WTF is the ”Eval Loop” - Evals Deep Dive, AI News, and Mastra Docs with Paul

**Abhi** (0:00)
What up?

**Shane** (0:02)
What is up?

**Paul** (0:04)
How's it going, dude?

**Shane** (0:06)
I'm doing well. How are you doing?

**Paul** (0:08)
Pretty good. Pretty good. Got my Celsius here. Getting ready for the day.

**Shane** (0:13)
Yeah, I got a little echo coming from your end. I see you have a stack of books. You know, that's good. I do as well.

**Paul** (0:24)
I'm using this microphone.

**Shane** (0:27)
Yeah. I mean, maybe check your input. I don't know. Can you hear me pretty well?

**Paul** (0:32)
Hello.

**Shane** (0:33)
Oh, that's a little better.

**Paul** (0:37)
Hello, hello, hello.

**Abhi** (0:40)
Is that better?

**Shane** (0:42)
I think so. We'll go with it for now.

**Paul** (0:44)
Roll.

**Shane** (0:46)
If you're in the chat, let us know. Let us know how the audio sounds from Abhi's microphone. We can always work on fixing it, I guess. But welcome, all you admirals of AI. This is AI Agents Hour. I'm Shane. I'm here with Abhi.

**Abhi** (1:00)
What up?

**Shane** (1:01)
Yeah, we're going to chat some AI today, like we do every Monday.
If you, this is the first time you've tuned in. You know, we do this pretty much every week. Sometimes more, sometimes, you know, almost never less. And we have a pretty good episode today. We're going to be talking AI news like we always do. We're going to bring Paul in from the Mastra team to talk about some Docs updates. And then we're going to go deep on Evals because Evals is a hot topic right now. It has been for a while.

**Paul** (1:30)
Yeah. Well, before we do that, how was your 4th of July?

**Shane** (1:35)
Well, it was great, but I did dislocate my shoulder.

**Paul** (1:39)
Oh, shit, I thought that happened.

**Shane** (1:41)
So I was, I have a, my left shoulder is pretty weak and I was tubing and I hit a massive, I mean, my friend sent me sky high on the boat. And I landed, tried to hold on, popped my shoulder out, stayed on the tube, shook it out, and it popped back in and it was fine.
But I've dislocated the shoulder like six times in my life. So like, I've gotten good at like, it didn't, I don't think it fully dislocated because that'd be kind of hard to pop back in. It was like a partial dislocation. I could just like shake it back in and then it would pop back into place. So it's a little sore. But besides that, it was an amazing weekend. Got to spend some time with family and friends and.

**Paul** (2:21)
Do you see fireworks?

**Shane** (2:23)
Yes, I didn't go to any specific show, but where we're at, there's just like a ton of people around us. So we could just like observe everyone else's fireworks. We did shoot off a few. What about you? How was your fourth?

**Paul** (2:35)
It was chill, but SF fourth is way weaker than LA for sure. Shots fired, dude, because there are not that many people just doing fireworks illegally everywhere. And that's what makes fourth of July fun. But there was a bunch of... I went up the hill here and I saw a bunch near Bloom Saloon and stuff. So if you had a good vantage point, but then it just ended and in LA, they go off all night. So it's just a different vibe.

**Shane** (3:08)
Yeah. Someday you'll have to come check out some South Dakota fireworks.
We go pretty hard here in South Dakota.

**Paul** (3:17)
I can only imagine if it's legal, how crazy the fireworks would be.

**Shane** (3:20)
It's not legal in cities. Well, I mean, it's still people do it all over, but it's not technically legal in city limits. But where I was was outside of any big city limits. So people are just shooting them off all over.

**Paul** (3:33)
Oh yeah. Well, should we get into it? Yeah, let's do it.

**Shane** (3:38)
So this is a live show. That means if you're watching this on LinkedIn or YouTube or X or wherever you may be watching, you can leave us a comment. We try to respond to most of them. So let us know what you think as we go along through some of this stuff. The first thing we're going to talk about is we're going to do some AI news. So why don't we just get into that?
And yeah, let me share my screen and we'll talk about the first bit. So I thought this was kind of interesting. XAI raised 5 billion in debt and then an additional 5 billion in equity. So 10 million or 10 billion with a B billion total. So that's a lot of money. Obviously, it seems to be all around building out data centers and they want to build the largest data center, right? And the idea is that if you have the most compute, you will have the best AI. I mean, I don't think GROK has the best AI today. Obviously, I don't think anyone does. But it is kind of interesting to see if the bet's going to pay off, because if that holds true, they're going to be well positioned to potentially have some of the best at some point.