Flapping Airplanes is the promise of research-driven AI

The new AI lab is called Flights was launched on Wednesday, with $180 million in seed funding from Google Ventures, Sequoia, and Index. The founding team is impressive, and the goal – to find a way to reduce data for training large samples – is very interesting.

Based on what I have seen so far, I would rate them as Level Two on trying-to-make-money.

But there is something very interesting about the work of Flapping Airplanes that I couldn’t put my finger on until I read it. This post from Sequoia partner David Cahn.

As Cahn explains, Flapping Airplanes is one of the first labs to continue to expand, the continuous generation of data and computing that has defined much of the industry to date:

The idea of ​​raising is against giving more human resources, as the economy can affect, to develop today’s LLMs, in the hope that this will lead to AGI. The research paradigm says that we are 2-3 researchers away from “AGI” intelligence, and therefore, we must provide resources for long-term research, especially projects that will take 5-10 years to complete.

(…)

Early calculations prioritize the number of teams over anything else, and will favor short-term wins (in 1-2 years) over long-term bets (in 5-10 years). The first search method can spread the bets for a while, and it should be willing to make many bets that have a small chance of working, but together it expands the search area for what is possible.

It could be that the computing people are right, and it’s pointless to focus on anything other than server architecture. But with so many companies already pointing in that direction, it’s nice to see someone heading in a different direction.