S3:E1: The hidden cost of AI

00:00:00:

00:00:02: Welcome to the ALSO AI podcast!

00:00:06: Discover the latest advancements in artificial

00:00:09: intelligence

00:00:10: through insightful discussions with industry experts.

00:00:13: Each episode will

00:00:14: bring a fresh perspective

00:00:16: on

00:00:16: AI and addresses

00:00:17: their

00:00:17: real-world challenges, opportunities & interests.

00:00:21: you as an IT reseller

00:00:22: might have or your customers could ask about.

00:00:26: Featuring experts from ALSO and

00:00:28: leading Vandors, we explore diverse insights on AI.

00:00:50: In our first episode, we're asking a very simple question that surprisingly... Very few organizations are asking today and discussing.

00:01:03: If AI promises to make us more productive what happens when this productivity becomes dependency?

00:01:09: And What other hidden costs of AI as a result?

00:01:13: Jake and Federico welcome!

00:01:14: Thank you Morgan.

00:01:15: Hi my name is Jake.

00:01:16: I'm here on behalf of HP and now lead the Amplify AI program at HP.

00:01:25: Our charter is to help drive AIPC sales in Markershire, we do this by working with partners like also really excited here today to engage in dialogue on AI.

00:01:37: We certainly all.

00:01:38: so thank you Jake welcome.

00:01:39: Hi, my name is Federico Ascoli and I'm head of platforms here also.

00:01:44: And responsible globally for our platform so heavy user of AI which we try always to think about how our partners can leverage AI through all platforms.

00:01:55: So yeah looking forward today's discussion very excited be here

00:01:59: Likewise here.

00:02:00: welcome Federico please to have you.

00:02:02: Also maybe break the ice before we jump into deep end.

00:02:07: let say Maybe you can both answer this question.

00:02:10: What AI task do you personally use almost every day?

00:02:14: So for me, I look back at my journey is really a personal experience.

00:02:21: each person uses it differently because each person has different tasks and job functions.

00:02:29: When I started my journey about two years ago, it was a lot of prompting information retrieval access organization type tasks.

00:02:38: And recently in the last six months we've really progressed beyond that into automation.

00:02:46: so trying to make AI work more effectively and efficiently by by offloading tasks and making them basically automatic.

00:02:58: I think i've focused most of my energy, it's not a single task that is more about the function of tasks

00:03:03: than

00:03:03: what we're working on.

00:03:05: So time-saving you are saying Jake?

00:03:09: Enhanced productivity in reducing the amount of time spent

00:03:13: on

00:03:14: things which add less value so I can spend more time adding value to my organization.

00:03:22: What you're saying, do you think it's saved your time and in the same time also improved the quality of your work?

00:03:30: Absolutely.

00:03:31: I think there are two parts to it.

00:03:32: first is that certain things It does really well for as far as improving like qualitative work being more accurate reducing on a time it takes to get information and organize it.

00:03:45: And then There was this idea of automating tasks.

00:03:48: so those tests but i used that I don't do anymore because there's a more efficient way to do them.

00:03:55: So, thats where i think you really supercharge productivity.

00:03:59: yeah um...I would see that!

00:04:00: I'm interested to know what Federico thinks?

00:04:04: Yeah..um....i guess my journey with AI has been similar too to Jake's .

00:04:09: I think it started me with AI as being mostly personal assistant like retrieving data researching and mostly I think your partner, right?

00:04:21: So starting from a non-structured idea to try and have this sparring partner build these structures.

00:04:29: Then it moved more.

00:04:31: what the hell like my toughest critics.

00:04:33: so trying to improve the quality of my work by using AI to challenge my own words, my own thoughts and my way of building content.

00:04:43: But then it moved necessarily to being part of my operating model, part in my workflow.

00:04:50: To really speed up how I'm working basically and that is where you know... It's getting fun so as to get the speed of AI but also think about how you can leverage it to be qualitative.

00:05:11: That is where I believe to put stress on one important point, which it's a personalized journey.

00:05:21: The value and the quality that you get out of it apart from spitting up your workflow really being intentional along how we use it and what you want out of this?

00:05:32: And that is where really you can unlock the value out there.

00:05:36: Maybe from HPs perspective do you believe dependent on AI faster than we realize?

00:05:45: Absolutely, I feel that any tool has come and sort of evolved.

00:05:53: So you think go back to the invention of the internet or electricity... You quickly become dependent as behaviors change.

00:06:02: but this behavioral changes takes time.

00:06:06: how quickly people adopt.

00:06:08: And so once again, I look back historically just in...I feel like AI.

00:06:13: it's dog years right?

00:06:14: It is so fast.

00:06:16: So eighteen months ago when we approached partners and customers i really felt that the adoption curve was actually very low.

00:06:27: We even see this internally at HP When first deployed within the own company.

00:06:33: What we realized is, we deployed it and what we found was that usage was actually very low.

00:06:39: And thats because there's a lot of awareness in training that is required to fully extract value of AI.

00:06:48: How long ago was this?

00:06:50: This was about eighteen to twenty four months ago.

00:06:53: It's right around the launch of the next gen AIPC hardware, and at that time Microsoft is really pushing co-pilot chat GBT making waves in new cycles And it was an era.

00:07:08: a lot of AI noise almost.

00:07:13: I think there were lots anxiety corporations because they wanted to jump on the AI train and extract this promised value, right?

00:07:26: So what we saw initially was a lot of demand in excitement but just uncertainty on how to unlock these gains.

00:07:34: And it's taken almost twelve or eighteen months for companies that have identified use case and then capability framework, right?

00:07:46: Which is really how you unlock the value.

00:07:50: And I think once you've done that You can start to get these gains.

00:07:56: Once it's hard to live without them.

00:07:58: so they become dependent.

00:08:00: Maybe a small debate for both of you.

00:08:02: Do you think dependency is necessarily bad thing in this case?

00:08:06: What do you think Jake?

00:08:07: So I look at dependency as function of benefit or risk.

00:08:13: I look at those spectrums.

00:08:18: Here's an analogy, if you have a car when cars were invented... and or use mass transportation like a train to get to work, or a car.

00:08:27: So when the car in trains were invented it actually really accelerated the ability of humans too move between places very efficiently right.

00:08:38: so you become dependent on Transportation for Work Like To Get Your Work Wherever Your Work Is You Can Move farther.

00:08:47: distance is much faster, right?

00:08:48: So this transportation really increased productivity and increased the size of cities like where people live.

00:08:57: Now if that transportation is reliable then it's okay to be dependent on it.

00:09:03: If its reliable you can depend upon all time.

00:09:07: If your transportation broke down every single day and so you were late for work or couldn't get to where needed go, then it's a big problem.

00:09:16: Your dependency creates massive risks in the organization

00:09:19: right?

00:09:20: So as long the dependency you have reliability and quality associated with thing that your dependent on.

00:09:30: And there's a big benefit, like this huge boost in productivity or revenue cost savings time saving

00:09:40: etc.,

00:09:40: then it is very good thing.

00:09:45: but if you have to mitigate risk

00:09:51: Do you think that AI is replacing a little bit critical thinking, and we are skipping the learning opportunities altogether?

00:10:02: Or do you think it's accelerating further learning opportunities.

00:10:07: I think its in very important part of this AI equation right which i think... In very simple terms you know, more dumb right?

00:10:18: Or like are we not being less intelligent.

00:10:22: Call of spade is spade absolutely!

00:10:24: Right exactly.

00:10:25: and I think this a sticky area because if you think about...I'm older than most folks here.

00:10:35: um i remember time on my life before the internet and before cell phones And I remember memorizing phone numbers Right, so I remember memorized my parents phone numbers.

00:10:47: My grandparents and friends you know And actually at any time i knew a lot of phone numbers.

00:10:53: once the cellphone became The standard everyone had one...I cannot remember phone number anymore.

00:11:00: It's only seven digit number.

00:11:01: it is not very long number in United States.

00:11:05: So did become less smart or intelligent?

00:11:10: From a memorization standpoint, yes.

00:11:12: Like my muscle the memory function is degraded.

00:11:16: however However I would say that because I didn't have to spend time memorizing phone numbers I could use that time to do something more productive right.

00:11:25: so i can Use That Time for Critical Thinking Problem Solving etc.

00:11:29: So does

00:11:31: networking other aspects

00:11:33: exactly?

00:11:34: In your in Your work life if you Can Take Two Hours of your time that you used to spend organizing information or indexing information, or recalling information and use it for critical thinking.

00:11:48: I think it actually makes me smarter right?

00:11:51: Even though some the simple tasks that i used to do...I don't do them anymore.

00:11:55: so..i'm not as good at doing those tasks!

00:11:58: As long they reinvest that say a higher skill set or high-skill level.

00:12:05: that's more valuable, it is very good use of time.

00:12:08: Now if I were to take the extra time and not reinvested then maybe what I am degrading myself right?

00:12:15: I'm degrading

00:12:15: my knowledge levels.

00:12:17: Yeah but the argument you're forwarding here Jake because this extra time that your using... You are using it as develop other areas in business than even for yourself personally.

00:12:29: You're asking yourself maybe better questions.

00:12:33: Yeah, I would say it's the idea is... The time you are not doing something that used to do as an older skill set or a redundant skillset and then has more value, incremental value and use it in a different way then you're becoming smarter or dumber.

00:13:04: It just depends on how you use your energy and use AI.

00:13:08: my opinion

00:13:11: is that it depends on having AI is my toughest critic.

00:13:21: And I found that's been very useful because, it's not only a matter of releasing time for yourself but also how to use the AI and you can leverage AI to empower itself.

00:13:32: be even more critical thinkers.

00:13:34: so challenge yourselves really tap onto this growth mindset always evaluating or how could do better?

00:13:44: Again its way more on.

00:13:49: I

00:13:50: want to add one more thing.

00:13:53: If you look at the use of AI in education, it's a great example which is if you use AI and stop thinking altogether its'a big problem right?

00:14:05: You hear these examples where the misuse by having it do your homework for.

00:14:15: And there's also the issue of accuracy and quality, because sometimes there is hallucination.

00:14:21: AI isn't perfect.

00:14:22: it doesn't actually answer them properly.

00:14:24: you see examples in where people submit their assignments and glaring mistakes that if they just read through material would have been noticed.

00:14:34: So because of this, it's important.

00:14:37: This idea being critical of yourself or a critical of the AI you have to evaluate the responses that are your given where You have to value them?

00:14:47: The conclusions that aren't giving you can completely rely if you surrender thinking two ai.

00:14:53: there is A huge I think like system or societal issue.

00:14:58: I think it's absolutely important for a layer of thinking and processing that sits above AI to evaluate the results.

00:15:08: You have to constantly evaluate the reports,

00:15:10: yeah?

00:15:10: Absolutely before...I would like us to explore further the AI hallucination topic just to close on this.

00:15:20: do you think in this case is more question.

00:15:24: businesses should really teach us how to think with AI as opposed to teaching is.

00:15:31: How do you use it?

00:15:32: One hundred percent, I think that's definitely the biggy aspect here again its eyes tool and that's what you make out of this.

00:15:41: and there's a fuller risks on not using that correctly.

00:15:45: And then i think that the tricky part of yes that It sometimes gives The illusion thinking well done.

00:15:54: So it generates, you know a well-structured document with everything in place and the call to action that everything looks fine.

00:16:02: but then if the assumptions are wrong If data behind is wrong The contact is wrong And there's not person who knows how do what to do With that material.

00:16:13: That can be risky.

00:16:15: That's where organization, I think really need to play a role into guiding people and employees that use that responsibly on the results.

00:16:25: Also getting them into this journey lower in their risks?

00:16:30: I would actually say both are important how you use it and how you think about it or they're sort of equal parts too responsible use of AI.

00:16:43: One other kind of perspective on this, the benefit and changes is that if you can stop doing certain work they used to take a lot of your time.

00:16:55: And refocus your energy.

00:17:05: behavioral changes over the course of my career.

00:17:08: So, I mean not to date myself but i started working in the mid-nineties and I remember...

00:17:14: You're revealing your age!

00:17:15: Yes yes I'm kind of an old guy.

00:17:19: so originally when we did web design and we did UIUX design you actually had to create wireframes which is basically obsolete now at this point, but we created wireframe designs which takes a long time and is very manual.

00:17:36: It's a very manual process.

00:17:38: And then you have coders develop code that implement the UI UX right?

00:17:43: Then just recently in the

00:17:45: last

00:17:46: three months ago We used AI to create an entire mock-up of UIUX for analytics dashboard.

00:17:57: you couldn't do that even four years ago, and today instead of hiring either internal employees or paying consultants to design and build UIUX models.

00:18:13: Or even build implementations in code You can do this through AI at almost the speed of light And then you can iterate faster, you can prototype and create a lower cost.

00:18:27: It's really quite amazing, actually.

00:18:29: So the work that you used needed to be done previously.

00:18:33: if you can take that energy and the cost of it develop faster or create faster with less expense obviously creates an incredible impact on your cost basis for value creation.

00:18:53: It's a very interesting idea that you're bringing on, Jake actually.

00:18:58: It opens up the question between how we use AI and how we think with AI?

00:19:06: What are your thoughts on this specifically?

00:19:09: As I said it is relatively easy to say what they were thinking or changing their way of thinking For sure.

00:19:18: And to build on Jake's example, the UX UI I think that it is really changing for example where we built platforms and how we built products.

00:19:30: because critical thinking behind this shifting from planning and developing what do you product could look like?

00:19:39: To unlocking new ways of collaborative thinking as well.

00:19:43: so by unlocking creating prototypes very quickly We can allow conversations for customers, so that we can co-create products or platforms with our actual users.

00:19:56: That is not only the case of AI replacing who are speeding up a step in the process but reengineering whole processes where it really impacts what you think and changes the conversation on all processes.

00:20:16: It's not just the delivering value at a particular step of process, but it has been increasing as a whole.

00:20:23: Earlier in this episode we were discussing trusting blindly the output about AI which highly depends on quality of the prompt and input.

00:20:35: that leads me to discuss with both of you the issue of AI hallucinations.

00:20:43: Jake?

00:20:44: In your professional opinion, why do you think AI hallucinations are still happening?

00:20:50: Is it something to do with the AI models that you were discussing earlier.

00:20:54: I feel that quality of AI has improved.

00:20:57: let's just start by saying compared two years ago and today its staggering how much better AI is functioning.

00:21:08: But even then, it's still imperfect and that is one thing everyone who uses and deploys AI has to understand.

00:21:16: And you have to be very careful what types of processes are going to automate with the AI for this same reason because they can make mistakes right?

00:21:26: So there is a concept of hallucination which just when responses get strange or incorrect with what's appropriate or they make inappropriate decisions.

00:21:41: So you asked, why does hallucination happen?

00:21:43: There is a technical reason for why it happens and usually when the model gets confused because he cannot retain all of that information in a series of questions or requests Or there just too much information It confuses itself.

00:22:01: so Is it possible that hallucinations will disappear completely?

00:22:06: I'll be candid, i don't think it will ever go away.

00:22:10: The frequency and severity of the hallucination will decrease over time just like already has.

00:22:18: but that's another really important reason why companies need to enact policy for where AI is used.

00:22:34: And it might lead to the issue when we become really overconfident with the output, right?

00:22:41: Which leads back to human judgment.

00:22:44: and let's say that governance in ethics behind use of AI.

00:22:51: What are your thoughts on this Federico specifically?

00:22:55: Yeah I definitely think that governance is a key word here.

00:23:00: I agree that hallucinations might always happen, but it's really a matter of how we handle those and kind of balancing the risk and benefits of processes.

00:23:13: And the governance that you put around.

00:23:15: this is just very simple example.

00:23:18: so if i'm an actor or sending out phishing emails, I might be correct.

00:23:24: only one percent of the time or my AI model behind can be correct.

00:23:28: It's only one person at a time and it'd be fine with that because relatively the cost of sending you know email is very low.

00:23:34: And then if we had this really high on the contrary If uh i don't know um more in the banking sector taking a one-percent risk on specific process and having an hallucinating not specific process could have huge consequences.

00:23:50: So it's also how we handle these hallucinations in the processes, the governance around that.

00:23:57: But the governance also depends on what do I want to actually achieve with AI?

00:24:02: And would this risk and benefits... How do I balance those policies behind

00:24:08: them?

00:24:08: It does!

00:24:09: This is a perfect transition to address AI as co-pilot not an autopilot.

00:24:18: there are reasons why airline pilots have a copilot, right?

00:24:23: Maybe you both want to give some examples about how we can leverage AI as reliable co-pilots.

00:24:34: What would the different uses and business applications specifically be Jake?

00:24:39: So I think the perfect framework for evaluating when to use it.

00:24:44: using as copilot versus autopilot is the criticality of the process, right?

00:24:51: So if you think about decision-making for AI.

00:24:56: A very simple decision could be to accept or reject a meeting.

00:25:02: You can set up an AI to act as your personal agent monitoring your calendar and accept meetings.

00:25:11: And If you accepted or rejected you know, but it made a mistake.

00:25:16: It probably is in the end of world because you could manually go back and fix it.

00:25:20: right now let's look at order processing for your customer.

00:25:24: so someone places an order For hundred PCs And the AI accidentally adds zero Or cancels the order or rejects.

00:25:38: That has

00:25:41: a revenue impact to your company's bottom line or in HR like employee terminations.

00:25:48: So just think of any task where there is a severe outcome if something, if a mistake is made and those circumstances you need being to monitor.

00:26:03: so you could set up AI, but you can't rely on it too.

00:26:07: To completely perform process without oversight right?

00:26:10: So there could be processes where use AI, But you have human oversight on top of it and then over time after you have a track record off performance You could consider reducing the amount of human overhead, So anything

00:26:31: related to automation and productivity around repetitive tasks, basically?

00:26:39: Because the risk does increase if we give AI more output to deal with when a human should be owning that decision.

00:26:48: Absolutely!

00:26:51: Federico what are your thoughts on this about what humans must own how AI can support us as co-pilots?

00:26:59: Yeah, for sure.

00:27:00: AI should definitely own repetitive work.

00:27:04: Things like information retrieval or even helping out with research.

00:27:09: but yeah I agree that the most important thing is that humans should anyways own the results of it so either ones who set up the processes and critical resources that need oversight.

00:27:23: So it's like recently I heard this nice example that right now with AI, we're becoming head chefs and the AI agents are our sous-chefs.

00:27:34: We were anyways responsible of the food that you get to customers.

00:27:38: The sous chef can execute but its on the head chefs too kind.

00:27:41: really.

00:27:42: set ingredients recipes then ensure quality what is being generated as actually good for a customer And I think that's kind of the key.

00:27:53: So AI can take on a lot of repetitive tasks, even content generation task but important thing is that humans own it and they have proper governance around them.

00:28:05: Yeah

00:28:05: i think that summarizes really well this question about using and leveraging AI as co-pilot.

00:28:12: let's talk little bit about productivity some more.

00:28:17: a good question actually, so one that I've been anticipating in this episode.

00:28:21: If every organization becomes i don't know thirty percent more productive how do we measure the impact of this and who actually gains?

00:28:34: uh competitive advantage?

00:28:37: Jake what do you think?

00:28:38: okay

00:28:39: well i would answer this in two parts.

00:28:41: The first part is time, which is first mover's advantage.

00:28:46: So obviously whoever adopts earlier gets a compound effect sort of like compound interest.

00:28:54: so they get the benefits earlier and assuming that the curve continues to increase and it doesn't plateau you get the First Mover Advantage Effect right?

00:29:06: And additionally there's this question of, okay let's pretend we're now past.

00:29:14: you know time has elapsed and everybody is on the same systems.

00:29:19: And they have the same benefits.

00:29:21: then the question becomes utilization of that productivity right?

00:29:26: So if you are thirty percent more productive and just decide I'm not going to do anything with this extra productivity say go home early every day from work.

00:29:39: Well, okay so then there's no competitive advantage right?

00:29:44: But if you take that it depends on how you utilize that thirty percent extra productivity.

00:29:50: what do you do with that?

00:29:51: and If can create some type of force multiplier or scaling effect Then its really about which companies you know, continuously adapt and magnify that extra productivity.

00:30:05: And where is it focused?

00:30:06: So I would say that It really becomes sort of a secondary equation as to how your using the incremental enhancement That really leads to competitive advantage

00:30:20: Especially with that context.

00:30:23: even more important The decisions that companies make so faster You can make sure strategic bets.

00:30:30: it's even more important what to use.

00:30:32: that you know, the remaining time that AI has relived for.

00:30:37: So basically its more important like make the right choices fast as the word is moving faster.

00:30:45: so definitely agree with that and I think actually puts a little bit of pressure on making the right bets and failing early but failing fast in building on that getting that exponential benefit out of this.

00:31:01: Yeah,

00:31:02: and this thirty percent gain of productivity becomes actually a motivation to differentiate from the competition.

00:31:09: right as we own A lot more decision-making power over This technology.

00:31:18: It all basically boils down To one idea that I think is the core denominator Of both your answers Federico and Jake Businesses and by extension people need AI they can trust.

00:31:36: Jake, maybe from HP's perspective how have you as a vendor changed that conversation on the device side of things?

00:31:46: Yeah, Marguerite it is great point.

00:31:48: one thing foundationally with HP this idea security, right?

00:31:54: This is a security manageability for endpoint devices.

00:31:59: And if you look at AI there are different places where AI's performed.

00:32:05: so you have AI that happens in the cloud.

00:32:08: You have AI That Happens on A Device and one things we've seen really clearly it not just One or The Other It Needs Both.

00:32:22: There are really good reasons to process AI locally on a device, and those include security offline processing.

00:32:32: You know there are times when you're traveling without internet or in place that doesn't have Internet and still need work done right?

00:32:40: So you'll need access to agents into AI.

00:32:45: so these different benefits?

00:32:49: can the models be as complex?

00:32:51: Absolutely not, right?

00:32:52: They cannot be as complex.

00:32:54: There's also certain functions that are performed better when you're using a local AI to offload

00:33:02: tasks

00:33:03: so that you can reduce your token costs.

00:33:06: As an example and there is high cost to AI compute in the cloud.

00:33:11: if I mean If you just look at OpenAI valuation they making alot of money because their charging for those services And people forget that when you make a request... That

00:33:21: data needs to go somewhere.

00:33:22: Yeah,

00:33:23: there's a cost of that request right?

00:33:25: So and when you process locally on the device those costs don't exist.

00:33:30: I mean they're basically just a function of your sunk cost of your hardware in energy.

00:33:36: so There is a cost advantage for certain types of processing.

00:33:41: There are efficiency gains by offloading some this AI onto local devices.

00:33:49: And there's this idea of data privacy and data security which can only happen when information never leaves a device, right?

00:33:59: So to us HP it is not one or the other.

00:34:03: It just what is appropriate in where if you optimize environment for both cloud AI compute and local AI compute

00:34:16: Security and privacy, especially in Europe is a very important subject that matters to a lot of people actually.

00:34:24: So what you're saying when it's all handled locally reduces the risk And maybe also at the same time security threats are becoming more intelligent.

00:34:37: so we need technology behind them To support these threats right?

00:34:44: Absolutely.

00:34:45: And Federico, maybe from Alzo's perspective do you see partners asking more and for us to onboard them on their AI journey?

00:34:57: Yeah.

00:34:57: For sure.

00:34:57: I mean that is definitely what we are saying there.

00:35:00: a lot of interest around then.

00:35:03: i think it was really exciting the fact that were moving from tool conversation or product conversations I said, an outcome conversation.

00:35:12: So it's way more common to hear from partners.

00:35:16: how do we solve that problem with the AI rather than what is best AI tool?

00:35:21: That I can import and as Jake was referring too there are infinite amount of options on using AI depending upon these specific problems that you want to solve.

00:35:34: So from how to decide between on-cloud or local deployments, for example as you mentioned security and compliance.

00:35:47: How do I make sure that my AI deployments are EU Act compliant?

00:35:52: These are the kind of conversations we're seeing right now in the market but actually trying to get those conversation with partners because again it's more about the matter of, you know, serving a problem.

00:36:04: Improving that process and leveraging the ecosystem we have both on hardware side or service side to really unlock the benefits of AI.

00:36:18: so yeah... We're seeing lots of interest there in many questions but they are good type of questions right?

00:36:24: So how do I embed this?

00:36:28: AI products and tools to really solve an end customer problem.

00:36:32: That's

00:36:34: the way it

00:36:34: gets us

00:36:35: there?

00:36:35: Thank you very much both, I would like to introduce a final section maybe to round off our discussions from this episode.

00:36:45: It is

00:36:45: called

00:36:46: myth or reality a question you will answer by myth or reality and maybe bring on a short explanation about this.

00:36:57: Jake, and Federico are you ready?

00:36:59: Ready!

00:37:00: Ready oh wow okay uh Myth or Reality.

00:37:04: AI Will Replace Programmers.

00:37:06: I would say it's myths.

00:37:07: so programmers developers will not disappear.

00:37:10: they will mostly change the role from writing lines of codes to really being builders.

00:37:17: I

00:37:19: agree with Federico, it's a myth.

00:37:21: The functions will change and the job description will change but they still exist.

00:37:27: you've already seen that we have prompt engineers.

00:37:30: now you'll have agent engineer or automation engineer so

00:37:37: maybe create new ones as well.

00:37:38: if i can add my two cents myth or reality AI always saves money.

00:37:45: definitely It could actually cost you a lot based on the frequency and use of your requests.

00:37:54: And also if it can cause problems, so

00:37:57: I agree with Smith's.

00:37:59: I mean he can save time but that doesn't mean at this time that you're saving is really adding financial value to your organization

00:38:06: OK?

00:38:08: Myth or reality AI is making us less intelligent.

00:38:13: Could be both, so it could.

00:38:15: Be a reality or meets depending on how you use it and that's why some also important to have the governance layer in trainings around it.

00:38:24: I

00:38:24: agree It's definitely both makes you forgetful Or less capable of certain tasks but You can take That same energy And that brain power and learn A new skill or New capability.

00:38:35: So it does Both.

00:38:36: myth or reality.

00:38:37: AI should never make final decisions.

00:38:40: So I would say it really depends on the risk.

00:38:43: It could be a myth in reality, depending how much risk the company is willing to take under specific process.

00:38:50: but ultimately whatever we decide that AI should decide on... ...it shouldn't be on.

00:38:58: I agree with the type of decisions making But if i had to categorize as whole today, but I think in the very near future you can rely on it to make certain decisions.

00:39:13: Let's just say

00:39:13: that way It comes back to the hallucinations doesn't it?

00:39:17: Myth or reality?

00:39:18: In five years AI will feel as normal As email is a big one

00:39:23: i would Say reality.

00:39:25: But there Is A lot of different ways To Think About AI and use cases Of AI.

00:39:31: so i Would Save Reality but it Will Move Too Something That Might Not

00:39:36: I'll say reality and it won't take five years, in two more years.

00:39:40: It will just be a standard business process.

00:39:43: Jake and Federico would really like to thank you for joining me on today's episode of the Alzo AI Podcasts.

00:39:49: We found our discussions thought-provoking so it was truly a pleasure to pick your brains during this session.

00:39:56: Thank You Morgan!

00:39:56: For the opportunity to come on this podcast.

00:39:59: If i can leave your audience with one final comment which is roll up your sleeves have fun.

00:40:07: Just explore AI, use it as much as you can because that curve starts when you start using it daily and the productivity you get out of is really incredible and amazing.

00:40:23: so thank

00:40:29: Thank you.

00:40:34: It's been great, and it is good to have this kind of discussions with these enthusiasts like Jaquie and Jake that are always enriching.

00:40:46: so it has been great.

00:40:47: thanks a lot!

00:40:48: Thank You And This concludes season three episode one of the Arzo AI podcast.

00:40:54: thank-you for listening in.

00:40:56: stay tuned as we explore the competitive side with different guest speakers this time.

00:41:03: If every organization has access to these technology, where does the competitive advantage actually come from?

00:41:10: This was Morgan Galan-Jones.

00:41:12: Thank you again!

00:41:30: For more resources, links and product details discussed today.

00:41:33: don't forget to check out our show notes and visit our podcast website.