From web browser to ChatGPT: how technology suddenly goes mainstream
Let's travel back to 1993: that's when the first web browser came onto the market, and with it, understanding of the then-new internet kept growing. Before that, interest in it was reserved for tech geeks. We're now seeing a similar phenomenon with ChatGPT. Machine learning technology, misleadingly rebranded as artificial intelligence (AI), has been around for ages, but for the most part only geeks were interested in it.
That comparison was made by John Naughton, Professor of the Public Understanding of Technology at the Open University. He's also the author of "From Gutenberg to Zuckerberg: What You Really Need to Know About the Internet".
In this post, we take a look at:
· ChatGPT's breakthrough,
· other AI models besides ChatGPT,
· how the breakthrough came about
· and what could happen next
ChatGPT's breakthrough
In The Guardian, Naughton describes the panic that gripped the tech giants Meta, Google and Amazon when they realized that rival Microsoft was overtaking them in AI.
OpenAI had trained a model called ChatGPT that interacts in a conversational way. The dialog format made it possible to answer follow-up questions, admit its mistakes, challenge incorrect premises and reject inappropriate requests. The chatbot that could only give a few prefabricated answers was a thing of the past. That triggered strong reactions around the world.
AI models? We've got those too!
And suddenly every one of the tech giants also had its own AI model – and not just since yesterday.
Google, for example, came up with "Bard". The technology is similar to ChatGPT. It's also meant to improve Google's own internet search so the company isn't knocked off its throne.
Meta, Facebook's parent company, naturally had something up its sleeve too: it presented its new AI software, "LLaMA", aimed at government agencies and the research community. It emphasized that LLaMA needs significantly less computing power than the competition.
China didn't want to be left behind either, and Google rival Baidu presented "Ernie". The company has been working on this AI software since 2019. According to media reports, Musk and Babuschkin had also already discussed putting together a team to advance AI research.
How did the breakthrough come about?
In the book The Nature of Technology, economist W. Brian Arthur explains that many of the biggest technological advances happen because there's a moment when a number of necessary but unrelated developments suddenly come together. That's how completely new possibilities – innovations – are created.
In the case of the generative AI the world is now obsessed with, there were four necessary components: the availability of truly massive cloud computing power; the unimaginable amounts of data the internet provides for training LLMs; significant improvements in algorithms, boosted by neural networks; and huge amounts of money from insanely profitable tech giants. So it was the combination of these four factors that led to the ChatGPT moment.
So what happens next?
If we look back to 1993 once more, it becomes clear: all these technologies, however complex they are at first, eventually become commodities. Gradually, numerous new services and products emerge that build on the original innovations.
An interesting example is Google Maps: at the beginning, despite criticism, the company invested a lot of money to develop what was then a new kind of product. And today? Every parent googles the location of a potential new daycare, we check reviews of the restaurant we want to go to in the evening, or we check out the neighborhood before viewing an apartment. Without the physical location provided by Google Maps, none of that would be possible.
Conclusion: the future is now!
So we can expect something similar to happen with generative AI. Microsoft certainly didn't invest $10 billion in OpenAI for nothing. In future, Microsoft 365 users will get a new large language model (LLM). In March, the company presented Copilot on its website. It combines the power of LLMs with the data in Microsoft Graph and optimizes it for you.
Want to learn more about the potential of AI chatbots for your company? We're here for you! 🤖
