text stringlengths 0 2.67k |
|---|
**Autumn Nash:** I think that's one of those things that is always going to be relevant in tech. We keep getting new shiny toys like the cloud, AI, all these things, and we're like "Okay, this is going to do everything for us." And then we're like -- the newest hot take, "We're going to do everything on-prem." All of t... |
**Anurag Goel:** \[48:25\] Yeah. Really well put, both of you. I think it's always -- and this is what has allowed us to scale with our customers as they've grown, as opposed to losing them to them building out their own infrastructure on AWS. It's always been about the customer for us, in a very deep, sort of meaningf... |
**Autumn Nash:** Any scaling software... |
**Anurag Goel:** Yeah, exactly. And things are breaking all the time, there's chaos, your plans that you made at the beginning of the year have suddenly fallen apart, because guess what -- |
**Autumn Nash:** What you test is one thing, and then as soon as you give it to real people, you're like "What did you do?" |
**Anurag Goel:** Yeah, exactly. And all of these problems get compounded as you add more people to the team. Because now you have to focus on the \[unintelligible 00:49:13.29\] communication problem. And I think ultimately, focusing on the customer allows you to move faster in those situations. Like, what is the best t... |
Another example of this is - yes, our customers want GPUs these days. People want to train their own models. We spoke to a lot of people who are running AI companies on \[unintelligible 00:49:37.19\] and we realized that most of them are using third-party APIs. They're not training their own models. So while Render cou... |
So I think our goal is -- and obviously, I don't think we'll build everything for everyone. We can't. But we want to make sure that for 80% of the apps out there we serve your core needs, and then we make it easy for you to connect to other things. So we are also have, for example, customers who are using Mongo Atlas. ... |
And to go back to the point I was making around just being really customer-focused, and not -- so object storage. We have started working on object storage now, and I think our goal will still be to get it out to market ASAP, so we might actually use S3 under the hood. But our customers, again, they don't care. We will... |
**Autumn Nash:** But see, this is why people use the cloud. It's getting rid of the undifferentiated heavy-lifting. It's because you can now worry about the top layer, and they're gonna worry about that storage, right? And you get to make it easier for customers, and they're gonna pay you for that, and you're gonna pay... |
**Justin Garrison:** \[52:25\] Well, a lot of times it's a solution in search of a problem. |
**Autumn Nash:** Yes...! |
**Justin Garrison:** "Oh, I've found this cool thing. Let me go find who will pay me for it." |
**Autumn Nash:** I don't know if it's because it was zero interest rates for so long, but now it's like anything that involves AI, and we're like "Sir, we didn't ask for this." |
**Justin Garrison:** We almost went a whole episode without mentioning AI. We were very, very close. |
**Autumn Nash:** Sorry. |
**Justin Garrison:** No, I mean, it came up earlier. I was like "Ohhh, that was a --" |
**Autumn Nash:** No, but I mean just in general though... I went to a very big tech conference about a month ago, and I just walked the floor, and everybody was solving the same two problems, or solving problems that nobody asked for. And I'm like "We have actual problems that people can start startups to solve. I don'... |
**Anurag Goel:** Exactly. |
**Autumn Nash:** That makes me so confused. I really love that you're actually trying to solve a problem... What observala-- Justin, what's the word I'm looking for? |
**Justin Garrison:** Observability. |
**Autumn Nash:** Yes... And monitoring do you use to like get more insight into your customers? Because that is so underrated. People, find out where your customers are struggling, and then go solve that problem, and not just make up a problem with no capital. |
**Anurag Goel:** Yeah. I think for us there's a lot of different ways in which we connect with our customers. And we've always been that way. But as a starting point, it's really important for us to give people email support, even if they're not paying us anything. And the reason we do that is because we get feedback. ... |
**Autumn Nash:** It makes you a better engineer, though... Because you are seeing -- |
**Justin Garrison:** Everyone's help desk. |
**Autumn Nash:** Yes. That is so underrated. I hate when people are like "I have this fancy degree, and you guys were just help-desked, or whatever. And I'm just like "Dude, those are the best engineers." The most fire principal engineers are the ones that started solving someone else's problem. Because for one, you le... |
**Justin Garrison:** I mean, it's a long feedback loop. The error message that the customer copies and pastes back to you in an email is more valuable than the log you see when you're running debug. |
**Autumn Nash:** Can we go back to that though? Because that is so underrated. I'm just like "That's where people should start. Go fix something." |
**Justin Garrison:** Autumn, I'm gonna send you all my emails. \[laughs\] I mean, the scaling factor is the hard part. "Oh, hey, we get these emails", and it's this asynchronous medium, and... |
**Autumn Nash:** But there's also ways to do that though. Like, you can have something crawl all your emails and find how likely it is for you to get pinged for the same thing over and over again... But having that open channel is so important. |
**Justin Garrison:** Autumn's new startup is Email AI, so we're just gonna... \[laughter\] |
**Autumn Nash:** Don't tempt me. |
**Anurag Goel:** I would like to say that even though we mentioned AI, it was in the context of not doing the obvious thing with AI, which is \[unintelligible 00:55:26.24\] GPUs. It was to actually build object storage. So I'm totally with you. |
**Autumn Nash:** But I think that's so smart, because it's something that not everybody is doing, and it's an actual need. Please can we - because it's not that AI is a bad thing, just like the cloud wasn't bad. It's just when you use one hammer for everything, you're like "Dude, why?" So I think that's awesome. |
**Anurag Goel:** Thank you. |
**Justin Garrison:** Anurag, thank you so much for coming on the show and talking to us all about Render, and the backend of it, and what you're building next for customers... That's so cool to hear. And for anyone else that is interested, go check it out, render.com. And we will be back in a bit. |
**Anurag Goel:** Thank you. Really great to be here. |
**Autumn Nash:** It was a great conversation. |
**Break**: \[56:11\] |
**Justin Garrison:** Thank you so much, Anurag, for coming on and talking to us all about how the backend of Render works, and kind of where you're going with adding value to customers. I think it's really neat to hear just that they're doing multi-tenant Kubernetes clusters at scale, at their scale, for a lot of custo... |
**Autumn Nash:** I like anything that makes engineering lives easier. |
**Justin Garrison:** I would love to see more of them open sourcing some of those controllers and operators and the things that he talked about... Like "Hey, how do you actually do this in practice?" |
**Autumn Nash:** Even if it's like publishing learnings, or just any kind of contributions... I just hope that startups, if you're going to use open source software, please contribute back, because we're barely hanging on in a lot of like open source communities to begin with, with all that's going on in open source...... |
**Justin Garrison:** You've been figuring out from their practical experience what layer you want multi tenancy, and how you implement that tenancy. Like, we talked about the load balancing layer, versus the compute layer. |
**Autumn Nash:** That's why I'm saying it doesn't always have to be code. You know what I mean? But like just some sort of way that you can contribute back to make it better for the community... Please, community is important. |
**Justin Garrison:** Engineering blog posts are fantastic. |
**Autumn Nash:** They are! The Netflix one is like fire. It comes in my inbox and I'm just sitting there with like ice cream at night... Because the best way to learn how to not mess up is find somebody else's... |
**Justin Garrison:** I just had this mental image... I could totally picture you on your couch, with Murder Mystery out in the background, eating ice cream, and just reading the Netflix blog on your phone. |
**Autumn Nash:** In fluffy pink robes, and fluffy rabbit socks... Just so much pink and just ridiculousness and ice cream. |
**Justin Garrison:** For today's outro we're gonna go with the segment that we've done before called "How stuff breaks" or "How things break." And in this case, the breakage is actually Google ran out of things to deprecate, so they deprecated one of their customers. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.