Story of the month
This month, we focus on the integration of Artificial Intelligence into consumer electronics products, in particular personal computers (laptops or otherwise) and cell phone handsets. Apple’s launch of the iPhone 16, its latest cell phone, and the Computex trade show in Taiwan (the high mass for personal computers) gives us the opportunity to talk about the adoption of AI in very large public markets. So it’s “AI on the edge” (Artificial Intelligence that enables smaller models to run on lightweight terminals, as opposed to “AI on the cloud”, which exploits larger models that can’t run on the edge and need the power of the cloud to operate) that we’re going to detail here.
AFTER INFRASTRUCTURE, THE AI BATTLEFIELD SHIFTS TO THE GENERAL PUBLIC
The opening of ChatGPT to the public in November 2022 revealed that AI technology (despite widely mocked cases of “hallucination”) was now ready for mainstream adoption. In our view, one reason for this was that querying an AI (via ChatGPT or other language models subsequently launched by OpenAI’s competitors) no longer required computer skills, as the interface was now in natural language.
However, in the months that followed, we mainly witnessed the construction of the foundations of the AI infrastructure, without which AI cannot develop. This covers both the computer servers that host the data (from which the language models are trained and answer the questions posed by AI users), the processors and electronic chips powerful enough to analyze growing quantities of data, and so on. It is therefore semiconductors and network and storage equipment that have concentrated the bulk of investment, i.e. hardware and software solutions installed at the heart of networks. Core AI therefore benefited from these investments and, as a logical consequence, from investor interest. The large language models (e.g. Chat GPT, Claude or Llama) form the “super structure” of Generative AI.
We believe that one of the polymorphic mutations of this artificial intelligence could soon be its displacement towards AI end-users at the end of these communication networks. This is where AI adopters are most numerous, and therefore where the potential market is the greatest. This shift to the edge of networks is therefore, unsurprisingly, the next step in the widespread adoption of AI, which is the ultimate goal targeted by all the major players in Tech. It will enable the use of smaller, more agile language models that consume less energy.
CELL PHONES: THE TANTALIZING PROMISE OF AI-POWERED HANDSETS BEGINS…
Several manufacturers have recently launched new cell phones enriched with AI-based applications, such as Google’s handsets equipped with Gemini (a virtual competitor to ChatGPT). Without wishing to favor any particular manufacturer, but given the marketing talent Apple has demonstrated in the past, here’s what we take away from the Californian giant’s announcements on September 9, and the conclusions we can draw for “AI on the edge”:
- New or enhanced applications: In particular, Apple detailed how the integration of AI into its tools would enable it not only to write a text (e.g. an email), but also almost instantaneously to change its style (more or less informal), translate or correct it. Similarly, AI will be integrated into iPhone 16 to enable the creation of emoticons and personalized images, i.e. based on photos taken and stored by the user in his or her phone, or even on a simplified diagram or summary description written by the user and describing the desired final effect. Apple’s web browser (Safari) should also benefit from AI to summarize web pages. Although these examples do not necessarily represent revolutionary new applications, they do confirm our conviction that AI is destined, at least for the next few years, to simplify the user’s life in simple but frequent needs, both professionally (productivity gains) and personally.
