Technologies in 2023 and Predictions for 2024
Decoding Tech Trends: Pierre Kirisci Explores AI and Blockchain Revolution of 2023
To talk about the most amazing innovations that have been shaping the technology landscape in 2023 and the technology trends for 2024, I invited Dr. Pierre Taner Kirisci, a technology enthusiast with a record of accomplishment of more than 20 years in coordinating and managing industrial technology projects. I met with Pierre at the GITEX conference this year where he was a keynote speaker and he kindly agreed to contribute to our vlog.
Pierre:
I think everybody would agree that 2023 was all about Artificial Intelligence. I remember quite well at the end of 2022, we had a research project and we were in a meeting in Vienna. At that time, ChatGPT had just come out. Not a lot of people had heard about this, and we talked about it. Some people heard it, some people didn’t. We were all not very sure what exactly to expect from ChatGPT. But I had the notion from the very beginning that this is going to be a very big game-changer in 2023, and yes, this actually happened.
In 2023, ChatGPT became more popular. But, of course, you should know, or the audience maybe, would be interested to know: Is this really a brand-new technology that came up, or did this actually evolve from the past? Did this also have a kind of history? And if we talk about ChatGPT, this is all about the topic of so-called generative AI. Generative AI is based on imitating human conversations. It’s based on large language models which are trained. This is, of course, a big effort. We should all know that this technology, and incororated models, did not just evolve in one day or within a couple of weeks or months. This all happened within the last couple of years because of the tremendous effort to train those models.
But the actual technology we are talking about here, generative AI, and the related applications ChatGPT, Dall-E, etc. which uses text to create images., We could trace that that back to 2017. In 2017, we experienced, at least for the people who are following AI topics, that there were already some research groups dealing with large language models. And in 2022, there was a kind of breakthrough with conversational large language models which came into being.
In 2023, when this came up, everybody was very excited about ChatGPT and what ChatGPT could do. At the beginning, you actually had almost the impression that it could do anything. But I would say the average person who got hands-on with ChatGPT tried a couple of things. Of course, it could create interesting texts based on text input. It was able to create images, and all kind of data. And during the year in 2023, we’ve seen a lot of applications evolve, which have been using the GPT model. You have GPT-3, you have GPT-3.5, which was then trained with hundreds of millions of parameters. And then during the year, just recently, a couple of months ago, we had the updated model of GPT-4, which was trained with trillions of parameters. This is not only a big effort, but fundamental. And this is the reason we are seeing a lot ofinvestments. If you look at the investment landscape, the venture capital companies, they were all going crazy in 2023. Everybody wanted to invest in AI, particularly in generative AI companies that are developing applications, services, and products. In my point of view, they were overseeing a lot of other interesting technologies. I would consider this as a typical hype, which happened in 2023.
I even had the impression that companies or venture capital firms investing in startups didn’t even look into the details, likeasking those companies, “Do you already have something? Do you already have a prototype?” Lots of companies even didn’t have any MVP. They only had a concept or an idea and said, “Okay, when we get funded, when we get the money, we would actually then start developing and training our models.”
Lots of people don’t know that training these large language models, at least if you want to do it in a professional way and have an impact on the market, costs millions of euros per month. This is the reason why these fund-raising rounds are commonly large. Agood example is Mistral AI, this French company, which was funded a couple of months ago. I think they also received 100 or 150 million. I don’t know exactly.
So the first question you usually ask yourself: Why does a company need so much money in order to develop AI? I have been in this space for many years, and we have been developing AI models for at least 10 years. However, we are not really involved in generative AI, but we are doing research, and our team is actually developing machine learning models for specific applications. For these purposes, you don’t need large language models, and it’s not that cost-intensive.
But when you talk about generative AI, yes, the costs are significant. It’s also interesting to note that it’s not only companies like OpenAI in this space, but other companies are also within this space. There’s a kind of competition going on in 2023. Some, of which, are, of course, the cloud infrastructure providers. These are the companies who host those servers where your models are implemented and trained. Of course, like, for instance, Amazon Web Services would be a prominent example or Microsoft Azure or Google. On the other hand, you have this competition going on with the chip developers because you need very specific chips to do the job. Nvidia, for instance, is a good example. Nvidia provides, for instance, specific AI-empowered chips to do this task. And you have Microsoft, you have Salesforce, you have Google, you have Amazon. We’ve seen in 2023, there was really a big competition between these companies. This is still going on, and we don’t actually know exactly where this will lead, but can only estimate where this journey might be going.
Next to AI, we also have to talk about other technologies as well. For instance, blockchain technology. There was a hype in blockchain technology. I recall that was during the last couple of years, similar to what we are seeing here in AI, which was happening in 2023, was happening likewise with blockchain technology during the last couple of years. I am in the blockchain space since 2017/2018, so I’ve been following exactly what has been going on.
When we talk about blockchain, we probably also have to talk about Web3 because blockchain technology very closely connected to Web3. The hype has been going on in blockchain from in 2017, 2018, till 2022. We had this hype, but then, after a while, there were some expectations regarding blockchain technologies, regarding cryptocurrencies and decentralized ledger technology, which actually was not fulfilled the way people expected this. They’ve probably not seen sufficient applications, namely blockchain-based applications, which have been implemented in the business space in the different sectors.
At the very beginning, we always discuss this: Okay, blockchain can do a lot of things for logistics, and we can optimize logistic processes with this innovative technology. We can optimize any kind of process because we eliminate intermediaries which are usually required for certain processes to work seamlessly. We can enhance trust and so on and so forth. But this did not materialize the way people have expected this. Therefore, there was a kind of disillusion happening , and everybody turned their attention to AI, while blockchain was more or less in the background.
Although, since I’m in the space, there were still, of course, a lot of projects being built, and there’s a lot of things happening. Therefore, I said at the very beginning, that lots of people, due to this hype of AI, did not really pay attention to what was going on also in the other spaces, within other technologies. I can say there were still a lot of things going on in blockchain technology. Blockchain technology is becoming more and more mature. There were, of course, during 2023 and2022, some negative examples.
Take, for instance, the company of Sam Bankman-Fried from FTX, which went through the media, and where lots of investors lost billions of dollars.. Lots of people who were following these events thought, okay, there are a lot of projects out there based on blockchain, based on cryptocurrencies, on stable coins or whatever, we really have to be careful here. There are a lot of scams out there.
So, these kind of projects were doomed to fail, and of course, in my point of view, it’s absolutely normal, it’s often like that when a new technology evolves, and comes into being, of course, there are always these negative examples. There are always scammers out there, who want to make a profit and fool the people. This is natural, I think. We can say that this happens almost in any new technology which is introduced at the very beginning.
But what is important here are the lessons learned. What type of lessons can we take home? Because here we learn a lot only by looking at these negative examples. It gives us some very valuable input and results which we could actually use to build better technologies, to optimize technologies, to rethink technologies. I think this is part of the journey. This is something which has to happen.
And, of course, related to blockchain technology, we also have to mention Web3. There’s a lot of fuss and a lot of talk about Web3. You should know, Web3, there’s no actually really sophisticated or standardized definition on Web3. I consider Web3 as a concept or as an idea.
And when we talk about Web3, we also have to mention what exactly was Web1 and Web2. So, when the internet emerged, when the internet was introduced in the beginning of the ’90s, I can remember that quite well because at that time, I was studying at a university in Germany. I remember when the first web browsers were introduced. This is the time we’re talking about. This was Web1.
Web1 was like people had access to the internet and were only able to read the data, read the information. It was read-only. And then came Web2. I think this went on for at least a decade. Then the concept or idea of Web2 was introduced. This was when all of those big companies came into being, like Amazon, Google, of course, was there before, but Google played a significant role as well. There was Facebook, and then after that, Instagram, and so on. All those social media companies, this was Web2.
Web2 was more interactive. This means you could actually read, but you can also write. You could create things, but you did not own the data. You paid for the data. How did you pay? The currency was your own data. And that was a kind of currency. We can say that this is how we paid for using their free services, such as Facebook is free, Twitter is free, and now it’s called X, but at that time, it was still Twitter. Twitter is free, and of course, we provide data for using those services.
Now, we are going to Web3. Web3 is more about owning the data. This means we actually come back to the original vision. Maybe this was even the original idea of the internet, that we become data owners, that we take the decision on what type of data we want to consume and what type of data we want to share. And when we provide data, we can even make companies pay for this data. This means the whole thing is more about decentralization. Everything is decentralized. That’s the whole idea. But blockchain is one enabling technology of Web3. This is how I see it. It’s one enabling technology of Web3, but you also have other enabling technologies. Take, for instance, a big topic in 2023, at least for the people who are following what was happening in the blockchain space: decentralized identities.
Digital decentralized identities. How do you do KYC – Know Your Customer? How do you do anti-money laundering in a way that is more trustworthy, in a way which is easier for the customer? Here, you have concepts like decentralized identities. This means I am the owner of my identity, and I only have to register once, and it is available in a decentralized manner, and everybody can use it. That means it’s very convenient for the customer.
This is something that will also come. I can talk about this for some trends in 2024, 2025, and beyond.