go crazy. Yeah.
Speaker 2:And after like three iterations, it came back with amazing answers. And I used a bunch of those answers on the podcast. So happy the podcast turned out super well. Really happy. But yeah, for...
Speaker 2:So business case, is it is it about value? Well, in this case, it is about value because I'm on the stage or I'm on the podcast. I'm delivering. I'm selling myself. I'm delivering value to people and I had Copilot just make my presentation go from to effing amazing.
Speaker 1:Not sure. Why do you not just get AI to write the podcast for you and just say here's an m p four file, it's all done.
Speaker 2:Because me and Rebecca were... I kinda wanna say good friends. I hope she agrees.
Speaker 1:Rebecca, if you want the truth, just let me know. We'll
Speaker 2:we'll get
Speaker 1:together. That's fine.
Speaker 2:Yeah. But we both have this same kind of weird dark sense of humor. So...
Speaker 1:And that was the magic. The human in the
Speaker 2:loop. Exactly. So on the podcast it was magic and it was just like us talking like we are talking when we're waiting on a flight on the airport. Yeah. And it was just flowing, but there was also some zingers and some good value in there.
Speaker 1:Nice. Well, let's talk about a whiskey that is from barley, wheat and oats.
Speaker 2:Mhmm.
Speaker 1:Finished in an American oak.
Speaker 2:All the things that we need in a food triangle. Oats and Absolutely. Yep. Mhmm.
Speaker 1:Spirited notes of pepper, orange people, and toasted oak. Orange people? Orange peel. Sorry. Yeah.
Speaker 1:I heard my say it, decided not to correct myself. But as much as this is drinkable, I think it will have a few things missing, but it is a different kind of whisky. This is from the same Distilled.
Speaker 2:Yeah. Okay.
Speaker 1:Alright. But you're gonna see two completely different places.
Speaker 2:Yeah. Because the first one was a single malt. This is a triple malt.
Speaker 1:Okay. Yes.
Speaker 2:So it's barley, oats and wheat. Correct. Yeah. Now what you find on this,
Speaker 1:no corn, no, it's not American. What you find is it's got a great nose.
Speaker 2:It's got a great nose. It's more grainy. Yes. It doesn't have that lightness and that soury earthiness from the French oak. No.
Speaker 2:So what kind of oak was... American. American oak.
Speaker 1:It's on the front. Sorry.
Speaker 2:Oh, sorry.
Speaker 1:Mhmm. But it's got that kind of roughness about it, isn't it? Yes. It's beautiful. Yes.
Speaker 1:And I've got a... I really hope that there's a contrast in this. But they're right about the orange on the nose. I love whiskies that sit there and give you that toasted skin of an orange.
Speaker 2:Yes. Me too. Yeah. Now I just had a sip. Also super drinkable.
Speaker 2:A little more kick in the back. Yes.
Speaker 1:More rye as well, but it's... There's no rye in here, but it's No. But like floral is there.
Speaker 2:Super, super smooth. I kinda miss that oiliness that we get in some other whiskeys that just
Speaker 1:That's a little bit like that in the first one. It's not as oily.
Speaker 2:Not that much.
Speaker 1:No. Not a big one.
Speaker 2:I I kinda like that in a whiskey, but this is once more, it's a great aperitif whiskey, would say. Super smooth.
Speaker 1:I think so. I mean, they're different than most things that are in this cabinet here, there's a fair choice.
Speaker 2:Yes. They are... They're definitely different than the rest.
Speaker 1:Yeah. Mhmm. Yep. I think might just leave a few silence here while we finish this one off.
Speaker 2:I I... Yeah. I I like this. Very drinkable. Yeah.
Speaker 2:And I think this is... So if I have to compare the French oak single malt with this American oak tree, triple malt, I think this triple malt is more for the masses. It's more the... For people that might drink a whiskey once every five years, this is more like how they would see what a whiskey would be.
Speaker 1:I take a different approach because they're both in the same house, so it's me. There's about four hours difference. This one we're drinking now, I might drop into while I'm preparing dinner sort of 07:00 at night and everything else. The one we did afterwards is at the end of the day, I've eaten, maybe I've watched a movie, I've had a chat with somebody, and I just... Whereas I might go for a heavy Peter's stuff to finish the day off, I would heat drink the first one, the French oak one.
Speaker 1:Oh, I would start with that.
Speaker 2:I would start it down at 04:00.
Speaker 1:Traditionally, would. I agree. Yeah. But when you finish work, you want something that's gonna go boom, you're in the house. And this has got
Speaker 2:that bite. Like a welcome home, daddy.
Speaker 1:Yes. Exactly. Whereas at the end of the night, you wanna go smooth and ease off like a brandy, then that's where that that first one tastes. But don't get me wrong. I get it.
Speaker 1:See. Yeah. The difference is four hours.
Speaker 2:Yeah. I like that. I like that. Okay.
Speaker 1:That is cool. Yeah.
Speaker 2:No. We didn't use AI to come up with that.
Speaker 1:Oh, shit. I'll delete it. No. We didn't use AI to go. Absolutely didn't.
Speaker 1:Be interested in whatever's we did if we ever fed our podcast scene to find out what would come out. We'll do that one day, I suppose. So We'll do that one day. Business case. Wanna talk about where where you would put AI in as a natural place.
Speaker 1:So I've told you that I've been doing some work with one of the analysts inside the business around sales teams. Yeah. And we we agreed, I think, on a previous podcast, sales guys, they're the default people to have AI. And I was breaking this down just recently and trying to work out, and I was breaking it down as, again, part of the keynote. So, I said that this is a problem opportunity.
Speaker 1:Yeah. Alright. We know that AI isn't really a problem, but it is a problem to get the real opportunity out of it. And the difference is you have to turn the freaking thing inside out, which is a lot heavier in terms of change. So what do I mean by that?
Speaker 1:So if I take the sales process, yeah? Yeah. You know, you've gotta prepare for a meeting with a new customer. You need to understand what you're gonna propose. You need to defend your case with the customer and convince them it's a good idea.
Speaker 1:You kind of The
Speaker 2:whole sell me dispense Yeah,
Speaker 1:that kind of stuff. You wanna finish off the meeting with some notes. You then wanna close it all down with emails and stuff. And of course, you can use AI to do all of those, and there's a sixth one actually which I'll cover in a minute or two. So you got all those things.
Speaker 1:You can use AI tools to do that. You can take You can start using Copilot or ChatGPT on the internet and you can say, Hey, tell me, what is the company doing? Have a look at their press releases. Do they have some new products? Find out whether my product will actually help them with their products.
Speaker 1:So, I can do that kind of research. I can look at all of the communications I've got, find out whether we've responded to everything they've asked for, find out whether there is any anger in any of those emails. You can kind of do all this all again by email. You can create the proposal that you need to do, alright? And one of the cool things that you can do, you can actually say, Hey, look, I'm gonna present this to this company, alright?
Speaker 1:Challenge me. What questions are they gonna ask me? What do I need to prepare for? You can actually have that conversation and role play which is really, really cool. And of course, you can then sit in the car after the meeting and then tell Copilot or your AI, hey, this is what we did and we talked about this and this is what they brought out and that's good and add
Speaker 2:to Well, have touch AI listening to the meeting.
Speaker 1:Well, potentially you could do if you've one of those hidden devices, which most salespeople wanna work face to face and they don't wanna sit there and go, Hey, I'm gonna record this meeting. They wanna be friendly and sociable. But doesn't matter. They can do it all in terms of getting the output. And then, of course, they can create their emails and include all the content, and away you go.
Speaker 1:So, that is a process where sales process exists and then I'm interjecting AI as a tool where I want it to do something for me.
Speaker 2:Yeah. And it's not just automation. Coming... It's gathering information from different sources potentially and throwing that into one block that we can read, that we can understand, that we can do
Speaker 1:something But it's still seven or eight tasks. Yeah. Alright? Yeah. And there's no consistency in there because if you ask the wrong question at the wrong time or if you get let down a different way, you're kind of losing that consistency.
Speaker 2:It's about reasoning. Yeah. It's about we need to let AI reason over a question
Speaker 1:that we That's need to what you have to do and that's the big change. And it's the one that will take time and effort that you need to get to in eighteen months time where actually the process is owned by AI. Yes. Right, so that's what we need to do. So, we need to have AI and say, hey, this is the process and how we do that, I don't know at this point, but it's gonna be a case of orchestrating a bunch of agents and stuff like that.
Speaker 1:But the human input is now advising. So they're saying, hey, this customer, I think the potential is for this, keep an eye on it. So you're defining what you want that customer to get. Yeah. Alright, then you walk away.
Speaker 1:Yeah. The next one is, you know, communications. They're tracking communications coming in. The AI is ignoring them most of the time until maybe it says, hey. There's an email arrived that's talking about that stuff you sold and how That's much my your email.
Speaker 1:Or Hey, I've just read in the newspaper that they're selling and building and doing, which means that potentially they want more business, there's an opportunity. And so, AI is in the core of the process. It takes over the pipeline and you're just advising it and rounding it and confirming as the human being that's adding that intuition and knowledge that AI can't. There's the human in the loop.
Speaker 2:The loop
Speaker 1:is driven
Speaker 2:by AI.
Speaker 1:The the human is not in the loop. AI is in the loop. So I termed this as human to AI is the first normal process, but AI to human is the one that you're really getting your value. Yeah. And the big reports, Gartner, McKenzie's, and everything else are all saying that the only way you're gonna get real, real return on investment is by allowing it to work out, hey, you need to make a decision, and I'm a bit stuck here, but I think you need to send this email because I've just found this and this, And you're just waiting for it to happen.
Speaker 1:And the value, of course, is that instead of a human having to manage 25 accounts and track them each month and make sure what they're doing, they can do 75 accounts because it's all being watched by AI. Yeah. And that's what I was saying, but it's a problem opportunity. It is. You identify the problem and then try and work out that opportunity and allow AI to do that opportunity.
Speaker 1:Get
Speaker 2:the human out of the loop. Trust AI to do the thing, verify the results every once in a while. Yes. I actually heard... So, you know, you know I'm a big podcast listener.
Speaker 1:Yeah. Of course.
Speaker 2:And mainly when I'm in the gym, I have my so when I'm driving to Brussels, for example, I've got a certain kind of podcast. When I'm driving to The Netherlands, I've got another kind of podcast. When I'm in the gym, I'm listening to...
Speaker 1:Is it all to do with location and direction of travel?
Speaker 2:No, it's more about state of mind I wanna be in. So when I'm in the gym, it's mostly business podcasts and things like that. And I heard a few really good stories on those podcasts. One that I heard yesterday was about a guy that... So you know Open Claw?
Speaker 2:Yes. So he connected everything he got with Open Claw. So he gave it medical, things and and and all these, like his, wristwatch that was giving him health data and all these things. And actually, the system said, you know what? You are systematically, dehydrated.
Speaker 2:You need to drink more water. So what the thing now does is because it's got access to everything in the house, his cameras in the house, his screens, his WhatsApp. Every now again he's he gets a message like I've seen you've been sitting there for four hours.
Speaker 1:Move your ass baby. Why
Speaker 2:don't you go grab a glass of water? And he went to the kitchen, grabbed the glass of water, drank it. He came back and the system said, well done. I'm proud of you. And I mean these these are things that we need to do.
Speaker 2:We need to... Because AI can do things so much faster, better, easier, smarter than us. Because we humans we make mistakes. That's why we have
Speaker 1:So is AI.
Speaker 2:I know. But less. That's why we also try to do self driving cars.
Speaker 1:Because... Love the way you said we try to do self driving Yes.
Speaker 2:Well, in in Belgium and I was in Belgium today, the word was that we need to speed it up because The Netherlands just made it legal to use self driving in on the highway. And well, self driving is doing pretty well. I... When I was in Orlando, I saw one of those Did you go in one? I didn't go in one, but I saw one and it was a Jaguar
Speaker 1:F Face. They are. Yeah. Yeah. They're all Jaguars.
Speaker 2:And with with all the cameras. And I mean, self driving, the idea of self driving is that if a computer does it, it just will make less mistakes than a human.
Speaker 1:Anna went to one when they were in San Francisco and she said, it's the weirdest thing. You wanna grab the wheel. Yeah.
Speaker 2:I can definitely imagine. Yes.
Speaker 1:Do you know the story about the way it goes when they first got created?
Speaker 2:Tell
Speaker 1:me. So, they set the basic programming, all the safety stuff and everything else that goes with it. I've just took the lid off this. Look at the colour of this cork.
Speaker 2:Oh, nice. Interesting.
Speaker 1:I'll come back to that whisky And in a they had a phone call from the police within a week of them going out there and they go, So what's happening? He said, Well actually we do have a problem. About 12:30 at night, they all congregate in the same place. What? Yeah.
Speaker 1:So they put people there and they all go and queue up at this particular location in the middle of San Francisco. And they had to literally program it in there for them not to do it, but they could never work out why.
Speaker 2:Awesome. I love I love those kinds of stories. Those are brilliant. But also, like, this this guy that I was talking about with the open cloth thing, it also got the control over his phone and his GPS. So he was in his self driving Tesla and the system said, oh, I see that you're running low on this kind of thing that you need to eat because you're trying to be healthy.
Speaker 2:So he redirected the GPS of his self driving Tesla to go to Whole Foods to or to go buy some whatever, wheats or milk or whatever. So he saw in the car that it was changing directions and the car was just going to somewhere else than he was envisioned.
Speaker 1:Imagine the... Either this guy is completely nuts. Oh, He's mentally illusioned. But
Speaker 2:the the idea is how far can I take this?
Speaker 1:Exactly. Trusting it. I agree. Not seeing Well, not really trusting.
Speaker 2:It's it's trying to verify. But it's it's it's an experiment. Like, what can it actually do? And how far can I take this? And You those
Speaker 1:talk about these podcasts and I know you go, yeah, okay, fine. But send me that one.
Speaker 2:Yeah, I will.
Speaker 1:That sounds
Speaker 2:like fun. Yeah. And I
Speaker 1:might try and find some time over the next few days because I'm away in Budapest over the two or three days. Oh, man. Because we've still got two more to drink and we're already fifty something minutes in, so we're having
Speaker 2:some Oh, We need to start drinking faster.
Speaker 1:Yeah. Well, to be fair to everybody, we're literally only taking really small tasters of this. Allegedly. Now this is gonna be an interesting mix. This is a rye malted whiskey in the French casks.
Speaker 1:We had barley in the French cask with floral stuff, and now we've got rye in the French casks. In? I'm just going to see whether it says specifically. Yeah. So it's it's rye from from The UK, from near the
Speaker 2:distillery. Okay. So it's the red bottle from Ardenham's.
Speaker 1:From Adenham's.
Speaker 2:Adenham's cyst. Sorry. Sorry.
Speaker 1:I don't apologize. It's all good. I'm gonna... I haven't stuck it under my nose yet, but...
Speaker 2:It's rye.
Speaker 1:It's rye.
Speaker 2:It's it's definitely rye. Yeah.
Speaker 1:It's nice color.
Speaker 2:It's a beautiful color.
Speaker 1:Very deep orange.
Speaker 2:Oak. No, not oak. Oaky orange. Golden... Orange.
Speaker 1:Yeah. Even red. Maybe. Maybe we're just color blind after the other two. Yeah.
Speaker 1:It's rye. It's never a strong rye, is it? It's actually quite a smooth nose.
Speaker 2:It is a smooth nose. And I am pretty sure this will be a smooth drink as well. Like, the other two were super smooth, but I'm I'm not getting any of those French oak notes on the nose.
Speaker 1:No. But the rice probably
Speaker 2:probably Rice probably all that. Blending them. Yeah. Exactly. So let's see how that goes down.
Speaker 2:Yeah. There's the smile.
Speaker 1:Oh, it is. It's got two layers really quickly, one after the other.
Speaker 2:Oh yeah. So you got the easy rye? Yep. And then the French oak kicks in with the soury, earthy... Oh.
Speaker 2:Interesting.
Speaker 1:That is really nice. I didn't notice it when I tasted it at the But then again, I was really just making sure they didn't taste like crap before committing Of to buying
Speaker 2:course. Yeah. Committing to buy three of them. Yeah. That is very interesting.
Speaker 2:Oh, okay.
Speaker 1:Don't
Speaker 2:know a two for one.
Speaker 1:Yeah. It is, yeah. I don't know who the distiller is now, the master distiller now, but whoever that blending is, this is beautiful. Some research needed.
Speaker 2:Yeah, but also super smooth.
Speaker 1:I mean, I've... Yeah. But it's between the two. It's not as smooth as the first one.
Speaker 2:It's not as No. No. So the first one, the single malt was super smooth. Yep. The triple malt was had a little kick.
Speaker 2:Yep. This has a little
Speaker 1:Not more like a scotch kick.
Speaker 2:No. No. And this slightly more, but still, I... It's not chewable. It's just very drinkable.
Speaker 2:It's very smooth.
Speaker 1:So we kind of got started main and dessert in these three whiskies?
Speaker 2:Maybe. Yeah. Yeah. Yeah. This is a little too much for an aperitif.
Speaker 2:This might be like in between or maybe even after. Yeah. I have to say, I've been drinking a few of those Finnish Keto
Speaker 1:I've got... You still got one to finish on there?
Speaker 2:Yeah. Rye ones because I've been trying to make Manhattans with some, what's the other, the red thing that you put on there, the thing that's in a martini, vermouth.
Speaker 1:Vermouth.
Speaker 2:Some red vermouth and some angostura bitters and I'm trying to work out what the good...
Speaker 1:Are you writing down your choice... Your mixes then?
Speaker 2:Sorry? I'm...
Speaker 1:Are you writing down what you're putting in there? How much of this and how A much is
Speaker 2:Yeah. Yeah. So I... So basically, you mean no? Well, I...
Speaker 2:Basically, when I did this, I had this little bottle of keto and I just put the whole bottle in it and then I just put a bunch of vermouth in it and I put like three drops of Angostura bitters in it and it was a little off.
Speaker 1:So basically... Pipettes needed.
Speaker 2:No. No. That's true. So
Speaker 1:before we we go to the other English distillery drink that I I wanted to put into, I'm not suggesting that you need to brush that. If you want a clean glass, just let me know.
Speaker 2:Got all time lines. I don't mind. I could do a little. Yeah. Not everything.
Speaker 2:I kind of want to have another sip.
Speaker 1:It is nice.
Speaker 2:It's very nice.
Speaker 1:Yeah. So, we were talking about use cases. And we've worked out that the real secret to this, the real secret to this is really turning around that AI is given the opportunity to make decisions based upon some kind of control or guidelines or guide rails issued by the person. Mhmm. Okay?
Speaker 1:Mhmm. Mhmm. So where the fuck do you start?
Speaker 2:Well, I I kinda wanna divide that up into three parts.
Speaker 1:Can you count to three after four whiskies?
Speaker 2:Can. One seven nine. Yes. Got it. Good.
Speaker 2:So basically, we got the reasoning engine, which is the heart of the AI
Speaker 1:bit. K.
Speaker 2:If we add a component to that, that will fire it off when it sees necessary. For example, when there's a new customer complaint coming in or when there's a new document being added or something. So a little automation to kick off that reasoning thing.
Speaker 1:K.
Speaker 2:For example, we got a new, on our website, we have an interest form and somebody filled out the interest form. Now he goes to research that company, see if the lead is worthwhile to pursue and... But definitely a sales
Speaker 1:though for a minute. Yeah. Let's just of something else as an example. But yes, Okay.
Speaker 2:But let's add a trigger component
Speaker 1:So what would to be a trigger? A new email?
Speaker 2:Don't forget that... Yeah, like, for example, customer complaints.
Speaker 1:Alright. But hold that thought for a second because people should be aware that AI don't always pick up triggers very well. This is where that combination of power
Speaker 2:That automation comes in. Yeah.
Speaker 1:Yeah. That that's kind of stuff.
Speaker 2:So I'm I'm talking about the reasoning engine, which is fabulous. Add a automation component to that, like a trigger.
Speaker 1:Sort of orchestration stuff.
Speaker 2:To say to us humans like, hey. I've seen this. This needs to be dealt with or something. And...
Speaker 1:Okay. I will just mute the closing story.
Speaker 2:Okay. There we go. Yeah. And then the third one, I think is also a very important one, is adding skills.
Speaker 1:You mentioned skills that
Speaker 2:you're mentioned skills
Speaker 1:And that you said you would get into Yes.
Speaker 2:So let me let me dive on.
Speaker 1:I'm assuming this is from some presentation.
Speaker 2:No. It's not. No. No. No.
Speaker 2:So basically, we've got a few AI engines out there. We got ChatGPT. We got Grok. We got Gemini. And 10,000
Speaker 1:of those.
Speaker 2:Yes. Exactly. And one of them is is the clot from a tropic. Yep. One of the things that that does is you can define a skill.
Speaker 2:And a skill is basically something that... Like, we have skills. I can play a violin. I can play... Can tap that.
Speaker 2:No. I can't.
Speaker 1:But... Really? Let's do that on the podcast today. I'll give you a week.
Speaker 2:Yeah. Let AI figure that out. Okay. But we can have a certain skill. And a skill is basically a posh text file where it explains all the things that it needs to do and needs to think about.
Speaker 2:For example, every time you create a PowerPoint presentation make sure it's in our brand colors. Make sure that our brand is mentioned every two slides. Make sure that the text that you put in is always in Dutch, never in English, but it also never uses words that have more than three syllables, uses the... Our key core values, things like that.
Speaker 1:So this is like the setting in ChatGPT about
Speaker 2:It's not a setting. It's basically... It's a combination of memory and a very specific thing that it could do.
Speaker 1:So how do you put it in?
Speaker 2:Well, you put it in as a as a markdown file, which is basically a posh text file. Okay. So all in it is write down what it needs to do, but it can also write itself what it needs to do. And it's basically a pretty long text file where you can just change all the things that you want. I'm using this or I'm gonna use this.
Speaker 2:Let me...
Speaker 1:Make a size.
Speaker 2:Yes. So I've got a customer and they wanna make project sites. They will do a few 100 project sites, in SharePoint, every year. And my task is to create a template for this, the SharePoint template. Now the problem with the SharePoint template is that I can't create a nice, homepage for the site in the template.
Speaker 2:It just is not there. So what I can do or what I suspect I can do Yeah.
Speaker 1:What I might be able to do.
Speaker 2:And the skill is a project site, creating a project or, configuring a project site. And in that skill, can say, you know what? I need five document libraries. I need these content types added. I need this and this and this.
Speaker 2:These kind of people need to have access. Look at this kind of system. Look at who's involved in that. Add them to this site and I can just write that skill. And every time I say to, Copilot, we need a new project site.
Speaker 2:It will look at that skill and it will follow all those instructions. And basically what that does is AI is generative. So it means every time I ask it the same question, it's gonna give me a different answer. With a skill, I can make that answer predictable.
Speaker 1:So this is in Claude, and there's a box called skill, and you fill in the text?
Speaker 2:It's not really a box. It's it's... How how it's working in SharePoint is there's a SharePoint skill library. It's like a document library that's called skills.
Speaker 1:Okay. Now I'm starting
Speaker 2:to see. And you put all those files in there and then it can just reference those items.
Speaker 1:And this happens because of the Claude front end that...
Speaker 2:Yes. So you need to enable Claude for that. But I mean it's coming... It's it's it's very new. It's been there for two weeks now or one week.
Speaker 2:So... No. But but let me...
Speaker 1:In AI terms, that's four years.
Speaker 2:Sorry. Sorry. Sorry. No. Claude has hit for nine months.
Speaker 2:But now with Claude coming to Copilot, we now have the skills library in SharePoint that's been there for a week. Correct. So if you ask Copilot to create a skill, it will create that library in the SharePoint site, and it will put those skills files in there.
Speaker 1:That's interesting. Does it work out which skills files it needs based upon the prompt or the request from Yes. The
Speaker 2:Yeah. So basically, I had a a really good demo from Edina Oleven, which is a Dutch MVP. Yeah. Who was playing around with that, and she made a presentation about that. That's cool.
Speaker 2:How she had Lego built manuals in the list, and she had all the numbers of all the the boxes. Yeah. And she said these are, ones I've already created. Look at manuals. See how many steps there are and we say 20 steps is an hour.
Speaker 2:So fill out the metadata column that says how many hours it would take to fill out this Lego set. And if I ask it, I've got two hours this week to create a Lego set. Which one would you recommend that I haven't built yet? And then it would give me an answer.
Speaker 1:I have one response to that, I think. Bore is fucking cool.
Speaker 2:Yeah. It's it's really cool. And talking about use cases, I think that's where this is gonna
Speaker 1:This start with whole use case thing, I don't know whether you and I are on the same page with that definition. Probably not. Think we're always five, No. That's fine. Because I agree with what you're doing, but all you're doing is adding guidelines.
Speaker 2:Yes.
Speaker 1:Yes. I know. Yes. And agree with that. But that's not going to allow AI to fly around everywhere and suddenly go, hey, guys.
Speaker 1:Your production process is missing a step.
Speaker 2:No. But you know what it will do? When there's... Let... Let's go back to that customer process.
Speaker 2:So this might be a high thing or a low thing and the probability might be high or low. But then we could all... If if we really let AI do its
Speaker 1:Which thing is what I'm trying to work out how to get So
Speaker 2:it might automatically start contacting that customer and say, hey, I just saw you did this and this and this. And it could ask follow-up questions. It could it it could do the whole sales process by itself. But with the
Speaker 1:Assuming that somebody's gonna respond to it.
Speaker 2:Yes. But with the skills, we can make the things that it's gonna ask and propose more of what we want it to do. For example, if we wanna focus more on selling product B instead of product A, we might teach that as a skill and the AI will try to sell product B more than product A.
Speaker 1:So where's human in the loop in this?
Speaker 2:The person writing the skill.
Speaker 1:Okay. So
Speaker 2:then And then just say So now the sales
Speaker 1:manager doesn't deal with people anymore. No. No. No. That's interesting.
Speaker 1:Is that a reality?
Speaker 2:It is.
Speaker 1:In your world? No.
Speaker 2:No, because I really have time to play with it. But I think this is gonna be a very big opportunity. So the So what did you
Speaker 1:call yourself?
Speaker 2:Engine, with automation, and those skills, those three combined, I think that is gonna make AI a superpower.
Speaker 1:Interesting. So what did you call yourself
Speaker 2:at the beginning of this podcast? The apocalyptimist.
Speaker 1:The apocalyptimist. Yeah. Interesting. So we'll have an apocalypse, but you're optimistic about it.
Speaker 2:Yeah. Yeah. The world's gonna go to shit, but it's fine. That's good.
Speaker 1:So so what about AI taking over your job then?
Speaker 2:Well, I I I had a few sessions about it. And in a few weeks in the collab days Netherlands, I'm I'm delivering that session again, How AI is gonna take over my job as a SharePoint consultant. And... Yeah. It it really looks like that because basically now we had the knowledge agent in SharePoint that could come up with metadata, which is okay.
Speaker 2:It's cool because we need to categorize our data for Copilot to be optimal.
Speaker 1:So so let's just take this as an opportunity for selling a cake. All right? As a big picture. But before we do that, I want to tell people about the whisky we're about to taste and then we'll close down with some weird things I've got in my head about selling a cake based upon your skills.
Speaker 2:Okay. Selling a cake. Okay, okay.
Speaker 1:So, I'll tell you about the whisky while you pour a little bit in. Okay. So, on the way out of Norfolk to my mother's, as I said earlier, then I came across a sign that said English Distillery. And, of course, the English distillery we have tasted before. Matt, don't know whether you're still listening, but great call out for that.
Speaker 1:And so I actually found the distillery and had a distillery tour, which was really very, very good fun. But then I found
Speaker 2:They're always good fun.
Speaker 1:They are actually.
Speaker 2:If people have never done a distillery tour, even if you don't like whiskey, do it.
Speaker 1:Do it. Go ask questions. Learn about it. It's pretty cool.
Speaker 2:And they have... That's something AI will never get.
Speaker 1:That is true. Yes. Although somebody did the AI whiskey, it didn't really catch on. No. Anyway, they basically have some really, really good whiskies now.
Speaker 1:They'd be... It was the first ever English distillery. But they pointed out and said, look, you know, one of the things that we've started doing is that if I have a barrel, of course, the angel shared 20% is gonna disappear or whatever the number is, and so I might get a hundred and two hundred and twenty five and a half bottles. Yeah. And that half bottle is useless.
Speaker 1:So what we started doing about six years ago was we, based upon whether it was peated or unpeated, we would add it to the barrel. So they had a barrel for unpeated and peated, and they just kept adding these half a barrels in, which for a distillery to be fair is a significant amount of whiskey. Yeah. Yeah. It's doing so well today.
Speaker 1:And then when the barrel was full, they bottled it and they called it?
Speaker 2:The infinity The infinity cask. So
Speaker 1:it's a cask that is constantly filled. Alright? And then when it's filled, they bottle it. I'm guessing they taste it first just to make sure it's not really, really shit. Now this is the tenth cask.
Speaker 1:Yeah. So the first cask was peated, the second one was unpeated, the third one, so we're actually tenth version in. The next one will actually be peated. So they've done 10 barrels of this at 250, 300 bottles a barrel. There's there's a fair few.
Speaker 2:I know Glenfidig does something like that as well with their Solara cask.
Speaker 1:Oh,
Speaker 2:right. Just they just top it off, and they just, yeah, take take bottles out of it on the bottom and they just keep topping it off on the on
Speaker 1:the top. Well, I thought the story was so interesting. I just had to. I got to taste this one and I noticed.
Speaker 2:Basically, it's it's what I wanted to do a few years ago as well.
Speaker 1:So... Instead of waiting for the last inch.
Speaker 2:Yeah. Exactly. So my idea was that for all the whiskies I like, if there's just one inch left, I will just throw that into one nice pitcher and leave that. And the result should be a combination of all the best whiskeys I have. So, the result should be amazing.
Speaker 2:Spoiler alert. That is shit. I threw it away because it was horrible.
Speaker 1:Well, we just tasted the the Andam's one where we... It was a rye one in a French cask, and we've already got this playing off with each other, but it's playing off with each other perfectly.
Speaker 2:Yes. And that's where the master distiller comes in.
Speaker 1:Remind me about that when I talk about cakes in a minute or two. Sure. Alright. But have a taste of this because this is... I mean, we like a Macallan.
Speaker 1:Alright? I actually got back into Macallan just recently. I know.
Speaker 2:I know. I get it. I get it. Funny story. In Orlando with Vesco.
Speaker 1:Yes. That I remember.
Speaker 2:He's ordering Macallan's and I'm sitting there. I don't really like a
Speaker 1:Macallan. I'm drinking it and I'm like,
Speaker 2:I still don't like a mechanical. So, Vesco goes like, That's okay. I've got you a second one. He's like, Okay.
Speaker 1:I had one the other day and I'm going, Oh, actually, you know, maybe I'm being a little biased on this. Yeah. It was quite nice.
Speaker 2:By the way, thanks Jabra for paying for our Macallan.
Speaker 1:Nice. But anyway, but but if we talk about Macallan, it's a grown up whiskey. Yeah. I mean, it's got age, it's got maturity, and it's got knowledge and everything else.
Speaker 2:Now, whether we... It's what it's... Well, yes. Let's say yes. Because I kinda wanna visit Macallan at some point and you can only get in with an appointment because there's a big gate around it and you can only get in if you have an appointment.
Speaker 2:I kinda wanna
Speaker 1:visit Yes. Them at some
Speaker 2:Potentially.
Speaker 1:We should. Let's start making noises. So
Speaker 2:I'll I'll be I'll be nice and gentle. But, yes, I kinda wanna visit Macallan. So, yes, it's mature and it's grown up and it's all those things. Yeah. Yeah.
Speaker 1:Yeah. Yeah. Alright. Cool. But anyway, this, I think, really is a grown up whiskey that that is coming out of this bottle.
Speaker 1:It's just freaking amazing in in terms of its depth and complexity. It is. Look at the sip. Wow. I had all of their whiskies to choose from and this is the one I took.
Speaker 1:It's really good.
Speaker 2:Is proper complex. Yes. This is layered. This... Yes.
Speaker 1:The finish is amazing. Yeah. It's a finish that builds though. It's not like, hey, This actually builds. So it gets better and better and better.
Speaker 1:Problem with this is not gonna last long.
Speaker 2:No. But this is not a whiskey I would drink every day. I mean, that that French oak one, I could have one every day. This one is is not for when I'm sitting on the couch watching a movie. No, this is when I'm...
Speaker 2:When I had like... When I've created something and I'm proud and I kind of want to celebrate it, this would be the one I would have. This is delicious. And it's delicious in a specific way. It's not delicious like I love beaded whiskies.
Speaker 2:I love whiskies with honey. Things like... Has nothing
Speaker 1:No. It's... It has depth. Has citrus. Freshness.
Speaker 1:Yeah. It it has all of those when it shouldn't really have any of them. It it's it's bold. It's forward and dynamic. And then what I mean dynamic, it's not constantly changing, but it actually surprises you in different levels at different times.
Speaker 2:Here, I've got apricots and earth tones and there's the tea, like, there's a tea that's a red red tea. It's called... In Dutch, it's called rose bottle. But...
Speaker 1:Oh, the rose petals, like the Turkish Delight.
Speaker 2:No. No. No. No. No.
Speaker 1:God. The phone is coming
Speaker 2:out. The
Speaker 1:phone is coming out.
Speaker 2:I need to look that up.
Speaker 1:This we said before, the English distillery is the first English distillery for whatever hundred years. Keep that for a moment. This whisky is showing that level of experience and depth that they have created over the years.
Speaker 2:Absolutely. Yes.
Speaker 1:Go on
Speaker 2:then. I'm talking about rose hip. Yes. Okay.
Speaker 1:There's rose petals. Okay. So that's
Speaker 2:not really rose petals. It's it's a different kind of thing. It's it's this.
Speaker 1:Yes. Okay.
Speaker 2:It's called rose hip But apparently,
Speaker 1:that's okay. You know, before they go...
Speaker 2:Oh, apparently it is the fruit coming from roses. Yes.
Speaker 1:You know that red behind you from the Mother's Day roses there? You see that? Oh, yeah. Now before they opened up into that flower.
Speaker 2:That's rosehip. Okay.
Speaker 1:That's rosehip.
Speaker 2:Oh, look. I've learned something. Awesome. Amazing. That's cool.
Speaker 1:Alright. So while we finish this off then Yeah. I I kind of wanted to throw something in. And to be honest, four whiskeys on, I'm not sure how valid this is going to be.
Speaker 2:But It's gonna be good.
Speaker 1:It's gonna be basic. But... So I wanna get practical about stuff. Mhmm. So a lot of the stuff around AI is about knowledge and information, which basically says we can only use AI in the thinking part or the knowledge part of an organization.
Speaker 2:Well, that's where it delivers the most value. We can use it everywhere. No.
Speaker 1:No. Okay. Baking a cake or a cake Sure. So we create a cake, a customer comes in, they goes, I'll have that lemon cake, please. And they put it in a box and they walk away from it and then the story.
Speaker 1:Where is AI gonna fit into that kind of process?
Speaker 2:Everywhere. Okay. Yes. So, first of all, we wanna... We we we need to have a shop.
Speaker 1:No. No. Before you go any further. Okay. I want AI to be the core process, not just a series of things where we can use AI to improve it.
Speaker 1:It needs to be the core process.
Speaker 2:Okay. So it can't create a cake? Nope. But it can do everything else around. It can do the sales, can do the marketing, it can do the...
Speaker 1:No. Again, you're using to... So one of the things that you have been doing a lot of is you're talking about using it as a tool, not driving the process. So your skill really is a way of of guide railing AI so it can go and do something within a restrictive area. And I'm trying to...
Speaker 1:I'm not trying to fix this by the way. No. I get it. If we talk about Atlas Copco. We worked both worked in Atlas Copco, an amazing company.
Speaker 1:Yes. Building your own products. Yes. We've both been in the production line and they get told what to do when and how many products.
Speaker 2:AI
Speaker 1:is not gonna take that spanner and zoop it in. So where is the user value?
Speaker 2:But if we... So let's say, at the end of the line, they come out and we test them before we ship them, those test results might be interpreted by AI to say, you know what? I think the left plate might be a little too thin.
Speaker 1:But by then, it's too late.
Speaker 2:Well, it it could be in the... Somewhere in the production process where we have a camera looking at it.
Speaker 1:Okay. So now you're talking about robotics on a production line.
Speaker 2:Yes.
Speaker 1:Which already is there. But no. No.
Speaker 2:No. But the robotics will only give you data, and AI will interpret that data.
Speaker 1:Yeah. But the same as it already knows whether something's too light or too loose or not tight enough. That's not AI. It will say, hey, that's not quite as close together. Tighten it up two more areas.
Speaker 1:That's automation, it's workflow.
Speaker 2:That comes back to the first version of AI which is hot dog or no hot recognition saying this is a hot dog or this is a daxon.
Speaker 1:Again, AI is a tool. It's doing just, oh, what does that look like? Okay.
Speaker 2:No. But it's
Speaker 1:true. I mean, if
Speaker 2:if you're if you're if you're young and you wanna have a job that is not gonna be affected by AI, be
Speaker 1:a hairdresser. A
Speaker 2:hairdresser, a cake baker, just a bakery. Things like that. Okay. Yes. For now.
Speaker 2:Because when robotics enter and I know there's some very smart people building very high end robotics that are gonna enter the market next year in two years. That is gonna be the next level shift and that is gonna be bonkers. Bullshit. I don't think so. No.
Speaker 2:And we are not ready for that.
Speaker 1:We're not ready because every
Speaker 2:household is gonna have one or two robots that just gonna shoot.
Speaker 1:Nobody... Okay. Here's what's gonna happen. Alright? They're gonna come on the market at €4,000 a pop.
Speaker 1:Yes. You can afford one. Yeah, can afford Two. We're both single. Yep.
Speaker 1:Alright, there are families out there that would never afford them.
Speaker 2:Of course. But that's... It's like it's like an electric car. It's more expensive than a normal car.
Speaker 1:Stop the process because this is getting into social stuff. My question to you is how do you put AI in the process? So, there are so many things where it's impossible at this point and maybe there's shit coming around the corner and as you say robotics is one, so production is another, okay? But it's a long way off. But there is still, how many of that is a percentage?
Speaker 1:60, 70% of stuff where AI cannot do? Only focusing around people based organ orchestration or automation where it can help.
Speaker 2:So basically, we're in your kitchen. We have a counter full of stuff. We have whiskies. We have plates. We have glasses.
Speaker 2:We have all these things.
Speaker 1:All my electrical shit. These are
Speaker 2:all created by humans. Correct. And there's not really ways where AI could be could bring that much extra in, I think.
Speaker 1:No. But it could scan this place.
Speaker 2:Yes. I had a conversation this week, with somebody, who had like a dead space in their garden, took a picture of that and say, you know what, AI? This is a dead space. How can I liven this up? And AI would give her images and recommendations.
Speaker 2:They put it on there and it's now a nice spot with a like, a fire pit and
Speaker 1:a chair. You know something? I love that, but it's still a tool.
Speaker 2:It's still a tool. Yeah. Exactly. Yeah.
Speaker 1:It's interesting. I I mean, I still think there's a lot to go. I mean, some of the weird stuff that I've been thinking about is things like reality, so selling a house because I can't say that word after this number of Whiskies. But, you know, throwing a little robot around the house that is taking the pictures, that's scanning the size, that is doing the management.
Speaker 2:Or you can use a drone for that?
Speaker 1:Whatever we fucking use
Speaker 2:is No. Gonna be driven by Absolutely. But
Speaker 1:again, it's still a tool.
Speaker 2:Yes. It's one
Speaker 1:part of the process.
Speaker 2:Those house robots will have AI in there. So no. No. Let's let's let's not let's not use house robots. Let's use...
Speaker 2:You know I'm a big fan of Elon Musk.
Speaker 1:Yeah? Oh, No. No. No. No.
Speaker 1:So... That's true. You come into the end of the podcast and you bring that subject.
Speaker 2:Wait. Wait. Wait. Wait. Wait.
Speaker 2:So basically, his idea is we need to become interplanetary. Yes? So first...
Speaker 1:According to the press.
Speaker 2:Yeah. Yes. First step, build a moon base. Apparently. And on the moon, we can just dig for minerals and whatever that we need.
Speaker 2:Yeah?
Speaker 1:Yeah.
Speaker 2:Okay. Who's gonna do that? Humans? F no. There's unions and there's all kinds of regulation and safety and people can only wait work that many hours blah blah blah.
Speaker 2:So what we're gonna do is we're gonna shoot a bunch of robots to the moon and they're gonna dig and they're gonna get all those minerals and we go... Just gonna shoot them back to Earth. Yeah?
Speaker 1:Okay.
Speaker 2:So those you... Those robots are gonna be AI driven. They're gonna be AI...
Speaker 1:Why are they? Sorry? Why are they?
Speaker 2:Why? Because why not?
Speaker 1:Because you will have already programmed them to do something specific
Speaker 2:on the Yes. But it's it's not a digger. It's a robot that needs to do 150 different
Speaker 1:tasks. So who's going to... Yes so it's... And if it's automation and
Speaker 2:it finds a big rock it will know oh I need to move that big rock but I can't push it anywhere so I need to find a different way to do it. It's not AI. It's not AI
Speaker 1:being Oh, my drill has stopped. Let me move three meters to the left and see whether that changes.
Speaker 2:No. No. That's a good example. It is reasoning about my drill has stopped. Why has it stopped?
Speaker 2:Is my machine broken? Is it this... Or it that...
Speaker 1:Because it's not moving forward anymore. It's something harder than what I used to be in. It's a standard trigger.
Speaker 2:Yes. But then the question is, how am I gonna solve this?
Speaker 1:I'm gonna take the drill back. I'm gonna move to the left and put it in. It's a little program I run. It's not AI. You know, actually, here's the problem.
Speaker 1:I don't know. It's... Know this is fun to put stupid. I don't
Speaker 2:know if
Speaker 1:you should be recording this. Anyway... Gives a shit? So I haven't turned it off yet. So the robot's gonna go, oh, fuck.
Speaker 1:There's a rock in the way. So if it does the human thing, it's gonna go, let me drill this a little harder. Let me push a bit. I just I then it's gonna go, oh, fuck. The drill bit's broke.
Speaker 1:Let me go. Let me tell you how the AI is gonna do this.
Speaker 2:Dude, had such cool, amazing, adventurous story about how the future's gonna look and you're just gonna go like, oh, it's gonna just
Speaker 1:go like, oh, no. No. Let me finish this because I think it's valid. So I think it's about reasoning and orchestration. I I don't disagree with that.
Speaker 1:But the first thing it's gonna do is gonna pick the robot that can scan the soil to actually work out where the best place to drill is.
Speaker 2:Yes.
Speaker 1:And then it knows that when it does drill, it's gonna work it out. So maybe I get where you're coming from. But all of that is not AI. It's still programmed by a person. And my I guess, the argument here and a bit with the cake shop as well is that it's not about just handing everything over to AI and say, fucking do it.
Speaker 1:It's actually about saying, look. What I would like you to do is this. It's the human saying, this is the logical process that gets us there, and then the AI can improve on that.
Speaker 2:Yes. But that's what you're talking about, about the cake shop. So I'm saying everything but making the cake is something AI can do today.
Speaker 1:It can help with. It can help with.
Speaker 2:Let... Yeah. Okay. Okay. It can't bake the cake, but if we have AI in a robot, then that robot can can make the cake.
Speaker 1:So why do we... Who's gonna eat the cake?
Speaker 2:I don't know. Whoever buys it?
Speaker 1:Okay. And then they say, I don't like this cake. Sure.
Speaker 2:And that is feedback that we can give to the
Speaker 1:AI. The same
Speaker 2:thing that we do with AI right now. I I ask you the question, it gives me an answer. I'm saying, that doesn't work. And then it says, oh, that's right. My bad.
Speaker 2:I'll make something else.
Speaker 1:So what are you gonna do as a job in twenty years time? I'll be dead and gone, mate. I'll... You'll be inheriting my 180 whiskey bottles. So maybe that's what you'll be doing.
Speaker 2:I'll be drinking whiskey all day. Okay.
Speaker 1:Do you know what? I think part of this really is we don't know yet. No. And it's...
Speaker 2:It is gonna change the society. And now we get back to that social aspect Yeah. Where we need some basic income, for example, because... Well, we need to find our purpose. That's basically what we need to
Speaker 1:do. God.
Speaker 2:I know this is so far off.
Speaker 1:I'm gonna go back fifty years. I wrote an essay.
Speaker 2:Okay.
Speaker 1:I wish I could... I wish I had a copy of it now.
Speaker 2:Uh-huh.
Speaker 1:But I remember going to college and writing an essay, and they asked us to talk about computers and what they would do. So it was that Yeah. Bit fifty years ago where I kind of pointed out that socially, the world is gonna fall apart unless the government is able to say, hey. Look. You know?
Speaker 1:Yes. This is now automated, but it will in some way be taxed higher so that people can afford to pay for the automation that is actually in operation because... Yeah. And it's a kind of loop. And and I remember getting top marks out of it and the lecturer going, very good point.
Speaker 1:Well thought through. Mhmm. But there is no real answer in my stupid 16 year old essay. But... Yeah.
Speaker 2:But it's not up to us to come up with answers. That's why we have world leaders.
Speaker 1:Hey. So one more question, and I only want Okay. Answer a or answer b, and then I'm gonna sort the... Then then we're gonna stop stop the podcast. Yeah.
Speaker 1:We're already at ninety four minutes.
Speaker 2:Okay.
Speaker 1:Alright. So here we go. Alright. You have two cake shops in the street.
Speaker 2:Yes.
Speaker 1:One cake shop has got a guy that's got forty years of experience cooking cakes. The cakes get updated on a regular basis.
Speaker 2:Okay.
Speaker 1:And he's got this model there and it's sitting in the window. And the other one is driven by some kind of box with a hole in it and you punch in what cake you want and out it comes. Alright? You have money in your pocket. Which shop will you use?
Speaker 1:A, the guy that is smiling at you and saying, hey, I'm gonna put my experience and I've made something special today or b, a hole in the floor?
Speaker 2:Can can I say it depends?
Speaker 1:You can know. You can say a or b and then I'm gonna... No.
Speaker 2:Well, if if I wanna have a cake, I'm going with the the old guy. If I wanna do this for Instagram likes, I'm gonna go for b.
Speaker 1:You don't have the answer, do you?
Speaker 2:No. No. I'm gonna go... Because I love cake, I'm gonna go with the old man who has been building cake for a very long time.
Speaker 1:And there we have some honesty, guys. Bye.