Korva is a scroll of the internet’s best ideas.
Every dominant platform gets displaced
From Microsoft to Google
There was a time when the question was who could possibly disrupt Microsoft, and the answer was Google. Then the question became who could possibly disrupt Google. Where a search starts is already shifting toward OpenAI and Anthropic.
Who leads today
Google still looks like a net winner and in pole position, because it comes from search and applies AI with context better than most. Microsoft's AI feels second rate next to the top three or four, and it is not clear it can get back.
The next displacement
OpenAI and Anthropic will be displaced too, and there is an excellent chance the challenger comes from China. Underneath the models, the same pattern runs through chips: Intel once looked as though it could never be disrupted, and then came Nvidia.
How did Google beat earlier search engines?
In the late 1990s, AltaVista and Yahoo led search, and results leaned heavily on the words found on each page. Larry Page and Sergey Brin's PageRank algorithm, described in a 1998 paper, ranked pages by how many other pages linked to them and how important those linking pages were. The approach produced noticeably more useful results than keyword matching. Google was incorporated in September 1998.
Why did Nvidia chips end up powering AI?
Nvidia released its CUDA programming platform in 2007, which let researchers run general calculations on graphics chips. Graphics chips suit neural networks because both rely on huge numbers of simple calculations done in parallel. In 2012, Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton trained the AlexNet image network on two Nvidia graphics cards and won the ImageNet competition by a wide margin. The cards were a pair of GeForce GTX 580 models.
Why do dominant companies miss new waves?
Clayton Christensen's 1997 book The Innovator's Dilemma argued that leading firms listen to their best customers, who ask for improvements to the existing product. New entrants start with simpler, cheaper offerings that look unattractive, then improve until they serve the mainstream market. By the time the shift is obvious, the incumbent's business is built around the old model. Christensen called this pattern disruptive innovation.
More from 20VC with Harry Stebbings
