text stringlengths 0 2.67k |
|---|
**Autumn Nash:** Because honestly, that's what I want to know. Because at some point, all these data centers that are being built, all this compute power that people are buying - what are we going to do with it when they realize that AI is not going to make everything better? We are going to have such a surplus of thes... |
**Tim Banks:** Yes and no, because people were using these GPUs to do inference before. It's not going away. We may not have the boom of gen AI like we do, but I don't think it's ever going to go away. But that's the beauty of buying it. I don't need it anymore - I can get some of this money back by selling it. |
**Autumn Nash:** But do you think it's going to get to the point where everybody is going to have it, and it's going to be like not even worth selling? You know what I mean? You know how cars, and -- |
**Justin Garrison:** It drives the price down. It's the supply and demand. And that's not new, because in the early 2000s, the dotcom boom, all these people got rid of servers and a new breed of people that could do stuff off this cheap hardware, like "Hey, I don't need that big fancy thing. I just need a couple of sma... |
**Tim Banks:** What do you think powered the first Bitcoin miners for old servers? People's old desktops that \[unintelligible 00:43:01.17\] stuff like that. The notion of buying it means that when I'm done with it, I can sell it and get some of that money back. Is there a secondary market for it? Absolutely. There's a... |
**Autumn Nash:** Yeah. I think startups should start in the cloud, and then figure out what you're using, and when you figure out your scale... |
**Tim Banks:** \[43:51\] But the thing that I think where startups mess up is startups build in the cloud as if they're always going to stay in the cloud. And what you should be is agnostic as possible, because it may behoove you to be able to move at some point. And you need to make good architectural decisions in the... |
**Autumn Nash:** That's not even just for cloud, though. I think that's for -- how many times have you seen horrible tech debt come from the fact that people just didn't really spend a lot of time on their design and architecture in the beginning, even with databases? You know what I mean? |
**Tim Banks:** Low-bearing tech debt is what we call it. It's too expensive to not move. |
**Autumn Nash:** Especially working in databases for a few years, seeing the way that people did their database schemas and stuff, and you're just like "But why? Why would you do that?" |
**Tim Banks:** Do you think Oracle database is still around because people are like "Hey, you know what would be great for this new project? Oracle." Nobody does that. |
**Autumn Nash:** Yeah. So I think that is just sage advice in general in tech. People always want to code and build something cool, but it really pays to spend a little time up front. |
**Justin Garrison:** Well, and I've completely changed my position on good cloud architecture, where I think that lift and shift is the best model for the cloud, because it's the easiest way to get out of it again. And the cloud should be temporary leasing, like you do if you go travel to someplace on vacation. You're ... |
**Tim Banks:** I think the most efficient way to run is by default to treat everything that you have as if it's spot. Because if it can go away, that means I can put it anywhere. |
**Autumn Nash:** Yeah, I think you should do that in general, even when you're running on your own hardware, just because you always have options. And I think also just keeping it simple, because then if you realize something no longer works for you, it's easier to change your architectures and how you do things. |
**Tim Banks:** You know how long it's taken some people to have to rewrite code just to be able to use ARM-based chips, because they were very, very -- |
**Autumn Nash:** I don't know. I feel like we have all seen how tech debt goes wrong, and being able to make things neat and simple enough - for one, it's easier to fix, but it's also easier to add, or to... You know what I mean? |
**Tim Banks:** You want to be able to modularize it. You want to be able to tear stuff out and replace it. Or like if I want to switch providers, it should be as simple as changing a target or a variable. That's the ideal. Now, the practicality is it's never that easy, but you want to try to reach that ideal, so that w... |
**Autumn Nash:** Yes. |
**Tim Banks:** That's the only thing that's expensive to move, and that's why they charge you so much to egress, is because if you do it right, that's the only thing that's going to take some time and hard to move. Compute should be agnostic. |
**Autumn Nash:** That's the crazy part. And I've learned so much going from a solutions architect and then being an engineer, because solutions architect - you're constantly telling people about like best practices, and what you should do... And then when you actually have to go build those things, best practices don't... |
**Justin Garrison:** Well, I'm sure -- like, Tim practices jujitsu six days a week, and then he goes out and fights someone... And the practices help, but it's not the real world. When he's about to get punched in the face, then he has to actually adjust and say "Oh, this is different than practice." |
**Autumn Nash:** Y'all, Tim has one talk about -- what is that one that you did and it was all jujitsu? The last one you did, or like one of the more recent ones... |
**Tim Banks:** Oh, the one talking about fundamentals versus techniques? |
**Autumn Nash:** Yes. And then you had all this -- you had a bunch of jujitsu... Dude, it was so good, because it was like this real life thing, and you were bringing this real life experience with these pictures and these visuals, but it was also talking about technology. And it was wild... Like, it was such a good ta... |
**Tim Banks:** \[48:09\] So we talk about techniques, and a lot of people in jujitsu, or anything, we talk about technique. Technique is the ideal way that something should work. So if I'm going to do what's called a scissor sweep... So I'm on my back and I'm in guard. I'm going to grab your lapel of your gi, and I'm g... |
So I'm going to move on my side, I'm going to scissor my legs, and you're going to fall over on your side. And that's going to work pretty well in a drill. If I do it like in a \[unintelligible 00:48:35.11\] it's never going to work, because I have to understand the fundamentals behind why a scissor sweep works. |
A technique is the application of a fundamental. The fundamentals behind why a scissor sweep works is that I'm going to move your body to one side. So I have to take away a post. That's why I grabbed your sleeve. I need to control your posture and pull it down toward the ground, so that your shoulder hits the ground on... |
**Justin Garrison:** And know when to not use it in a real world... Like, is a scissor sweep the Zookeeper of jujitsu? |
**Tim Banks:** No. No, no, no. |
**Justin Garrison:** Sorry, I was just trying to bring it back... |
**Autumn Nash:** What's the Zookeeper of jujitsu? We're about to get spicy right here... |
**Tim Banks:** The jujitsu guys are going to be -- I'm going to get yelled at; I'm going to get screamed at in Portuguese, but the Zookeeper of jujitsu are guard pullers. The people that start a fight by sitting down. And under perfect, ideal scenarios that works just fine. Until you get a guy who knows how to pass gua... |
**Autumn Nash:** That's really interesting, because I feel like that's really simple and not what you would think... But also, it's -- see, my son does jujitsu, and I really wanna do it with him if I wasn't so scared all the time... But I love the fact that so much of jujitsu is in your head and it's strategy. |
**Tim Banks:** It is chess. High stakes chess. |
**Autumn Nash:** Yeah. But it's also -- like, you can be so much smaller and still be very competitive at it with someone bigger, because it's so much strategy. And I think that that is what makes you such a very unique -- like, you are so effective in technology, because for one, you have the fundamentals, you've done... |
**Tim Banks:** This is what it comes down to, is system and process analysis. Jujitsu is just a system. It's just a collection of processes. |
**Autumn Nash:** It is. It's amazing. |
**Tim Banks:** \[52:00\] So is tech. And so when I talk about -- going back to having context for what you're doing... If you're just doing a scissor sweep and you can do these moves and hope someone falls over, it's never going to work, because you don't have context for what's actually going on. When you understand t... |
So a coder who just knows coding is -- all they know is going to type code. But if they understand the fundamentals behind how the infrastructure works and how your architecture is laid out, and why things are the way they are, you're going to be able to make better choices about how you write code and how you fix prob... |
**Justin Garrison:** I think that's a great place to wrap it up, because we went full circle on this. So Tim, thank you so much for joining the show. Where can people find you? Where should they reach out online? |
**Tim Banks:** Oh gosh, I'm on Twitter at @elchefe. I talk adjacently to tech. I mostly don't talk about tech. I'll talk about dumb stuff, but... I like to talk about whatever. I don't have a professional and a personal Twitter account. I just have the one. So be advised. |
You can find me -- I'm writing a blog, it's at tim-banks.ghost.io. It's on Ghost. It works out well. I'm also on LinkedIn. You can find me if you know what I look like. You can go to Twitter if you're not sure and then go find me on LinkedIn. And that's mostly it. Feel free to talk to me out at the conferences if you s... |
**Autumn Nash:** If you want someone to make your whole day from inside of your phone, follow Tim on Twitter. He'll make your whole day. If you're having a bad day, go read Tim's tweets and you'll have a better one. |
**Tim Banks:** Yeah. No, I really appreciate it. This was a lot of fun. |
**Autumn Nash:** It was so much fun. I had so much fun talking to you. I'm so glad to finally virtually meet you. |
**Tim Banks:** Yay. |
**Break**: \[53:51\] |
**Justin Garrison:** Thank you so much, Tim, for coming on the show. It was a lot of fun. I don't think we've laughed that hard on any interview. |
**Autumn Nash:** Oh, my God. It was so much fun. I think I got three ab muscles from laughing at Tim. I don't have words... I think this is my new favorite episode. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.