Waiting for Netscape

What will bring large language models to the masses?

The Zvi quotes from an exchange on Twitter.

Timothy Lee: The last year has been a lot of cognitive dissonance for me. Inside the AI world, there's non-stop talk about the unprecedented pace of AI improvement. But when I look at the broader economy, I struggle to find examples of transformative change I can write about.

As of today, large language models (LLMs) are exciting the technology elites, but they have not had any mass-market impact. Ordinary households are not making use of the models. The business world seems to think of the models as research tools, whereas I think of them as human-computer communication tools.

Ethan Mollick is expecting democratization of innovation to emerge.

The lesson of innovation is that technology is only really useful when it is used. This might seem like a tautology, but it isn’t. When a new technology is introduced, people adapt it to solve the needs they have at hand, rather than simply following what the manufacturer of the technology intended. This leads to a lot of unexpected uses as the new technology is pushed to solve all sorts of novel problems. Right now, AI is in the hands of millions of educators, and they can use it to solve their own problems, in their own classrooms.

[but] I have yet to see robust communities of educators (or many other professions) coming together to share what they create.

Emphasis added, because the sentence echoes Amar Bhide’s concept of venturesome consumers.

But are these models really “in the hands of millions of educators?” I suspect that fewer educators have access, and of those hardly any have the time and the imagination to make good use of the models.

LLMs have not yet broken through to the mass market. This reminds me of the World Wide Web before Netscape.

Tim Berners-Lee proposed the World Wide Web protocol in 1989. His vision was for a text-based linking system suitable to academics sharing research.

Until 1993, the Web remained in that state, uninteresting to ordinary people. Then the National Center for Supercomputing Applications at the University of Illinois released Mosaic, a web browser with a more exciting vision. It could be installed on personal computers, and it included graphics.

Serial entrepreneur Jim Clark¹ soon hired Marc Andreessen and other Illini to create a commercial version of Mosaic, which came to be called Netscape. This company had a vision of making the Web commercial and bringing it to the masses. They succeeded in that vision, even though they faltered as a company.²

But there were other important companies that enabled mass participation in the Web. eBay allowed millions of people to become merchants. GeoCities encouraged many households to build web sites, and in 1997 it became the 5th most popular web site.

I think that LLMs are missing these sorts of intermediaries. We are waiting for someone to come up with a Netscape, an eBay, and a GeoCities for large language models.

substacks referenced above:

@

AI #60: Oh the Humanity

@

Innovation through prompting

1. I recommend the Michael Lewis book The New, New Thing about Clark.

2. I think that it failed because it had too many creative types and too few management types. I think you need a balance between the two.