The future of logging and error resolution with Jim Mullady

How are large companies that employ thousands of engineers easing the pain of everyone suddenly working from home?

Learn all about it from Loren Carvalho, Staff Software Engineer at LinkedIn. 

We talk about what a developer container is, why containers are so important in developer tooling, and the history of containers that brought us to this point. 

Obviously, the world has changed with so many developers now unexpectedly working from home due to COVID-19. At LinkedIn, everyone formerly had their own Linux towers at their desk to do their development work. How do you replace that hardware remotely? 

Loren and the team are harnessing a Kubernetes cluster dedicated to running remote development containers, integrated with VSCode’s native DevContainer support. 
 
 Here are the links mentioned in this episode: 

• Loren’s website 
• Loren’s Twitter 
• Loren's LinkedIn

A content delivery network, or content distribution network (CDN), is a geographically distributed network of proxy servers and their data centers (Wikipedia). Basically, it speeds up the process of loading web pages by reducing the distance between your computer and the server where files are stored.

The use of CDNs is now ubiquitous. But, according to Skypack (a JavaScript CDN), "Despite the last 10 years of progress, webpage speeds have not improved over time." Matthew Phillips, an engineer at Skypack, joins the show to teach us more! 

Why is that happening? How are CDNs changing, and how is the Skypack CDN helping Node.JS be compatible with browsers?  And most importantly, what's coming next in the future of CDNs?  

Spoiler alert: Do we all become roving human CDNs? 

Here are the links mentioned in this episode: 

• Skypack
• Skypack's Twitter 
• Matthew's Twitter
• Matthew's website, with recipes 

What makes video APIs more complicated than web APIs?

Learn everything you didn't know you that you didn't know about the video API space with Wesley Faulkner, Developer Advocate at Daily.  

Video becomes more prevalent every day as COVID keeps many people indoors. While before, folks might have thought twice about turning on a video chat in their home, it's now commonplace. 

This has led to a big boom in startups that integrate video calls in their apps. But how will they deal with streaming real-time data and the rising standards for quality of those videos? Wesley has some thoughts. 

We close out with a fun side tangent on AI, deepfakes, and Wesley's optimism for the future of video and connection in the world. 

Here are the links mentioned in this episode: 

• Wesley's Twitter
• Daily's website
• Daily's video API demos 
• ThisPersonDoesNotExist.com 

Charity Majors is the CTO of Honeycomb. Listen in to learn about what "observability" is and how your team can slice and dice your way to higher uptime. 

Charity discusses figuring out how to talk about observability before it was an industry term, why many organizations still don't understand it, and why monitoring and logging will soon be rendered obsolete.

We close out with a fun conversation on where we'll be in 100 years: Will everything be AI, or will there always be someone with a wrench figuring this out? 

Links mentioned in the episode include: 

• Honeycomb's website
• Charity's website
• Charity's Twitter

Mrinal Wadhwa is the CTO at Ockam. Listen in to learn why secure communication in IoT is the scariest thing you're not thinking about. 

There is an increasing number of items in our house that connect to the internet: our TV, our car, our doorbell. These IoT systems both sense and control... and you don't want someone other than you controlling them. 

We discuss massive security breaches, how Apple finds your phone without being creepy, remote work accelerating the need for security, and what's coming next in this space. 

Stay tuned until the end for Mrinal's take on which connected devices in your home you should go unplug right now. 

Links mentioned in the episode include: 

• Mrinal's Twitter 
• Ockam's GitHub
• Ockam's primary repository
• Mirai Botnet
• Hackers take over a tornado warning system

Dan Moore is the Head of Developer Relations at FusionAuth. Listen in to learn how the future of authentication may look different than it ever has, and the tools driving us towards that future. 

These systems have always existed in some form: Authentication is any form of verifying who you are, authorization is any form of verifying what you can do. These systems are getting more complicated to make your life more simple.

We discuss changing standards, the guarantee of more security breaches, and the surprisingly near future of biometrics for authentication. 

Dan is confident you'll give up privacy for convenience– Because most people already do. 

Links mentioned in the episode include: 

• Dan's Twitter 
• FusionAuth site
• FusionAuth blog
• FusionAuth expert advice
• Download free FusionAuth community edition
• JAAS paper Dan wrote

Roopak Venkatakrishnan is a Staff Software Engineer at Bolt (Previously Google, Twitter, atSpoke) and a DevTools aficionado. He is building Swissknife on the side, making "big company tooling" available to all. 

Join us for a conversation about why local development is so important, why ~80% ability to test locally is the right ratio, and the startups trying to give you a plug-and-play local development environment (Porter and Blimp). 

We then discuss what is coming next for DevTools, especially platformizing local developer setup so it’s even more simple to get started.

Tweet us your thoughts at @DevToolsTopia or at @roopakv!

Diana Pfeil is a startup CTO with 15+ years of experience with machine learning and optimization, including a PhD from MIT.

She's seen machine learning developer tools grow from non-existent to a complete Python ecosystem, pre-trained networks, and numerous out-of-the-box paid solutions. That said, her position is to stick with open-source cloud provide tools– hear why!

Diana and I also discuss equality and fairness in machine learning, and how it can be used for good vs. evil in society. We can use machine learning eliminate bias in society, but we have to be intentional. She's a fan of the tool Shap to check your outcomes. 

We close out with discussing the future of DevTools for machine learning and whether or not I should be scared of Skynet running the world. 

Jim Mullady is a Sales Engineer at Coralogix, where they build machine-learning-powered logging integrated with your CI/CD pipelines.

He's seen logging and monitoring evolve throughout his career. Jim says as an industry, we're moving from looking at logging only after an incident to looking at logs and squashing bugs ahead of the incident. 

Machine learning is a major part of this shift in both detecting an upcoming issue from anomalies and detecting what's a real issue vs. noise. 

Shawn Harris, Software Engineer at Google, joins Kate for the second episode! We discuss how process is the ultimate DevTool because the hard part of software development is not the coding, it's working together.

Agile Scrum and Jira are Shawn's personal favorites, but we also talk about how to tweak processes depending on the team you're on and even define a step-by-step plan to help you get new processes rolling on your team. 

We close out the episode with talking about how software development processes will change in the future, including Shawn's hypothesis that it will incorporate psychology to accommodate for modern pressures like pager duty and CI/CD.