Containers got rid of all that duplication of the guest OS. It provided the process level isolation. It’s more lightweight and agile. Serverless or ‘Functions as a Services’ (FaaS) is even more exciting because you're now able to take code functions and run them at the granular level. All of the underlying machinery of having containers being spun up using communities is hidden behind the scene. That’s why ‘FaaS’ is becoming so popular.

Watch on YouTube: https://youtu.be/7vnRbiTT3mQ

For this episode of The New Stack Makers, Host TC Currie is joined by Steve Herrod, VC at General Catalyst where he focuses on infrastructure. He recently published an article on LinkedIn called Cloud vs. Edge – This is Not a Cage Match.

“It comes down to the use cases,” Herrod said about the difference between Cloud and Edge computing, "You don’t want your car’s IoT chip to have to go to the cloud before it decides to make the turn, but it’s faster to go to the cloud to crunch a bunch of sensor data from thousands of IoT devices."

Listen in to hear how existing companies are like elephants, where Herrod looks for engineers to staff his startups, recommendations for existing companies wanting to update their stacks, what he sees as the trends for 2018, and which two things about serverless fascinate him.

For this episode of The New Stack Makers, Host TC Currie is joined by Steve Herrod, VC at General Catalyst where he focuses on infrastructure. He recently published an article on LinkedIn called Cloud vs. Edge – This is Not a Cage Match.

“It comes down to the use cases,” Herrod said about the difference between Cloud and Edge computing, "You don’t want your car’s IoT chip to have to go to the cloud before it decides to make the turn, but it’s faster to go to the cloud to crunch a bunch of sensor data from thousands of IoT devices."

Listen in to hear how existing companies are like elephants, where Herrod looks for engineers to staff his startups, recommendations for existing companies wanting to update their stacks, what he sees as the trends for 2018, and which two things about serverless fascinate him.

For this episode of The New Stack Makers, Host TC Currie is joined by Steve Herrod, VC at General Catalyst where he focuses on infrastructure. He recently published an article on LinkedIn called Cloud vs. Edge – This is Not a Cage Match.

“It comes down to the use cases,” Herrod said about the difference between Cloud and Edge computing, "You don’t want your car’s IoT chip to have to go to the cloud before it decides to make the turn, but it’s faster to go to the cloud to crunch a bunch of sensor data from thousands of IoT devices."

Listen in to hear how existing companies are like elephants, where Herrod looks for engineers to staff his startups, recommendations for existing companies wanting to update their stacks, what he sees as the trends for 2018, and which two things about serverless fascinate him.

Bryan Jacquot of HPE tells a story in this episode of The New Stack Makers about a company in the healthcare space he visited recently that has an application dating back to the early 1990s. Now this application from the big hair era faces competition from companies deploying more sophisticated, more modern services.

So what are these companies to do that have these relatively ancient software infrastructures that have less relevance than ever before? The challenge comes with how developers in larger companies that are largely building out monolithic technologies.  Part of the answer comes with building ancillary services. But that's just barely touching on the issue.

Bryan Jacquot of HPE tells a story in this episode of The New Stack Makers about a company in the healthcare space he visited recently that has an application dating back to the early 1990s. Now this application from the big hair era faces competition from companies deploying more sophisticated, more modern services.

So what are these companies to do that have these relatively ancient software infrastructures that have less relevance than ever before? The challenge comes with how developers in larger companies that are largely building out monolithic technologies.  Part of the answer comes with building ancillary services. But that's just barely touching on the issue.

Bryan Jacquot of HPE tells a story in this episode of The New Stack Makers about a company in the healthcare space he visited recently that has an application dating back to the early 1990s. Now this application from the big hair era faces competition from companies deploying more sophisticated, more modern services.

So what are these companies to do that have these relatively ancient software infrastructures that have less relevance than ever before? The challenge comes with how developers in larger companies that are largely building out monolithic technologies.  Part of the answer comes with building ancillary services. But that's just barely touching on the issue.

Bryan Jacquot of HPE tells a story in this episode of The New Stack Makers about a company in the healthcare space he visited recently that has an application dating back to the early 1990s. Now this application from the big hair era faces competition from companies deploying more sophisticated, more modern services.

So what are these companies to do that have these relatively ancient software infrastructures that have less relevance than ever before? The challenge comes with how developers in larger companies that are largely building out monolithic technologies.  Part of the answer comes with building ancillary services. But that's just barely touching on the issue.

Here’s something that probably doesn’t cross most developers’ minds:  In a distributed system whose components don’t share state with one another, how does one produce an application whose stated goal, if you will, is to create and maintain a state — specifically, something learned?

We already know that artificial intelligence routines such as decision trees don’t need to have state pre-assessed for them, to render results that seem rational enough.  Chess move algorithms, for example, don’t have to retain a concept of the active chessboard in memory, to rate the quality of a potential move.  They may not produce the best chess programs on the market, but they can evaluate moves, and they can beat amateurs.

Well, by definition, a machine learning application is expected to retain something — that’s what the entire “learning” part is supposed to mean.

“The thing that you really want to do from the get-go is, make a couple of decisions,” Bonsai co-founder and CEO Mark Hammond told The New Stack Makers.  “The first is going to be, are you going to have a system where online learning matters — that is to say, where the system needs to learn concurrently, with users utilizing the system?  Or can there be a lag in that?  Is it possible for you instead to record the data that has come in, as users use the system, and then learn on those recordings in a more batch-style operation layer?”

Here’s something that probably doesn’t cross most developers’ minds:  In a distributed system whose components don’t share state with one another, how does one produce an application whose stated goal, if you will, is to create and maintain a state — specifically, something learned?

We already know that artificial intelligence routines such as decision trees don’t need to have state pre-assessed for them, to render results that seem rational enough.  Chess move algorithms, for example, don’t have to retain a concept of the active chessboard in memory, to rate the quality of a potential move.  They may not produce the best chess programs on the market, but they can evaluate moves, and they can beat amateurs.

Well, by definition, a machine learning application is expected to retain something — that’s what the entire “learning” part is supposed to mean.

“The thing that you really want to do from the get-go is, make a couple of decisions,” Bonsai co-founder and CEO Mark Hammond told The New Stack Makers.  “The first is going to be, are you going to have a system where online learning matters — that is to say, where the system needs to learn concurrently, with users utilizing the system?  Or can there be a lag in that?  Is it possible for you instead to record the data that has come in, as users use the system, and then learn on those recordings in a more batch-style operation layer?”