They’re essentially drawing parallels between the current technological landscape and past waves of innovation that brought us the internet, big data, cloud computing, and other advancements.

It’s important to note that none of these developments emerged out of thin air. They evolved gradually — built step by step. One could argue that the internet itself set the stage for artificial intelligence to thrive. At a minimum, it created the digital ecosystem where AI can operate — such as when models like Claude begin using the internet autonomously.

Periodically, it's useful to examine in more detail what has transpired, and why.

This theme was explored in a recent TED talk by Sungjoo Yoon, who started with a personal anecdote about early relationships and how he came to understand the volatility of a former partner who left him for his closest friend.

He likened this to shifting consumer trends and market transformations, where companies experience brief moments of popularity before being replaced by something new. He referred to this as “tectonic shifts” in markets — for example, how command line interfaces from 1969 were eventually replaced despite their foundational role in early personal computing.

“We saw in 1969 the development of the Thompson Shell, the first Unix shell that allowed users to input commands if they understood terminals, paving the way for modern terminal operations,” he explained. “You could type instructions and get far more complex computations in return. This marked the beginning of replacing the old with the new.”

He also highlighted the emergence of GUIs in the 1980s, web-native interfaces in the 1990s, and the clear technological progress of the 21st century.

What Drives Innovation?

Throughout his talk, Yoon discussed what he called the preference principle — the idea that having more insight into individual preferences allows for the creation of more powerful and attractive technology.

“I worked as part of a research team at Caltech in a behavioral economics lab, and this is actually a fundamental concept that applies globally,” he said.

Looking at welfare systems, it's essentially about identifying people’s needs and efficiently distributing resources accordingly. To do this well, you need accurate information on what those actual needs are.

He touched on auction theory and behavior based on preference principles, including concerns around overbidding. He also referenced employment and hiring decisions:

“If we want efficient global labor distribution, we need detailed knowledge of people’s actual skills so that roles requiring high-level expertise are filled appropriately,” he explained. “We shouldn’t assign someone unqualified for a task, nor should a skilled person be underutilized.”

Then there’s interface theory.

“This is an area that remains largely overlooked but is vital in consumer technology,” he noted. “When I say ‘overlooked,’ consider how people discuss the physical design of consumer tech. It’s odd because it’s not just about hardware size, right? That’s a definition many tend to use — focusing solely on product dimensions.”

He used the iPhone as an example of how a device has evolved over time.

“It's not strictly about physical measurements, or else we’d see a more predictable pattern. Nor is it only about the experiences those measurements produce.”

Developers, he suggested, are increasingly focused on these concepts in order to better connect with expanding user bases — particularly younger generations with evolving tastes.

“Greater participation means collecting more aggregated preferences from diverse groups,” he said, “and more aggregated preferences lead to higher quality information — which translates into better products.”

Yoon pointed to Google, YouTube, and LinkedIn as platforms actively utilizing aggregated data to enhance their offerings.

He also emphasized that there are more effective ways to gather this kind of information.

The Role of Language

Yoon also discussed natural language versus unnatural language.

In essence, asking someone directly for information often feels awkward and forced. In contrast, using natural language and social interaction tends to be much more effective overall.

According to Yoon, natural language fosters trust.

The Era of Autonomous AI

How does this translate into practical AI applications?

We're now witnessing AI agents becoming proactive — capable of initiating actions independently. These systems are loosely guided by humans toward specific goals but figure out the steps themselves.

This means we’re moving toward a future filled with non-human participants operating within systems previously controlled exclusively by humans.

That’s likely a large part of what Yoon is highlighting with his preference principle, and the gradual evolution of technology we must adapt to if we aim to build a better path forward.

The above is the detailed content of How To Build Products People Can't Live Without In The Era Of AI. For more information, please follow other related articles on the PHP Chinese website!