Below is a post explaining the usefulness of personal agents and some use-case examples from the existing market.
In quite a lot of ways, it overlaps with my own thinking. But personally, unlike the writer of that post, I think the issue is not a form in which a company or product monopolizing information collection or distribution channels takes the lead. Rather, we need to look for ways for individuals to directly build their own agents and, through them, easily streamline the whole process of collecting their own information, accumulating it, and distributing it through distribution channels. That is my view.
Reference
The ultimate form of private data is probably personal data.
Google,
The company that has secured the largest amount of personal data in the world is Google. Google knows the searches I have made over the last several years, the web page visit history accumulated through Chrome, what emails I send and receive through Gmail, and what content I watch, cry over, laugh at, study with, or relax with through YouTube. Google, which directly stores the most concentrated data on each individual person, will ultimately be able to leverage my data the most in the future on top of standardized algorithms and hardware.
Apple,
In that sense, I think Apple still has one big move left. Apple does not directly store data as a matter of principle, but it holds the smartphone, the passage through which all my data flows. In other words, it controls the distribution path of information. If it only gains my consent, Apple is the place that can access all of my personal data from the closest point. When LLMs light enough to run smoothly on phones become mainstream, an iOS integrated with an LLM could directly launch apps on my behalf, read reviews on Baemin, choose a menu that reflects my taste, and even finish the payment for me. No, it almost certainly will.
Then how should ordinary companies approach AI business?
When you look at companies these days, you often see cases focused on titles like, "We use GPT too. We build services with AI too." For now, individual technical elements may seem important, but if you think about the broader direction of the future, I think the most important question for a company becomes "How can we create a channel for acquiring independent and exclusive private data?". We should look carefully at whether somewhere in our business we can become a collector of information like Google, or a monopolist of information distribution channels like Apple. Exclusive private data is discovered around the core business our company is already good at. Look at your users and your business. Think about what kind of data only your company can possess.
