Welcome back to another episode of TJ\'s Technology Tuesday. In this episode, I am sharing an excerpt from a keynote I recently delivered on the topic of artificial intelligence. Enjoy.
Beginning of the keynote
McKinsey is publishing a state of AI every year. And these are the latest results where they are comparing top performing companies and less performing growing companies in terms of use AI. Both are adding up to 100%. And so when they look, the first thing you\'re always doing is cutting costs. That\'s not the focus of the top performers using AI. The focus is more on your business models. And talking about business models, I hear a lot that marketing agencies are dying. In the US, people are not waiting to die, so they are just rethinking their business model. And there\'s a growing number of solo entrepreneurs coming out of the agency world. What are they offering? Their core offering is to say, you pay me $10,000 a month, and for $10,000 a month, you get every agency job within a business day. You need new business cards, you need new paperwork, you need a presentation, you need a poster, everything you need, you get it within one business day. AI generated, but of course, curated, communicated with a pro. Because give me the best artificial intelligence tools in terms of designing something, it would be a completely different result.
And these solopreneurs are working half a day and making a revenue of 1.5 million a year. I don\'t know how your business is structured. I still have some potential to reach that number. And yesterday in the bus, we had a nice discussion to say where I talked to one of you and we were talking about the Big Four to say they are competition, yes, What\'s their big advantage? They have the resources and the processes and the standardization that you can even set an to the system, and they can do everything because they have all the systems in place. With artificial intelligence, you have all that also. You have the resources that only the big four had in the past. You have that with artificial intelligence in every of your companies. The challenge is not to wait until the Big Four have decided to do that, too. One of the main reasons why we are slow in adopting the technology is always that we say, Okay, I\'m asking ChatGPT, and the results when I\'m getting them are some like when we ask our colleague after the fifth beer to say, How is that with IDV as one?
How does that work? Probably even without any beer, you shouldn\'t ask me on IDV, for example, because I have no clue on that. That\'s a good idea to have an understanding to say, Which tool and which system are you\'re using. Coming back to Amazon, the ugliest web page on the planet. So a nice thing where you can see how artificial intelligence is used because they use it since years. For example, when you order a TV, what you always get, you get a recommendation for a rack to mount it. Who has heard about the concept of temperature in the context of AI, artificial intelligence? So now you get a special super techy tip for the next cocktail party. Because very few people know the concept of temperature. You can explain it very easily with Amazon. When you order a TV and you get just one recommendation for rack, this is temperature zero. That gives you that one most probable answer. Temperature one would give you a cup or to put it out, a blanket for the sofa, a packet of crisps, for example, and probably even a pair of A pair of sneakers? We say, A pair of sneakers?
What does have a pair of sneakers to do with the TV? Probably you don\'t dress up in sneakers when you watch TV, but probably you ordered at the same time because you said, okay, I need a new pair of sneakers. When I\'m ordering at Amazon, okay, I don\'t have to pick up things twice, so I order together. This is the reason why sometimes you\'re getting some recommendations where you say, that doesn\'t fit. Hopefully, that gives you an idea that it\'s unrealistic to expect a perfect result of artificial intelligence because it\'s statistics on steroids. Also important to understand the difference between automation and calculation and algorithms and artificial intelligence. A nice way to understand the difference is looking into a bad day that Garry Kasparov had in 1997 when he lost against Big Blue. A lot of people say this is artificial intelligence. No, that was pure computing power. Now you can play that on your watch. And talking about watches, probably some of you have realized it. I\'m the crazy guy with the analog watch and the analog ring and the digital watch and the digital ring. So I love both worlds. And when I have a look at these smartwatches, the computing power you have in these smartwatches, it extends the computing power that the Americans had when they flew to the moon in 1969 by millions of factors.
What are we doing? The same what we did with this watch since we carried the first one, just reading the time. Most of us are using technology like driving a Porsche 911 with 60 miles an hour on the slow lane on the Autobahn. My mission is to gear that up. What is artificial intelligence now? Looking into Lisa Dole, who lost against AlphaGo in 2015. Alphago is a very simple game in terms of the rules, but it\'s very complex in terms of all the probabilities. And even today, it\'s not possible. You won\'t have any Go apps where you can really play against PCs nowadays. That will change with quantum computing, but nowadays, it\'s still not possible. What did they do? They just put in the rules of going to the system and let the system play 3 million times against each other. That was the idea. And the interesting thing is there was one move of the system where every go expert said, this is a complete stupid move of the system. Every expert said, Now, Lise Doell is going to win. Two moves later, he lost. Lise Doell was very contemplated and looked at that move and said, Respect.
This is a move no human being would have made. When you watched two keynotes of Sam Altman, he said, the next generation of artificial intelligence, now we are in the transformer ones that are combining no knowledge. We say the next step, which is already in the laboratories, are inventors. And inventor systems will be able to solve technological problems and human problems that humans were not able to solve yet. This is the bright side of technology. Looking into AI, most of people say, Yeah, it\'s ChatGPT\'s AI. This is like if you would have a toolbox with just a hammer. And of course, you can use a hammer to drill a screw into a wall, but it\'s probably not the most elegant solution. It\'s probably a good idea to have a well-assorted toolbox. Probably it\'s also not a good idea to have an overwhelming toolbox where you don\'t find anything. Probably it\'s a good idea to have a curated set of toolbox.
End of the keynote
Conclusion
If you are looking for an inspiring, motivating speaker who will enthuse your team, your sales partners, and your service partners — and, most importantly, drive them into action — I would be delighted to hear from you. I am available as a keynote speaker, workshop facilitator, or in other formats for and with your clients.
When speaking with your clients, I naturally do not limit myself to the one hour of the keynote itself — I am present at the event throughout the entire day, so that your participants leave saying: “Wow, that was outstanding service.” That has been my vision since I started at Nixdorf in 1988: to bring together digitalisation, sales, and productivity — but always with energy and motivation, so that new technology does not cause anxiety but instead makes participants genuinely excited about digitalisation and AI. I would be very glad to enrich your next event in exactly that way.
Yours, Thorsten Jekel.
Key Takeaways
- Top-performing companies using AI focus primarily on new business models rather than mere cost-cutting, according to McKinsey\'s annual State of AI report.
- Solo entrepreneurs from the agency world are building $1.5 million per year businesses by offering AI-curated deliverables within one business day for a flat monthly fee.
- AI gives every company access to resources and standardisation that previously only large firms like the Big Four possessed.
- The concept of "temperature" in AI determines how creative or predictable the output is — temperature zero yields the single most probable answer, while temperature one produces more diverse and unexpected results.
- AI is best understood as statistics on steroids — expecting perfect results is unrealistic, and understanding its probabilistic nature is essential.
- The 1997 chess match between Garry Kasparov and Deep Blue was pure computing power, not true AI; the 2015 AlphaGo victory over Lee Sedol — achieved by letting the system play 3 million games against itself — is a better illustration of what AI actually is.
- The next generation of AI systems, already in laboratories, are "inventor" systems capable of solving technological and human problems that humans have not yet been able to solve.
- Relying solely on ChatGPT for AI is like working with a toolbox that contains only a hammer — a curated, well-assorted set of AI tools produces far better results.
- Most people use modern technology — smartphones, smartwatches, AI — far below its potential, similar to driving a Porsche 911 at 60 mph in the slow lane.
- The key challenge for businesses is not to wait until market leaders adopt AI but to start leveraging it now to remain competitive.
Frequently Asked Questions
What does McKinsey\'s State of AI report say about how top-performing companies use AI?
According to McKinsey\'s annual State of AI report, top-performing companies focus their AI efforts primarily on developing new business models rather than on cutting costs. In contrast, less successful companies tend to prioritise cost reduction as their main use case for AI.
How are solo entrepreneurs making $1.5 million a year using AI?
These solo entrepreneurs, many coming from traditional marketing agencies, offer clients every agency deliverable — presentations, business cards, posters, paperwork — within one business day for a flat fee of $10,000 per month. They use AI to generate the work efficiently while applying professional curation and communication, allowing them to work half a day and earn around $1.5 million in annual revenue.
How does AI level the playing field between small businesses and large firms like the Big Four?
AI provides every company with the resources, processes, and standardisation that were previously only available to large organisations like the Big Four consulting firms. Any business can now access these capabilities through AI, removing the traditional resource advantage that large firms held over smaller competitors.
What is the concept of "temperature" in artificial intelligence?
Temperature in AI controls how predictable or creative the system\'s output is, and can be illustrated with Amazon\'s recommendation engine. A temperature of zero produces the single most probable answer — such as recommending exactly one TV rack when you buy a TV — while a temperature of one produces more varied and sometimes surprising suggestions, like sneakers alongside a TV purchase.
Why should you not expect perfect results from AI?
AI is best understood as "statistics on steroids," meaning its outputs are based on probabilities rather than certainty. It is therefore unrealistic to expect a perfect result every time; understanding this probabilistic nature helps set appropriate expectations when working with AI tools.
What is the difference between Deep Blue beating Kasparov and AlphaGo beating Lee Sedol?
Deep Blue\'s victory over Garry Kasparov in 1997 was achieved through pure computing power and is not a true example of artificial intelligence. AlphaGo\'s 2015 defeat of Lee Sedol, on the other hand, is a genuine AI example: the system was given only the rules of the game of Go and then played 3 million games against itself, developing strategies — including moves that surprised every human expert — that no human player would have conceived.
What are "inventor" AI systems and why are they significant?
Inventor AI systems represent the next generation of artificial intelligence, already being developed in laboratories according to Sam Altman. Unlike current transformer-based models that combine existing knowledge, inventor systems are designed to solve technological and human problems that humans have not yet been able to solve themselves.
Why is it a mistake to rely only on ChatGPT for all AI tasks?
Relying solely on ChatGPT for AI is like having a toolbox with only a hammer — technically you can use it for many jobs, but it is not the most effective approach. A curated, well-assorted set of AI tools, each suited to specific tasks, produces significantly better results than using a single tool for everything.
How does the smartwatch analogy illustrate the underuse of technology?
Modern smartwatches contain computing power that exceeds what NASA used for the moon landing in 1969 by millions of factors, yet most people use them simply to check the time — the same function as an analog watch. This analogy illustrates how people routinely use powerful technology far below its actual potential, like driving a Porsche 911 at 60 mph in the slow lane.
What is the main reason businesses are slow to adopt AI?
One key reason for slow AI adoption is that people ask AI tools poorly framed questions and then judge the technology by the weak results they get back — described as being like asking a colleague after their fifth beer for specialist advice. Choosing the right tool for the right task and understanding how to use AI effectively are essential steps to getting genuinely useful results.
Tools & Resources Mentioned
- Artificial Intelligence (AI) – Central topic of the keynote; covers ChatGPT, AlphaGo, and next-generation inventor systems
- ChatGPT – Mentioned as the tool most people default to, but highlighted as just one tool in a broader AI toolbox
- Amazon – Used as a practical example to illustrate the AI concept of temperature in recommendation engines
- McKinsey State of AI – Annual report cited to contrast how top-performing companies versus others use AI
- AlphaGo – DeepMind\'s AI system that defeated Lee Sedol at Go in 2015, used to explain what true AI is
- Deep Blue – IBM\'s chess computer that beat Garry Kasparov in 1997, cited as an example of computing power rather than AI
