Latest / Influential Visions / S11 Ep7: Ai Productivity Secrets From Sony's Robotics Pioneer Nicolas Babin

Transcript

S11 Ep7: Ai Productivity Secrets From Sony's Robotics Pioneer Nicolas Babin · Influential Visions

• 0:00
  80% of AI implementations fail.
• 0:02
  How- how do we become more productive under the pressure of this AI Act, Nicolas?
• 0:08
  These are really the two key elements of the AI Act, transparency and trust.
• 0:12
  My own town here in France, in Arcachon, uh, is p- is, um, paying for their, uh, citizens to use the ChatGPT version of ChatGPT.
• 0:22
  Well, I am very privileged today to be joined by Nicolas Babin, and we're gonna talk about AI productivity secrets from Sony's robotics pioneer.
• 0:31
  And Nicolas actually brought the AIBO to life at Sony Europe.
• 0:36
  Uh, he is now an EU digital ambassador, and he will reveal the productivity framework that bridges cutting-edge AI innovation with regulatory compliance.
• 0:47
  Essential insights for leaders navigating the 2025 and onwards AI landscape.
• 0:54
  So, it's, uh, it's really nice to, uh, nice to see you, Nicolas.
• 0:59
  Um, yeah, it's- it's very interesting the AI, uh, AIB- AIBO legacy because, uh, that is- I remember when that was launched.
• 1:09
  Actually, w- I was probably, what was I?
• 1:11
  I was probably about 20, maybe a- maybe a little younger.
• 1:15
  1999.
• 1:16
  Right.
• 1:17
  So- so yeah, so it's about 20 some- 20 something, and- and I remember seeing- seeing that on television.
• 1:24
  But unfortunately, it didn't cover the bit where someone came along and ripped AIBO's head off because I think that's a particularly interesting story.
• 1:32
  I- just- it just made me laugh.
• 1:34
  The last time we spoke, you kind of talked about that, and it's- it's just a classic.
• 1:38
  That was in Manchester though, and we didn't put it on TV.
• 1:40
  It- it would have been brilliant.
• 1:42
  It would have been- it would have been a viral clip.
• 1:44
  Yeah.
• 1:45
  Mm.
• 1:46
  Yeah.
• 1:47
  Right now, absolutely.
• 1:48
  These- in these days, yes, but in 1999 we didn't have, uh, social media.
• 1:52
  No.
• 1:52
  Yeah, of course.
• 1:53
  Of course.
• 1:54
  And- and- but AIBO was- was really actually very cutting edge technology, right?
• 1:58
  Yeah.
• 1:58
  Because it- it- it- it- it- it had, like, you could read emails with this robot that was- that was kind of in your- in your home, right?
• 2:07
  Yeah.
• 2:07
  And it...
• 2:08
  Nowadays, uh, you know, we- we- we're trying to get to grips with being productive, but actually, we're probably less productive than we were 25 years ago if you- if you ask me.
• 2:18
  I'm- I- I'm struggling.
• 2:20
  I actually switched my phone to grayscale because I found out that I was just, uh, being, uh, manipulated by all of these applications with the great colors, and I thought, "You know what?
• 2:30
  I'm gonna- I'm gonna switch that, uh, to grayscale and see how that affects, uh, the impact of my device on- on my brain." And it was- it was surprisingly interesting, uh, if I'm honest.
• 2:45
  Mm-hmm.
• 2:45
  No, I- I- I can understand that.
• 2:46
  Uh, to say that back in 1999 when we launched AIBO, we were the first AI-based product.
• 2:52
  I mean, robot in this case, but anything AI-based.
• 2:55
  As you said, it could read emails with emotions.
• 2:58
  So we're not at all what we are today with AI because today AI doesn't show emotions.
• 3:03
  But AIBO, we had made a robot with eyes that could change colors.
• 3:08
  So when AIBO was happy, he had green colors.
• 3:11
  When AIBO was upset, he had red colors.
• 3:13
  When AIBO was confused, he had orange color in his eyes.
• 3:16
  So you could tell, you know?
• 3:18
  And this is something that you can't see today because obviously emotions, empathy, uh, intuition, all this is really proper to human beings and that's what's important.
• 3:29
  But back in 1999, it was the first, right?
• 3:32
  And- and so- and we wanted people to be more productive.
• 3:35
  As you said, you had AIBO sitting on your desk and he would say...
• 3:38
  You know, you would work on something and you would say- uh, AIBO would come and say, "Oh, Nicolas, you have an email." You'd say, "Okay, AIBO, please read it." And AIBO will read it with emotions.
• 3:48
  So that mean like he would say, "Hello Nicolas," and he will bow like a Japanese person would.
• 3:54
  Uh, and at the- at the end, he will say, "Best regards," and he will wave at you.
• 3:58
  I mean, it was like- it was unbelievable and he would read you the email and he would say, "Would you like to answer?" And you say, "Oh, yes, please." Say, uh, "I'll- I'll deal with it later." You couldn't have a very, uh, elaborate discussion because, uh, you know, at the time NLP was just at the beginning, natural language processing was just at the beginning.
• 4:16
  So what we had to do is we had to reg- record sentences for AIBO to understand.
• 4:21
  So i- you know, the technology wasn't there yet, but it was really based on productivity.
• 4:26
  We had some plans that was unbelievable.
• 4:28
  I mean, Sony at the time was really the best company in the world and the most innovative one.
• 4:33
  And we had plans to say that...
• 4:34
  So AIBO would stay at home, obviously it has a camera in its nose, it will film everything, you could see everything on your PC because at the time, uh, smartphones were not invented yet.
• 4:45
  So like we have today with Nest and you can see what's going on in your house even if you're 10,000 miles away from it.
• 4:52
  It was exactly the same thing, but on top of it, you could move AIBO around the house.
• 4:56
  What we had not anticipated is the fact that a burglar would come in, would see a robot that weighs less than a kilo, and will take the robot with- with under his arm, and- and will go.
• 5:08
  So in that case, you know, you would lose everything, the- the robot and- and your whole house.
• 5:12
  But anyway, so what was important is we really wanted to have people more productive.
• 5:16
  And so we, uh, had AIBO, uh, and at the time again we had answering machines, you probably remember that time.
• 5:24
  Mm-hmm.
• 5:24
  And so AIBO would also be an answering machine.
• 5:26
  So when you would come home from work, you would say, "Hi AIBO, what happened today in the house?" And AIBO would say, "Well, somebody rang at the door, uh, the phone rang 10 times, I have four messages for you- for- for you." Exactly.
• 5:38
  And you- and, uh, you could say, "Well, please read the messages to me." And AIBO will read you the messages.
• 5:44
  You could say, "AIBO, turn on the light in the living room." You would say...
• 5:47
  So it was really like even the start of domotic because we had...
• 5:51
  And it was really the beginning of Wi-Fi and, uh, and for that it was really terrible.
• 5:55
  We had a PCMCIA card that was plugged in the robot and- and linked to your PC.
• 6:00
  I mean, it was really complicated to put everything in place, but it was working....
• 6:05
  and it was really-
• 6:05
  Right.
• 6:05
  ...
• 6:05
  with the idea of two things.
• 6:07
  It's, the first one was entertainment, because Sony is an entertaining company, and a company that makes entertainment products.
• 6:14
  Uh, and the second one was really about productivity, and ensuring that at least we would go into a new era of, uh, of AI, uh, uh, powered tools, like what we see today.
• 6:27
  But honestly, it was back in 1999, and we had hope that it would come up, w- our business plan at the time was that by the year 2005, you know, everything will be AI powered.
• 6:39
  Right.
• 6:39
  Obviously, we're over 20 years, uh, or 17 years, uh, behind what we had planned.
• 6:44
  But that, that was, that was the plan.
• 6:47
  It's fantastically interesting the evolution of these things, and how it was actually there back then.
• 6:52
  And it, and it could've, it could've really blown up, but the, the supporting tech around it just didn't, didn't quite kind of hold up, um-
• 7:01
  Mm-hmm.
• 7:01
  ...
• 7:01
  to what was needed, right?
• 7:02
  It's, um...
• 7:03
  I mean, there was not even 2G at the time, 1999, remember?
• 7:06
  Of course.
• 7:07
  Of course.
• 7:07
  Of course.
• 7:08
  It was still analytics.
• 7:09
  I mean, we didn't have digital phones.
• 7:10
  We didn't have any an- any, any of this.
• 7:13
  So, th- that's what we had to wait for, is all the technology.
• 7:16
  And, and, and unfortunately it was a bit, uh, about Sony history.
• 7:21
  Sony was always ahead of everything.
• 7:23
  Mm-hmm.
• 7:23
  Sony was the first with, uh, VHF, uh, uh, technology, uh...
• 7:30
  Well, actually it was not the Sony technology VHF, but Sony had its own technology then VHF came up and, and Sony lost it, and DVDs, uh, then Blu-ray.
• 7:40
  Uh, and then when Sony finally got a product that worked well, which was Blu-ray, uh, then we got streaming out and Blu-ray pretty much disappeared.
• 7:50
  I mean there's, you still have some Blu-ray discs, but nobody uses them anymore 'cause you have-
• 7:54
  Yeah.
• 7:54
  ...
• 7:54
  Netflix.
• 7:55
  Uh-
• 7:55
  Yeah.
• 7:55
  ...
• 7:55
  so it was always, S- Sony was always ahead of the game in every single department of electronic innovation.
• 8:03
  And unfortunately, it was too early each time.
• 8:06
  In-car navigation system.
• 8:08
  I launched the first-
• 8:09
  Yeah.
• 8:09
  ...
• 8:09
  one back in 1996.
• 8:11
  It was-
• 8:11
  Yeah.
• 8:11
  ...
• 8:12
  ex- the first, first one we worked with, uh, with Mercedes at the time.
• 8:16
  Mm-hmm.
• 8:16
  And, uh, you sh- it was un- unbelievable.
• 8:18
  I had friends, I mean, I think I told you the story many, many times.
• 8:21
  But friends who would ring at the doorbell saying, "Can we have a, a tour in the talking car?" I mean, literally-
• 8:27
  Yeah, yeah.
• 8:27
  ...
• 8:27
  I lived in Wimbledon lived in Wimbledon and you had neighbors lining up on the weekends, um, in front of my house.
• 8:34
  My wife was like, "Okay, you have to stop telling people you have a talking car." It's like, it was unbelievable.
• 8:41
  So I, I used to go around Wimbledon and, and, uh, Kingston and all these areas so at least I could show people how a, uh, an in-car navigation system would work.
• 8:49
  But then, you know, we were again too early.
• 8:52
  And when we stopped it was in 2000, and, uh, it started really to pick up in 2005.
• 9:00
  Yeah.
• 9:00
  Yeah.
• 9:00
  It were, but it-
• 9:01
  So it's like a market-making thing.
• 9:02
  ...
• 9:02
  all about productivity.
• 9:03
  Because-
• 9:04
  Yeah.
• 9:04
  ...
• 9:04
  even, even in-car navigation system is about productivity.
• 9:06
  It's about going to a place you don't know fast, you know, we're working with ambulance services and all this.
• 9:13
  It's everything is about productivity and that's what's important.
• 9:16
  It's an human augmented tool and the augmentation is about the productivity.
• 9:22
  Yeah.
• 9:23
  But that's, that's where we're failing, a lot of people are failing.
• 9:26
  I mean, I, I did a, did an interview, uh, the other day where we talked about 80% of AI implementations fail, right?
• 9:33
  Yeah, 80.
• 9:33
  And a lot of that is to do with basically not working out what's, what they actually need properly in the first place, uh, and there are just so many reasons.
• 9:44
  But generally it's a culture thing, isn't it?
• 9:46
  And, uh, now we're-
• 9:48
  It's a mindset.
• 9:50
  Yeah.
• 9:50
  It's a mindset thing.
• 9:50
  It's a, yeah.
• 9:51
  You're absolutely right.
• 9:53
  Yeah.
• 9:53
  Um, and, and now we sort of have this, uh, this, uh, AI Act, the EU regulation, uh, reality of that is that there's, you know, yes, we can be more productive, but we have to, we have to think about the compliance aspects.
• 10:10
  And I know you're quite, quite hot on that because you're an EU digital ambassador, so you know a lot about that.
• 10:16
  So what, uh, what's the deal?
• 10:19
  How, how do we become more productive under the pressure of this AI Act, Nicolas?
• 10:24
  Yeah, it's because AI is a very different product than anything we've seen before because, it, it's...
• 10:30
  Even if it's artificial it has a little bit of intelligence, meaning, you know, it, it can automate processes and sometimes, on some, uh, occasions and some tasks it can really augment and a little bit replace at least the mundane aspect of your, of your work.
• 10:49
  So an, a very important element of this is the ethics around AI, because, you know, you, you could have some, um, s- some areas of AI that can be dangerous.
• 11:02
  For example, uh, when you have AI combined with face recognition.
• 11:06
  Mm-hmm.
• 11:08
  Then, you know, you could...
• 11:10
  And we see that in some countries, uh, where you have social scoring, so you do something good in the street where you have cameras everywhere, then you get plus 100 points.
• 11:19
  You do something bad, you do, get minus 100 points.
• 11:22
  And at the end of the month's quarter year, basically you, based on your score, you can have access to a hospital or not.
• 11:31
  You can have access to, uh, the, uh, city hall services or not.
• 11:36
  Uh, and that, that, that is really bad.
• 11:38
  And so what Europe has decided to do, because today you have basically three big polls around AI.
• 11:43
  You have the US, you have China, and you have Europe.
• 11:47
  Right.
• 11:47
  And Europe is was, is the first, um, uh, s- uh, number of countries, you know, uh, continent if I could say, even though China is not a continent, but, uh, um, that has decided to, uh, have a legislation around it in order to promote transparency and trust.These are really the two key elements of the AI Act, transparency and trust.
• 12:13
  So based on that, Europe has, has, uh, looked at four different types of, uh, risks from the non ac- not acceptable risk, the one I mentioned, the, the, uh, the fact that you have a social scoring, that's not acceptable in Europe, to the least, uh, to- to the least impactful, uh, basically risk where you're talking to a chatbot and basically what we're asking the companies is to say you're in communication with a chatbot here, you know- ...
• 12:43
  or with an AI algorithm.
• 12:45
  Yeah.
• 12:45
  That's it.
• 12:45
  And in the middle you have two different type of risks.
• 12:48
  One that's, uh, average and one that's a little bit, um, more advanced.
• 12:53
  Uh, and for these two risks we ask that, uh, companies get a, um, a, a logo, you know, EU approved, uh, type of logo, uh, that, that we see everyone, every single product.
• 13:06
  You're in Croatia, you have every- everything you buy has an EU, uh, basically logo on it, meaning that it's-
• 13:13
  Mm-hmm.
• 13:13
  ...
• 13:13
  allowed to be sold in, in, in Europe.
• 13:16
  And it's the same for AI basically, whereas for outside companies or third party companies to providers to come look at your algorithm, decide if it, if it should be going to the non-acceptable risks or if, you know, within a, a framework, it can be accepted but y- with total transparency.
• 13:38
  So we're avoiding basically black boxes and that's what-
• 13:41
  Right.
• 13:41
  ...
• 13:41
  everybody's wear- awar- worried about is the fact that you can have an AI and a black box whe- where the algorithm sits.
• 13:49
  You have no idea what kind of data comes in.
• 13:51
  You have no idea, uh, what's, uh, bias, uh, can be input in it.
• 13:57
  For example, there was a very famous HR product in the US that they found out later on, and it was not intentional in any way, shape or form.
• 14:06
  But for some strange reason, the, the person who programmed that HR product had a bias against colored people.
• 14:14
  Hmm.
• 14:14
  And so resumes that would come in from colored people had less chance of being picked by the company than resume from white people.
• 14:23
  Mm-hmm.
• 14:23
  Mm-hmm.
• 14:23
  And, but i- it took a while, you know, because, uh, they've suddenly realized that even colored people with great backgrounds were not selected.
• 14:33
  And so it took a while.
• 14:34
  In Europe, we want to avoid that.
• 14:36
  We want to make sure that-
• 14:36
  Right.
• 14:36
  ...
• 14:36
  before the product is on the market, it has been proven that there is- there are no bias in the, uh, in, in the algorithm.
• 14:44
  And so th- that's really where, where we come from.
• 14:47
  Uh, at the time when we launched it, so that was last year, August 1st, uh, so it's over a year ago, we had, uh, we were working with the state of California, we're working with, uh, Brazil, many countries basically to, um, to make it, uh, more of a global standard.
• 15:05
  Unfortunately, things have changed a bit in the US, uh, but we're still working with Canada and other countries, uh, because they realize that actually it's very important, uh, for everyone, for everyone's productivity, for everyone's, uh, benefit to actually have a, um, a, uh, a legislation that- that's gonna protect businesses and citizens alike.
• 15:28
  Right.
• 15:29
  Right.
• 15:30
  And, and, and in terms of productivity, what's-
• 15:34
  Mm-hmm.
• 15:35
  ...
• 15:35
  what sort of, what sort of things are you, are you seeing people using AI for?
• 15:41
  So a- a- as, as, as you know, 'cause we talk about it regularly, you and I, uh, there are many different types of AI.
• 15:49
  So the reason why AI has come up in front of everyone is because OpenAI has come up with ChatGPT and which is based on a model called LLM, the large language model.
• 16:00
  So this is really the one that we see the most today, uh, in, uh, on the market because everybody has heard of ChatGPT.
• 16:07
  Everybody's using it.
• 16:09
  My own town here in, in France, in Arcachon, uh, is p- is, um, paying for their, uh, citizens to use the ChatG- the paying version of ChatGPT.
• 16:18
  Wow.
• 16:19
  So it's, it's to help people address, um, administrative issues, uh, to help people also, you know, be part of this, uh, ever-changing world.
• 16:30
  Uh, and so it's, you know, everybody has heard about it.
• 16:32
  So it's true that companies now in terms of content creation, that's what LLMs do, they, they use a lot of, uh, of, uh, ChatGPT and, and, and likes and Gemini and, and, and Clou- Cloud, which is also really good and, and others.
• 16:47
  Um, so that's really where I see it, but I see it as an entry point to AI because AI is so much more, as I mentioned, you know, you can have predictive algorithms, you can have recognition of images, recognition of videos, or you can have, uh, images recognition, especially in healthcare, which is an area where I work a lot.
• 17:08
  So any, uh, imagery, uh, is, uh, is actually, uh, uh, we have databases of billions and billions of images.
• 17:16
  So, um, uh, doctors can find, um, disease way before a human eye can see it because a human eye can only see, for example, a tumor when it's at least two millimeter hi- uh, growth.
• 17:30
  And if it's less than two millimeters, then the human eye cannot see it, but the AI can because it has been trained for it.
• 17:37
  That's the, uh, the element that's really common to all AI systems is the fact that you have to train the, uh, the algorithm.
• 17:46
  So ChatGPT has been trained, and I'm sorry, I mean, I should have said LLM models have been trained on data that, you know, you find on the web and things like that, that are available.
• 17:57
  Uh, images, uh, basically doctors have entered every scanner, every, uh, X-ray, every, uh, e- everything that basically you do as a, as a patient-...
• 18:08
  they've been entered and, and the algorithm have been trained on it.
• 18:11
  And that's the reason why, uh, uh, a, uh, an AI-based system in a hospital, for example, can say, "Well, I have in my data your parents', uh, file, your grandparents' file, your great-grandparents' file, files." And so I know basically that potentially, based on your historical information, based on what I see today, is predictive- we can predict that in the next five, 10 years, you would have, uh, an issue there.
• 18:39
  And so you can be, you know...
• 18:42
  The, the key in healthcare is the fact that if- th- the sooner the problem is found, the better it is in terms of your prognostic.
• 18:50
  Right.
• 18:50
  And that, that's exactly what AI does, is it, it creates a productivity for the, uh, uh, radiologist that's absolutely, uh, second to none basically-
• 19:01
  Mm-hmm.
• 19:03
  ...
• 19:03
  because, uh, because they can see things that in the past they could not see.
• 19:06
  And so it...
• 19:08
  Y- you know, obviously you still have cancers that are very difficult to predict.
• 19:11
  But AI is really here to optimize, uh, and, and to augment the, uh, the doctors.
• 19:17
  And, uh, my son is a doctor and he says, you know, "You can really see now the shift between older doctors over 50 who say, 'I've been doing this my whole life, I know what I'm seeing-
• 19:26
  Mm-hmm.
• 19:27
  ...
• 19:27
  and I know what I can do.' And younger doctors who say, 'Well, let's, let's combine our knowledge with AI-
• 19:33
  Yeah.
• 19:34
  ...
• 19:34
  and in that case, uh, be a little bit more, uh, much more productive than, than anything we've seen before.'" And that, that's basically-
• 19:42
  Mm-hmm.
• 19:42
  ...
• 19:42
  w- what I see in, in companies today, is really the entry point being LLM.
• 19:47
  So because w- we talked earlier about mindsets, and that's exactly it.
• 19:51
  You need to have people change their mindset in order to accept a digital transformation like AI.
• 19:57
  Mm-hmm.
• 19:57
  And so, when they realize that the mundane work can be done by AI, as long as they are behind it and they validate the outcome...
• 20:06
  Because that's, that's what the most important, and that's also what the EU AI Act does, is ensure that humans are at the center of everything.
• 20:13
  Well, once you realize that your mundane work can, instead of being done in five hours, can be done in 10 minutes, uh, and for...
• 20:21
  I have ano- another example which I just finished a project for, for a lawyer, um, who, uh, really didn't like to write contracts.
• 20:30
  I mean, he loved going to the bar and, uh, and defend his client in front of the judge.
• 20:36
  But he didn't like the, the mundane work of always doing p- as you know, legal...
• 20:42
  All contracts look very similar.
• 20:44
  Obviously you have paragraphs that are different but the, the overall bulk of the, of the contract is the same.
• 20:51
  And so this guy really would say...
• 20:54
  It would take him two or three days to write a contract, and then after that, you know, explain to the client, make changes and go to court.
• 21:01
  Well, now he uses an LLM model that does it in less than an hour.
• 21:08
  Uh, then he spends about-
• 21:09
  Yeah.
• 21:09
  ...
• 21:10
  half a day reviewing everything because you-
• 21:13
  Mm-hmm.
• 21:13
  ...
• 21:13
  really don't want to produce something that you haven't reviewed, and then he goes to court.
• 21:18
  Yeah.
• 21:18
  So he gains about two days, which you can imagine, 48 hours, I mean, it's like unbelievable, of, of his time-
• 21:26
  Yeah.
• 21:26
  ...
• 21:26
  by using an LM model.
• 21:28
  One thing that's very important, there was a big case in the US of a, uh, of a, a s- a lawyer who, uh, used a, uh, an LM model, uh, to, to, to go to court and he didn't review anything.
• 21:43
  And LM models, as you know, can invent things or can find stuff in their training data that is not correct.
• 21:51
  And so that model had actually made up some of the legal, uh, or some of the laws, and used some laws that didn't exist.
• 22:00
  And so the guy was disbarred.
• 22:02
  Wow.
• 22:03
  Wow.
• 22:03
  Disbarred.
• 22:03
  He was, he was, uh...
• 22:05
  He had to leave the profession.
• 22:05
  Well, it can be, can't it?
• 22:07
  The risk-
• 22:08
  Yeah, no-
• 22:08
  Uh, the risk is, is huge, isn't it?
• 22:10
  For not, not
• 22:10
  Yeah, that's exactly it.
• 22:12
  So, to my point of the fact that you absolutely need to make sure that human is at the center of everything, as a doctor, as a lawyer, as a, as a businessman, as a finance person, anything.
• 22:23
  You need to review everything.
• 22:25
  But reviewing is much faster than creating, and that, that's the story.
• 22:30
  Yeah.
• 22:30
  Yeah.
• 22:30
  So that, that's the whole story around productivity, and that's why we had to figure it out when we launched Aibo, you know?
• 22:36
  We thought Aibo could do the work that, you know, nobody...
• 22:39
  Was a pain in the neck in the past.
• 22:41
  Like you would enter home, you would go to your answering machine, you would press play.
• 22:45
  "Oh, sorry, I didn't hear." You would press rewind, you would press play again.
• 22:48
  Mm-hmm.
• 22:48
  Here you could just tell Aibo, "While you're cooking, please read me my, my messages."
• 22:53
  Right.
• 22:54
  Right.
• 22:54
  So, so-
• 22:55
  That's, that's Productivity.
• 22:57
  Yeah.
• 22:57
  So, so it's basically just mirroring the tasks that you're already doing and finding a way to make them faster without creating errors because-
• 23:05
  Absolutely.
• 23:05
  ...
• 23:05
  you can create huge errors, right?
• 23:07
  But it's interesting all of this, because, um, I, I was reading something yesterday about, uh, how actually using language models, uh, to write things is making us dumber, which was very interesting.
• 23:20
  And there's a bit of research into that which I'm gonna dig into in the coming weeks, um, and have a, and have a look at.
• 23:26
  But...
• 23:27
  So I, I, I don't like, I don't like to use it to write things.
• 23:32
  I like to use it to correct my grammar and maybe give me some structure or swap some words out here and there.
• 23:38
  Because otherwise I find that my writing ability goes down, the quality of, of output as well goes down actually, and it becomes, uh, uh, homogenized, right?
• 23:50
  So it's kinda like, it's kinda like...
• 23:53
  Uh, and by...
• 23:54
  I think they said by 2030, there's not gonna be anything original....
• 23:59
  on, in those language models because we will have run out of creativity because we are creating, like, an ecosystem of continual, uh, spinning of, of-
• 24:09
  Yeah.
• 24:10
  ...
• 24:10
  content which someone else has written.
• 24:12
  So, it's a-
• 24:13
  Yeah.
• 24:13
  ...
• 24:13
  it's, it's, it's, it's an interesting thing.
• 24:15
  But you're right.
• 24:16
  There's so many applications that can be used.
• 24:19
  I mean, Daryl, uh, Mann, who we were talking to the other day, he uses, now they use AI to look at patents, to scan all the patent, uh, databases to find specific things that they can use for clients, right?
• 24:34
  Yeah.
• 24:34
  So there are so many applications that, that really it's, it's not about, uh, giving everyone a list of, of how they can be more productive.
• 24:43
  It's about actually using, probably using a language model or a Google search, which is now powered with a language model, right?
• 24:51
  Absolutely.
• 24:51
  To find something specific-
• 24:53
  Yep.
• 24:53
  ...
• 24:53
  which will connect something else together.
• 24:56
  But unfortunately, big, big companies are, uh, their innovation is stifled by, uh, the, the framework, right, that they operate in.
• 25:08
  So-
• 25:08
  Yep.
• 25:08
  ...
• 25:08
  the smaller, more innovative companies, uh, and independent businesses can actually move a lot faster in this regard.
• 25:16
  And whilst they don't have the enterprise tools, many of them, because they're very expensive, they actually have the ability to bolt together lots of different tools, which could in theory make them, you know, uh, ahead of a lot of these big companies, right?
• 25:31
  Yep.
• 25:31
  Absolutely.
• 25:31
  And companies, uh, I found, uh, at least with my clients, the companies that use...
• 25:35
  And I'm sorry, I'm gonna sp- preach the choir here, but companies that can, that use outside consultants also, uh, at least in terms of digital transformation because they are not bogged down into their daily, you know, problems.
• 25:49
  Right.
• 25:49
  Mm-hmm.
• 25:49
  And so you need to...
• 25:51
  As you, exactly, uh, what you said, you need to get a little bit outside of the box, right?
• 25:56
  Yeah.
• 25:56
  In order to, to move forward.
• 25:59
  Because most of these companies that fail, and the, the, the percentage is actually very scary 'cause a lot of them fail because basically what they say is they go to the IT department and say-
• 26:09
  Yeah.
• 26:09
  ...
• 26:09
  "Oh, I need you to implement AI." And the IT department, you know, even if the guy has, uh, 30 years of experience, he has experienced in normal- normally within their, uh, industry.
• 26:21
  And so they-
• 26:21
  Yeah.
• 26:21
  ...
• 26:22
  cannot think about outside of the box, and that's what's important.
• 26:24
  Even if you don't use it, an internal, an external, sorry, consultant, um, for the whole project, at least you need to have somebody who can just open a little bit of that box and say, "Hang on, guys.
• 26:36
  This is the real world and, uh, how it works," right?
• 26:40
  Just don't-
• 26:40
  Yeah.
• 26:41
  ...
• 26:41
  stay within your, your own framework.
• 26:43
  Yeah.
• 26:44
  Yeah, absolutely.
• 26:46
  So I think it's a question of implementing, uh, a nice three-step framework, right?
• 26:50
  Of just-
• 26:51
  Yep.
• 26:51
  ...
• 26:51
  you know, working out what's going on, deciding on the technology you're gonna use, and then implementing the innovation with, uh, with both external and internal teams, right?
• 27:03
  That kind of scene-
• 27:04
  And I would add a fourth one.
• 27:05
  ...
• 27:05
  type thing, right?
• 27:05
  Yeah.
• 27:05
  I would add a fourth one-
• 27:06
  Mm-hmm.
• 27:06
  ...
• 27:06
  which is basically, um, doing it within the legislation, es- especially if you're in Europe as, as you and I are.
• 27:14
  Of course.
• 27:15
  Yeah.
• 27:16
  Uh, because it, it's really important today.
• 27:17
  You, you can't suddenly launch a, a digital information, uh, a digital transformation project without having this, uh, legislation in mind.
• 27:26
  So what I do a lot for my clients today is I work on charts.
• 27:30
  Mm-hmm.
• 27:30
  Uh, that it's, it's hard because most companies, you know, try to implement them, but, uh, it's, it's difficult that everybody needs to understand the chart.
• 27:37
  And so it's, it's, it's, the key is to make sure-
• 27:39
  Yeah.
• 27:39
  ...
• 27:39
  that the charts at least are, are clear and, and-
• 27:42
  Mm-hmm.
• 27:42
  ...
• 27:42
  concise and precise.
• 27:44
  Uh, but it's, it's, it's, the, the fourth leg is also extremely important, especially in Europe.
• 27:50
  100%.
• 27:51
  100%.
• 27:52
  Well, thank you, Nicholas.
• 27:53
  It's, it's-
• 27:54
  Thank you very much.
• 27:54
  ...
• 27:54
  always a joy to speak to you.
• 27:56
  And, uh, yeah-
• 27:57
  Likewise.
• 27:57
  ...
• 27:57
  please, if you, if, if, uh, if you're watching this, we'd love to hear your opinion, share some, some information, tell us what you're working on, and, uh, we will, uh, and do look into what we're doing at Monday Influencer.
• 28:08
  So, uh, it's an interesting-
• 28:10
  Really impressive stuff.
• 28:11
  Really impressive.
• 28:12
  Thank you.
• 28:13
  Thank you.
• 28:13
  I've known you for 10 years, and what you do today is like, wow.
• 28:18
  You know, it's, uh...
• 28:18
  It's coming on.
• 28:19
  It's coming on.
• 28:21
  Yeah.
• 28:21
  Thank you.
• 28:22
  You're wel- very welcome.
• 28:24
  Cheers.
• 28:25
  Thanks very much for listening to Influential Visions.
• 28:31
  Please make sure you share this episode with your friends and business connections.
• 28:35
  And don't forget to drop us a review wherever you listen.
• 28:38
  Thanks.