The speaker argues that building AI agents or companies focused on building AI infrastructure is a waste of time, likening agents to infrastructure rather than a product that can form the basis of a company. They support this claim by drawing historical parallels with the internet's development. 

In 1998, the internet's infrastructure was established. Companies like Google created a new layer (search engine) on top, while Amazon leveraged existing infrastructure to build a new type of company without building its own infrastructure. Later, Amazon launched AWS in 2006, becoming an infrastructure provider itself. Companies like Netflix and Spotify then utilized this infrastructure.

The speaker extends this analogy to the period around 2012, highlighting companies like Uber and Airbnb that used existing infrastructure to build successful businesses. They assert that the situation in 2020 and the present remains the same: there is established infrastructure, and companies are built on top of it.

The core message is that attempting to compete directly with those building the foundational AI infrastructure is a less viable strategy for success. Instead, the speaker advises focusing on building companies that utilize the existing infrastructure developed over the past five years, as these are more likely to achieve significant success.

Jake Van Clief

Building agents, building an AI company is a waste of time. Agents are infrastructure, not a company. I can prove it with history.

In 1998, the infrastructure of the internet was already built. Google created a layer on top, but built their own type of search engine. Amazon didn't waste time building infrastructure. They used what was already there and created a new type of company.

2006, actually switched over to the building side with AWS. And now we see companies like Netflix and Spotify using the infrastructure.

2012 it's the same thing. We have the companies building it and the companies that are actually using it. Uber and Airbnb.

2020 is no different. We have the infrastructure and then we have the companies using it. Today we have the companies building the infrastructure and then the companies using it.

You think you can compete with the builders? Go ahead. Go for it. There is far more successful companies created using the infrastructure than there was building the infrastructure. Everyone trying to build better agents or harnesses or all these special things are competing with these people. But if you build on top of the ones who have already built stuff over the last 5 years, those type of companies are more likely to be huge and successful.

stop building AI companies. #coding #aitools #buildinpublic

Building agents, building an AI company
is a waste of time. Agents are
infrastructure, not a company. I can
prove it with history. In 1998, the
infrastructure of the internet was
already built. Google created a layer on
top, but built their own type of search
engine. Amazon didn't waste time
building infrastructure. They used what
was already there and created a new type
of company. 2006, actually switched over
to the building side with AWS. And now
we see companies like Netflix and
Spotify using the infrastructure. 2012
it's the same thing. We have the
companies building it and the companies
that are actually using it. Uber and
Airbnb. 2020 is no different. We have
the infrastructure and then we have the
companies using it. Today we have the
companies building the infrastructure
and then the companies using it. You
think you can compete with the builders?
Go ahead. Go for it. There is far more
successful companies created using the
infrastructure than there was building
the infrastructure. Everyone trying to
build better agents or harnesses or all
these special things are competing with
these people. But if you build on top of
the ones who have already built stuff
over the last 5 years, those type of
companies are more likely to be huge and
successful.

Get Our Mobile App

stop building AI companies. #coding #aitools #buildinpublic

Transcription