Ben Gilbert: evaluation

Transcript context

…You don’t have to build tons of data centers. You can use S3 to store all your data. Yup. So, a variety of factors contributing to this. There are tens or hundreds of defined machine learning methodologies and I think something that is pretty interesting is there’s less than five that are actually used right now by a lot of people and have actually had a lot of research on them, a lot of time and practice. It’s interesting that like relative to a decade ago when there was still lots of green space of this is an emerging feel that’s been researched for a while but not commercialized, we don’t know which things work best. Now there’s like a fairly understood scope and scale and understanding of what they can be used for the few basic types or few most used types. What we’ve seen is the platformization of those. If you have machine learning tasks to do at your company that are not like wildly different than something somebody else is doing or don’t require any sort of like new research or actual mathematicians to be hammering on a methodology that’s not well understood yet, these are often available from Google, from Amazon, from Microsoft as cloud services that are built on tensor flow at Google or any of the other platforms at these other companies. What you see is like companies that might be 30-40 people don’t really need to build out a data science, a machine learning practice in a company. They just need to find a way to create really clean data and then feed it into these sort of like off-the-shelf systems that the big companies created. That makes me really excited as an investor because it’s going to enable so many more companies and so many more industries to take advantage of these tools. I think this will be one of the enabling factors that further pushes Aggregation Theory out into the world beyond the traditional borders of the “tech industry” that we talked about earlier.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.