Simplifiers podcast - Season 1 - Episode 3

Episode transcript:

Intro: Cutting Through the AI Noise

[Amel]
Today we're diving into one of the most talked about and often misunderstood topics in the business world: artificial intelligence. AI is already shaping how we work, sell, hire, and make decisions, but it's also wrapped in hype, jargon, and massive expectations. That's why we're here-to cut through the noise and explore how AI can actually deliver business value when it's approached with simplicity, clarity, and responsibility.

Joining me today are two brilliant minds. First, Gabriel dos Reis Estivalet, a data scientist at Solteq who works with organizations to turn data into smart, usable systems-not just dashboards or black boxes. Also, Mikko Alasaarela, an internationally recognized AI entrepreneur, keynote speaker, and thought leader on algorithmic influence and ethical tech.

Together, we're going to unpack what real-world AI success looks like-and where it often goes wrong. Let's jump in. Welcome, Gabriel

[Gabriel]
Thank you.

[Amel]
Welcome, Mikko. I want to get started with both of your journeys into AI and digital. Gabriel, let's start with you. What led you into this field and how did you end up in Finland working at the intersection of data, engineering, and business advisory?

Gabriel’s Journey: From Law to Data Science

[Gabriel]
So, I started my career actually in law. I went to law school in my country of origin, Brazil. I worked for a couple of years in the field, but I wasn't entirely satisfied with how fast-or rather how slow-the whole system worked. I decided to branch out and expand a little bit more.

Initially, I had good intentions to continue with law and then take on some business-related studies. But I eventually landed in San Francisco, California, and got a big exposure to a whole new world of technology and data-something I didn’t really know before. That was around 2014–2015. Since then, I’ve been shaping my career and studies in this new direction.

I’ve worked in different companies and industries, and now I’m in my third country-hopefully to stay and make Finland my home. I started as a data analyst, moved into data science, and since moving to Finland about six years ago, I’ve been increasingly exposed to data engineering.

Since joining Solteq and its advisory team, I’ve continued this work. The team is made up of people from different roles, working together to address the real challenges our customers face.

Simplifying Data for Impact

[Amel]
You've talked a lot about simplifying data for impact, not complexity. What does that mean in practice, from your point of view?

[Gabriel]
In many cases, it's actually easier to come up with something complicated and maybe even overpromise. But when you're trying to solve a business challenge-something specific, something that’s not optimal right now-often a simpler solution should be the first option explored.

It doesn’t necessarily mean jumping into the AI hype. Sometimes a simple fix is enough. We shouldn’t use AI just for the sake of it-it’s a great tool, but it's not always the right first step.

Mikko’s Path: From Games to Ethics in AI

[Amel]
Mikko, your career has taken you from San Francisco to Berlin, from startups to global stages. What keeps you so deeply invested in AI, and particularly in ethical, responsible use?

[Mikko]
It all started back in 2006–2007. I was working in the game industry. I’ve been a founder in multiple companies and an angel investor as well.

Over time, I’ve worked in many areas of tech-from nano-optics to mobile games to blockchain to AI. But AI really caught my interest in San Francisco when I got into behavioral algorithms-huge at the time due to the rise of social networking.

I was hanging out with people like Mark Zuckerberg and others in the social gaming space. Through those conversations, I realized behavioral algorithms were going to be massive. The game industry shifted from fun time-fillers like Angry Birds to creating algorithmic addictions in free-to-play games.

That direction didn’t sit right with me. So I left the industry and focused on using behavioral algorithms ethically-for the good of humanity. And looking at the rise of mental health issues and societal polarization now, I feel it’s more important than ever to be a counterforce.

Why Ethical AI Became His Mission

[Amel]
Would you say that's even your mission?

[Mikko]
Absolutely. That’s been my mission since I pivoted to this path. I’ve been working in AI for over 15 years, long before the ChatGPT hype. I built my first AI-powered chatbot in 2014-so I’ve been around.

When AI Goes Wrong

[Amel]
Let’s talk about where AI goes wrong. Maybe it’s a very Finnish way to start with the problems! But failed pilots, inflated expectations, and expensive learning curves are everywhere. Mikko, you’ve talked a lot about the imminent arrival of AGI-artificial general intelligence. What does it mean for us and our societies?

[Mikko]
Both Sam Altman and Dario Amodei have said we’ll have AGI by early 2027-that's only about 19 months away. To me, it’s like someone saying, "Jesus is coming-are you ready?" The shift will be that big.

As AGI approaches, everything accelerates. It feels like being in a rocket ship. The pace is overwhelming, and I don’t think society or individuals are ready. That kind of change, with an unprepared population, can be dangerous.

[Gabriel]
Yeah, it’s not easy. It’s a radical shift in how we work and live. Even for people in the field, it can feel overwhelming. New models and technologies are coming so fast.

[Mikko]
Exactly. I work closely with top AI hackers. I’ve seen ChatGPT’s price-performance ratio increase 150x from early 2023 to mid-2024. Now we’re seeing 10x performance improvements yearly. That’s faster than Moore’s Law.

Meanwhile, Finnish software companies tell me only 4–10% of their projects involve AI. That’s shocking to me. It should be 95%. At this pace, nothing else matters.

The Foundations Companies Miss

[Amel]
Gabriel, what do companies typically get wrong when they start their AI journey?

[Gabriel]
Many try to skip steps. These systems require structured data and foundations, but often data is in silos. Some companies lack business-side involvement, so it becomes a purely technical effort.

That disconnect is risky. These models are meant to impact the real world-operations, customers, outcomes. Optimizing for a technical metric isn’t the same as creating business value. Business must lead the initiative.

[Amel]
And not over-complexify.

[Gabriel]
Exactly. Simplicity works. If you can solve a problem simply, don’t introduce complexity. Include people early, make them part of the journey. It reduces friction.

Chasing AI for the Right Reasons

[Amel]
Speaking of humans-Mikko, do you think people are chasing AI for the wrong reasons?

[Mikko]
AI has definitely been used for the wrong purposes throughout its history. But now, people are starting to understand what it's actually capable of. Many who used to think of it as just a chatbot now see the potential of agents that can amplify their work.

This realization is changing mindsets. But not everyone is on board yet. Too many people are focused on incremental improvements-like a 10% or 30% boost using ChatGPT for search or summarization. That’s a very low ambition.

I’m helping companies build AI-native strategies. We set ambition levels to at least 3x. That’s where you start seeing true transformation. One Canadian friend of mine shipped 10 million lines of code last year by himself using an agent farm. That’s like the output of a whole mid-sized development agency. Last month alone, he did 2.5 million lines. That’s the rocket ship we’re in.

[Gabriel]
Do you think this is creating a split between those jumping into the rocket ship and those hesitant or lagging behind?

[Mikko]
Absolutely. The power law is extremely steep right now. Most people aren’t even using AI daily. Maybe 10% are-getting those incremental gains. A tiny fraction, maybe 1%, have figured out how to use agents for 2–3x productivity. An even smaller group is operating at 10x or 1000x productivity.

It’s worrying. If this continues, AI will become an accelerant of infinite power for a very small group. We have to prepare society for this and figure out how to share the benefits equitably. That’s why I’ve committed to building technologies that address this imbalance.

Debunking AI Myths

[Amel]
What’s a strong myth about AI you’d like to debunk for the general public?

[Mikko]
One trendy myth is that AI is going to consume all the energy on the planet. But let’s look at the facts. ChatGPT’s cost-performance ratio improved 150x in 18 months, and it's now improving at a 10x yearly pace.

You can’t get those kinds of gains unless energy efficiency also improves. If energy consumption per task didn’t go down, the models would become financially unsustainable. So no, AI isn’t going to suck the planet dry. Most of the energy in AI will be used in inference, not training, and inference is getting cheaper all the time.

[Gabriel]
I see this too when talking to customers. There's a lot of hesitation-people see AI as a bogeyman coming to take jobs and automate everything.

Yes, some roles might phase out, but others will emerge. When I started, data scientists were expected to do everything-analysis, modeling, deployment. But now we have more specialized roles: machine learning engineers, AI engineers, prompt engineers. So it’s not all doom. The mindset just needs to shift toward continuous learning.

[Mikko]
I’ll take a bit of a counterpoint. I think many of the worries are justified. We’re on a vertical trajectory with AI, but human evolution is nearly stagnant. In fact, average IQs have declined slightly over the past decades.

We’re heading toward artificial superintelligence-1000x more capable than humans. What happens to labor when AI outperforms us across the board? I think we’re moving toward a post-labor society, and we need to prepare for that.

[Gabriel]
True, if left unchecked, AI could take over more and more. But people adapt slowly. If everyone becomes replaceable, who’s left to buy what companies are selling?

[Mikko]
That’s why I say: get yourself to the forefront. Anyone listening to this podcast-don’t wait. Learn fast. The ones who adapt early will reap the rewards.

What Does Good AI Look Like?

[Amel]
Let’s talk about good AI-from the listener’s perspective. Gabriel, can you share a success story where AI made a measurable difference-not because it was flashy, but because it was smart and simple?

[Gabriel]
Yes. We had a customer for whom we built a portal where their end customers could access services, transactions, and information-all in one place.

Nowadays, most data is digital, which is easy to manage. But they also had a paper trail going back years. Manually digitizing it wasn’t feasible. And because many of those forms were handwritten, a traditional program wouldn’t handle it well.

We used a vision AI model to read and extract data from the documents. It wasn’t a massive, expensive project. It was simple, quick to implement, and accurate enough to be useful. Not sexy, but impactful.

Ethics Meets Business Value

[Amel]
Mikko, you often emphasize human-centered AI. What does that look like in a business context?

[Mikko]
We live in a world where algorithmic behavior-shaping is everywhere. Many players are trying to influence our decisions-some even call it mind control.

But we can use algorithms for good. Take Aura, the Finnish health tech company. Their ring collects biomarker data and uses algorithms to synthesize that into a simple sleep score. That’s incredibly impactful.

I talked to their founder, and he doesn’t see Aura as a hardware company-but as a healthy algorithm company. That framing makes sense. We could do so much more if regulation allowed broader use in health care. We already have tech that’s as good as doctors in some areas, but it’s being blocked to protect outdated systems.

Underrated Benefits of AI

[Amel]
Many benefits of AI have been discussed already. But what are some of the most underrated ones?

[Gabriel]
For me, it's the human connection that happens during implementation. Tech and business teams come together, complementing each other. You learn about other departments, build bridges across silos. That collaboration can be very fulfilling-if done right.

[Mikko]
One underrated benefit is the multilingual capability of language models. They can speak every major language-and many minor ones. You can use them in Finnish, Hindi, whatever-and they still work. It’s a remarkable outcome of how these models were trained.

Cultural and Psychological Bottlenecks

[Amel]
Let’s talk more about the cultural and psychological side. You’ve both mentioned that tech is rarely the bottleneck. So what is?

[Mikko]
When I work with company or government leaders on building AI-native organizations, we often talk about expectations that have moved beyond human speed.

Take customer service-people expect instant replies now. No human can satisfy that anymore. So we need to redesign organizations. Assign tasks requiring instant responses to AI, and let humans focus on slower, more complex, more empathetic work.

You stop asking what tools can support your people and start asking what tasks are better for agents, and which ones still require people

[Gabriel]
Exactly. People and organizations have to shift focus toward where they still hold strength-like empathy and emotional intelligence. That’s not something AI can truly replicate... at least not yet.

[Mikko]
I actually disagree. I used to believe emotional intelligence was a uniquely human strength. But after testing emotionally intelligent AI early in my career, I found that humans-on average-score pretty low. AI can outperform that baseline.

I ran experiments across hundreds of thousands of people online. The ease with which people can be emotionally manipulated was eye-opening. So yes, AI can be emotionally intelligent too.

Still, human connection matters. Sharing space with other people-talking, exchanging-that has intrinsic value beyond logic or performance.

[Gabriel]
True. Like with art or music-AI can create it, but you miss the connection to the creator. You’re not likely to idolize an AI musician the way you would a human artist.

[Mikko]
Let’s look at some numbers. At the start of 2024, 57% of all internet content was AI-generated. Now it’s over 70%. By early next year, it’ll be more than 90%.

Last July, bot traffic on the internet surpassed human traffic. In Western Europe, places like Ireland and Germany are already above 70%. Even social media has followed. In October, over 50% of influencer content on LinkedIn was AI-generated. You’re probably liking posts written by bots.

That’s why I think WhatsApp is the best social network today-no bots, just people and their actual friends.

Building Trust in AI Within Organizations

[Amel]
I want to take it back to the organizational level and the human factor. What’s the best way to build confidence in AI systems across a company?

[Gabriel]
Transparency. In some cases, I’ve shown failure cases right up front-where the system could go wrong. It helps set expectations and builds trust.

If people understand the limitations as well as the benefits, they’re more likely to embrace the tech. You have to be honest. Show what it can do, and also where it might struggle.

[Amel]
Interesting.

[Gabriel]
It’s an important piece of the puzzle.

[Mikko]
Transparency also applies to the models themselves. We spend so much of our lives interacting with models-they shape our thinking and worldview. So we need open-source models that we can validate and trust. This is especially important for digital sovereignty in Europe.

Designing Responsible AI with People in Mind

[Amel]
Where do users-customers, employees-fit into responsible AI design?

[Gabriel]
I believe it's about collaboration. Everyone should be brought in early, kept informed, and engaged. You don’t want to build something just because it’s AI-powered. You want to create something that meaningfully solves the original problem.

Everyone has different perspectives based on their roles. Involving people helps catch blind spots and makes the solution more effective and accepted.

[Mikko]
Exactly. And I think we’re going to see a shift in what makes someone valuable at work.

As AI becomes more capable-more intelligent than most professionals-what sets people apart won’t be technical knowledge. It’ll be human traits: empathy, emotional presence, care.

We’ve undervalued these skills for too long. Think about nurses-they’ve been underpaid for decades, but their human role in healing is vital. I believe AI will help bring that kind of work the recognition and value it deserves.

[Gabriel]
Yes, and trends are already pointing toward soft skills becoming more essential. Companies and CEOs are starting to highlight and reward them more.

Looking Ahead: Strategic AI Trends

[Amel]
As we wrap up, let’s look ahead. What are the most important AI trends business leaders should keep an eye on-not just technically, but strategically?

[Gabriel]
The tools are only getting better. I think we can all agree this wave of AI isn’t going anywhere. Models are improving. Agents are becoming capable of running workflows we never imagined a few years ago.

New use cases are emerging constantly. AI is already embedded in most apps we use, whether we realize it or not. Companies are at different maturity levels, but no matter where you are-it’s time to start. It’s dangerous to wait.

[Mikko]
Exactly. We’ve seen many tech waves in our lifetimes, but this one is uniquely obvious.

Metaverse and blockchain had potential, but there was uncertainty. With AI agents, I haven’t met a single person who doubts they’ll be essential by 2030. If you’re ignoring agents, you’re ignoring your own future.

Excitements and Concerns

[Amel]
Second to last question. What excites you most-and what concerns you most?

[Gabriel]
I’ll start with the exciting part. I’m positive by nature. The improvements we’ve seen in just a few months are amazing. What wasn’t feasible a few years ago now works beautifully-across languages too.

Personally, I use AI to help study Finnish. It’s like having a personal tutor. That kind of accessibility is powerful.

What concerns me is how algorithms can amplify the wrong metrics-clicks, impressions-and shape behavior in harmful ways. We’re beginning to realize the negative effects of screen time, especially on kids. Even I find my screen time too high sometimes. I don’t want my children falling into that same pattern.

[Mikko]
Same for me. AGI is coming, and it can either solve almost all our problems-or be extremely destructive if mismanaged.

If only a few individuals control all the benefits, that’s a disaster. We need to rethink our societal structures. A technology that powerful should lift everyone up-not leave the majority behind.

Final Advice for the Overwhelmed

[Amel]
Last question. If someone’s feeling overwhelmed by AI, what’s your one piece of advice?

[Mikko]
That’s easy. At the start of this year, I took on more work than ever-because I could. AI scaled me. My wife was worried I’d burn out, but by spring she was asking if I even worked anymore.

I told her: that’s exactly what AI does. I delegate to my agent farm. I get more done, but I’m more relaxed than ever. So if you feel overwhelmed-don’t be. Let the agents do the heavy lifting.

[Gabriel]
I’d say: start small, but start. Focus on a real business problem, something that brings value. If you're early in the journey, don’t try to do everything at once. Build momentum. This wave is here to stay-we have to embrace the change.

Closing

[Amel]
Thank you, both of you. This has been inspiring, honest, and refreshingly human. Thanks for bringing not just your expertise, but also your heart and clarity to this conversation.

And to our listeners: if you're wondering how to move forward with AI, here’s your takeaway-
Start small. Stay clear. Focus on value. And don’t forget the human side.

This was Simplifiers, brought to you by Solteq. If you enjoyed this episode, share it with a colleague, follow the show, and stay tuned for the next one.

Next up: we’ll be diving into transformation failures-and why so many change programs miss the mark.

Until then, keep simplifying your digital world.