It's been a while since I specifically delved into generative AI beyond my LinkedIn posts discussing the importance of incorporating varying degrees of knowledge or systems that provide references for the "raw" generation of a Large Language Model (LLM). See: What is an LLM (Large Language Model)? We're aware that large language models only "know" as far as the training data allows, so the race is on to continually update and expand their capabilities.
The impact of generative AI on business and society dominated the global conversation in 2023 and continues to do so in 2024. To support sustainable and inclusive growth, organizations must not only adapt to the digital age and refine their approach to talent and inclusion but also integrate this technology as a critical component of their communication strategy. They should utilize their repositories in a way that allows data and customers to effectively communicate with both internal and external users.
In some instances, entire companies may need transformation, especially as the urgency of the energy transition becomes more pronounced. Leadership and resilience are now more crucial than ever.
One of the most notable pieces of news from the past week is OpenAI's release of its "store." While for Sam Altman, his wedding was equally significant, both events hold their importance. A wedding, however, isn't typically a topic for a technology blog. Yet, there are notable aspects: the intimate guest list, symbolizing commitment and stability, which is always a cause for celebration. This follows the success of his app, reaching 100 million downloads in a month, and high-level meetings with world leaders and U.S. government representatives due to the technology's impact.
Four images from Sam Altman's wedding in an unknown location
OpenAI is a technology project created by Sam Altman, which combines GPT, a linguistic model, and LLM, an AI technology that generates new ideas from a given idea. Its mission is to create an AGI (see What is Artificial Intelligence (and AGI)). But before it gets to that goal, OpenAI now says it wants to offer consumers access to an AI that is more powerful than ChatGPT, because sometimes there can be limitations to asking questions of the model. For example, if one is asking ChatGPT questions about the weather tomorrow in their location or another region, it may not be very accurate and may not have answers to all questions. To avoid those problems, OpenAI wants to make the model more accessible through its GPT Store with specialized models for specific tasks.
The advantages, on paper, seem innocuous: OpenAI seeks an AI that is accessible and open to competition, allowing anyone to make use of it through a marketplace. Why does it do that? Because there are companies like Mistral that compete for the GPT license and have a business model based on licenses in exchange for money.
To compete with these companies, OpenAI can offer these companies free licenses or even offer its services in exchange for participation in their models, which would be a step towards building an ecosystem that offers these companies alternatives to the GPT license. In case there are problems for OpenAI, OpenAI can offer other companies related services, e.g. cloud-based services in exchange for participation in their models.
This is why there is so much talk about the potential GPT ecosystem and why OpenAI needs to offer its competitors an alternative to its licenses: OpenAI can offer its competitors services in exchange for participation in its models, offering these companies alternatives to its licenses.
The GPT Store is the closest thing we have known so far to the concept of successful app stores, such as the App Store and the Play Store. Once the app ecosystem is created, some die and others survive. Some make money, but the one who always wins is the store owner. The products, well, that's a little bit the same: they depend on the popularity in the market and the work those companies do to create, their marketing, if they really solve problems or not... Some of the GPTs I've tried are really bad: basically they are little more than refined prompts thrown out by a student hoping to create community and investment.
In a move similar to the other two stores (Android / MacOS). OpenAI has also announced that it is going to start a revenue sharing program with the creators of the GPTs, without offering many more details yet.
GPTs entail making a technology available to everyone so that everyone can create their own AI systems.
On the one hand, you leverage a very powerful technology for your own use. And on a global level, OpenAI is going to take advantage of the depth of use and creativity of the community, of anyone who knows how to set up a GPT that imaginatively writes business emails or tells stories. This is far from the mission of creating an AGI, but very close to the commercial vision of the current CEO of OpenAI. This explains the reasons for his defenestration and triumphant return to his post, hand in hand with Microsoft. Had Sam Altman not returned as CEO to OpenAI, Satya Nadella would have kept him at Microsoft. Since February 2014, Satya Nadella has been at the helm of Microsoft. Under his leadership, the company has achieved significant milestones, including increasing market capitalization and strengthening its presence in cloud computing, particularly with Azure. He has been responsible for several strategic acquisitions, such as LinkedIn and GitHub, as well as strategic investments in areas such as mixed reality and artificial intelligence (OpenAI among them). Satya Nadella took a company that sold licenses and brought everything to the cloud, and we have all become accustomed to using Office applications in the cloud, with renewable licenses every year. No more piracy. Now Office is available with any web hosting.
It wouldn't be surprising if OpenAI looked at which GPTs are used the most and developed them further internally. In fact, that's already something it has done in the past.
In an environment where GPT technology is becoming commoditized, with Mistral.ai or Meta Flame, OpenAI is taking a firm step to differentiate itself from the competition, because this is no longer just about the technology, but the whole ecosystem that carries that technology: Building that ecosystem will give OpenAI a huge barrier to entry against competitors. It's a move to make sure they're going to be around for a long time.
Because OpenAI plans to stand out from competitors in several ways. Let's look at their pitch: some things we can believe, others are more difficult.
"Democratizing AI development": By providing easily usable tools through the GPT Store, and with a familiar interface, OpenAI potentially allows individuals, companies and institutions to "develop customized solutions tailored to their needs". This approach lowers the barriers to entry for AI professionals who are starting out but lack extensive resources or technical expertise. Ultimately, the democratization of AI development drives cross-industry growth and helps spark other innovations in the field. (Let's buy into this altruism and goals for the global good).
"Foster mass creativity": As more people interact with the GPT platform, diverse ideas emerge, leading to further experimentation and exploration. This can go a long way in refining existing models and inspiring novel applications, pushing the boundaries beyond traditional limitations. OpenAI therefore benefits greatly by mining collective knowledge and ingenuity and creativity generated by its user base. (This is all the more credible because the business models that succeed are the ones that leverage people's free labor or time, from Gmail to Facebook.)
"Identify popular use cases: As OpenAI aims to closely monitor popular GPT models, identifying popular use cases as they emerge is critical. Exploiting information gained by tracking in-demand models allows OpenAI to prioritize its own internal research and development efforts toward popular, profitable and impact-maximizing areas. As a result, it is all about investing to preserve market dominance.
"Differentiate by building ecosystems": To stand out against other fast-growing rivals, focusing exclusively on technological breakthroughs is not enough. Instead, building a broad ecosystem around core products distinguishes industry leaders from those who follow in their footsteps. For example, Microsoft's acquisition of GitHub clearly demonstrates a solid understanding of the importance of cultivating strong relationships within the developer community. Similarly, OpenAI seeks to position itself as a preferred destination for AI innovation thanks to a thriving ecosystem that supports it, not only because of superior technology, but also because of a vibrant ecosystem.
"Ensuring long-term success": Finally, creating a robust ecosystem serves another crucial purpose: ensuring lasting relevance in a rapidly evolving industry. Even established players such as Apple or Microsoft constantly face many challenges. Therefore, taking steps aimed at consolidating brand loyalty pays off. In this context, OpenAI's moves function as insurance against future threats, shielding its position while simultaneously attracting talent and investment.
ChatGPT is still a very powerful technology and OpenAI continues to try to monopolize access to it. But, if we were in that position who wouldn't? Meanwhile, when you try the Store's "custom solutions", you can't help but feel like year 1 after the iPhone: in 2008 there was no Instagram, no Uber, no WhatsApp with stickers.
Let's see: The percentage of people who have integrated generative AI into their daily business is still small. Radical changes were promised and the changes... are not being substantial. The most massive use is as an aid for students when doing homework or writing college essays or drawing pictures. This shows that one of the sectors that will be most affected by a wide deployment of artificial intelligence will be the education system, as we discussed last year (How the new ChatGPT will bring radical changes to the world as we know it).
Last week, the Reuters Institute at Oxford University released its report on "trends and predictions for 2024" in which there was a survey about the use of Artificial Intelligence. 65% have "heard" that AI bots exist (ChatGPT leads with 59%, followed by 18% for Microsoft's Copilot and another 18% for Meta AI respectively). There seems to be little left for the rest. However, only 9% say they use AI on a weekly basis, a percentage that rises to 25% among 18-24 year olds. This is evidence that generative AI is very popular as a knowledge repository and as a text-generating tool, and "consumers" in the education sector (students) have found a practical use for the technology. This does not seem to be the case for the managers in the sector (schools, faculty, academics, etc.). This points to a challenge and a trend in other sectors.
That is, as 2024 begins, the three main uses of GenIA are to tinker, ask it questions, and ask it to write something for us. The use that will make AI ubiquitous like cell phones is yet to come. If it arrives, that is. This February, Apple's VisionPro mixed reality glasses will be released. There may be a way there, although it will be a long road: they will cost more than 3,000 dollars.