AI Killed the Podcast Star

Hello, John. Hello. I feel like we have a little bit of a down mood today. Life is just getting to you. I don't know. Everyone's ground down a little bit, I feel like. I think it's just going around. Plus, everyone's been sick. I actually didn't tell you this, but remember when you got sick? A few days after that, I got sick, but it wasn't flu. I actually think it might have been RSV or something.

Didn't seem like a cold or flu or COVID. I actually tested negative. I didn't test for RSV, though, because I don't have one of those. That was just really just terrible chest. And here it's like three weeks later now, and I still like in my chest, you know.

Probably have some bronchitis now from it. Yeah, I'm still dealing with a cough and probably clearing my throat quite often, so hopefully I can do that away from the mic. But yeah, that was the nasty little bug I got. Yeah. Yeah, it's been a while since we... Recorded, man. I know. Two months? One month? It's not been two months. Six weeks, probably. It's not been six weeks. We recorded early January, didn't we? Let's see. It's like the first week. December 31st.

So December 31st. No. Yes. Perfectly spelled. Yep. Six weeks. There goes my predictions. Yeah, exactly. Let's see. What do you want to start with? There's only one thing I want to talk about. Short show. We can end the show. Okay. You're quitting. You're quitting. Good day, sir. Quitting the show. Done with it. No one listens anyways. Don't make me laugh. You're going to make me choke here. Cough a blowing.

This is just for my ego. I just put a mic in front of me and make me feel important. Right, yeah. No, that's the Super Bowl. Oh, okay. I was in such rage about the Super Bowl ads. Well, the Super Bowl was fun because I was really rooting for the Eagles because I was so anti.

Swifties. Oh, okay. Yeah. That's the only reason I'm against the Chiefs is because I just got so tired of all the media attention over the Taylor Swift and what's his name? Kelsey? No. Kelsey? Travis Kelsey? I guess. Whatever her boyfriend is. I just got so sick of that being such a focal point of the Super Bowl. I know. I don't quite get that. It would be cool for a team to get three in a row, but still, at the same time, I'm like...

See, I was kind of against the Chiefs because I don't think anyone wants a three-peat except for Chiefs fans. Well, yeah. I don't think that's cool. You don't? No. I'm kind of tired of it. And I feel like, you know, I'm in this camp that I feel like the league has not done a very good job of officiating. And it's not just... Yeah, I mean, the Chiefs are an obvious one. Like when you just watch how that quarterback is treated and some of their other players, it's just insane.

It's like no one can watch that. We all have 4K TVs and they have replay from 80 different angles. We can see what happens. We can see the truth. So it's just, I'm surprised they're still trying to get away with this. It's just bad. But it's not just the Chiefs. It's a lot of games. And I don't think it's necessarily a conspiracy or intentional. I think a lot of it's just bad officiating. Yeah. There's a running saying that the...

NFL is rooting for the Chiefs so they can get some of the Swifty money. Yeah. I got a friend who lives in Kansas right now. Or is from there or something. Anyways, that's his team. He made the joke. He's like, yeah, we all we all pay our our fee to the NFL. Yeah. Well, I mean, you know, the one thing that's the one thing you can say about the NFL is that they're all about money. It's whatever.

I think that's what annoys me about it the most, because you can see, maybe not subtly, maybe I'm seeing it from my own point of view, but you can just see how skewed it seems to be. Because of that type of attention that it was getting, that it brought in this whole new audience, and it was just like, oh, we want that. Well, that's why they gave Taylor Swift so much attention, because it was bringing in a big chunk of a new audience to these games. Yeah.

There's always a reason for everything that happens in the NFL, and it all boils down to money. Money like ads. Yeah. The ads were kind of disappointing, I thought. There were a couple that were good. I don't know. I thought like no one makes funny ads anymore. It's like, come on guys. I mean, there were, again, there were a couple that were, there were some that were pretty cringy. That tongue one. Oh yeah. That was grossing me out the whole time. I'm like,

And then there were parts of them like, my kids are here. Yeah, exactly. Like, I don't want to explain that to them. Yep. Halftime show was a love or hate for people. My kids loved it. I mean. That was their type of music. They were happy about it. They were enjoying it. I had not even heard of that guy before. Oh, I have. I don't understand.

I mean, you couldn't understand anything he was saying. Obviously, there was, you know, there was not a single musical instrument on the field. So it was, this was not about, it was not music. It was not about music. It was some kind of, again, I didn't know this at the time, but there was some kind of messaging about, you know.

His grievances with this country and with his girlfriend's ex-boyfriend or something, some other dude, Drake. And so much of it was about that. And I was just like, the more I learned about when people try to tell me, you know. teach old man, tell the old man what's going on. The more I disliked it. I'm like, how do we, how did the NFL let them get used for this? I just, I don't know. It was weird. But again, I think that shows the difference between the way the younger.

crowd viewed that performance in the way the older crowd viewed it so the older crowd is reading into everything you know what is he saying why is he saying that why is why are we seeing this type of imagery but the younger kids they're just listening to music and enjoying it you know so i think that's

That's a big difference in how people consume that. If you think about music that you listened to when you were younger, you weren't analyzing it. You weren't even really inspecting the lyrics. At least I wasn't. I was just listening to what it was. It's only until later, I'm like, oh, I didn't know what that song was about.

You know, and I think for the younger generation, the kids, that's them right now. They're listening to it. Yeah, they know the words, but they're not really comprehending them or soaking them in or, you know, doing anything other than just kind of the... passive listening i think a lot of the kids the message the main messages he had that were completely lost on people like your kids like i don't think they got that part of it yeah no they don't okay they don't so i don't know again it's it's

His music is not my thing, so I'm not the target audience at all, I guess. And I'll just leave it at that because people get really upset if you say you didn't like their Super Bowl halftime show. I've learned. Oh, really? Oh, yeah. I didn't think it was that controversial. I thought it was just like one of those 50-50 splits where the younger generation, that's their generation of music. They enjoyed it and they were fine with it. And then you had the older generation that, I don't know.

wanted country music for some reason people are like let's put country music let's put country yeah i don't know like country music isn't even country music anymore it's pop now yeah well again we just sound like a bunch of old

I'm talking about modern music, which is exactly what we're doing right now and what we are. Anyways, to the ads, because the ads really pissed me off. And I'm not the only one that thought this. I found an article that someone wrote that was just like, they must have been in my head. Because obviously I want to talk about the Matthew McConaughey ads. Who did he do ads for? Was it Salesforce?

Actually, I think he did a different ad as well. Yeah, I don't remember a Salesforce ad. No, he did a Salesforce ad. There was a Salesforce ad. They spent a ton of money to do a Salesforce ad. But it was like...

It was an ad that we've seen before, at least I've seen before. And maybe it's one of those I've seen it before because they do the pre-release of the ads like on YouTube. Like you used to have to wait till the Super Bowl to actually see the ad. Now you can see all the ads before then. But anyways.

I'll tell you a couple of things that made me mad. The ad was a tell, not show. And when we're talking software, a software company, I think the ad should be show, not tell. Yeah. The focus was on the celebrities. And the scenario had made zero sense. So the scenario is that Matthew McConaughey is out to dinner and he gets reservation for this, for this.

place to eat and he's dining by himself and he's sitting outside, but it's downpouring. Oh, yeah. And then it looks across the street and there's Woody Harrelson with his buddies there chumming it up and laughing at McConaughey and said, hey, come over here. And the voiceover...

Is that, you know, oh, this restaurant didn't use AI, so it didn't know to change my reservation from an outdoor seating to an indoor seating. And I'm like, I didn't see anyone attempt to make any kind of defense of that ad. No one did. It was universally panned. Like, this is the dumbest ad. Well, first of all, that's not how the AI tooling works. It's not that active. And if it is, it's because you...

You did a lot of work to make that happen. It's not an out-of-the-box feature. And the way they told it made it seem like, get Salesforce and it'll manage your reservations for you. This reminds me of... The person who thought it was a good idea when they demoed, was this during COVID or whatever? I can't remember. When they demoed their latest AI, which was the thing that looked at the Coke machine to see if it was.

Low on Cokes, I would reorder Cokes. You're right. I remember that. Oh, man. I wish we still had clips of that stuff. We had a nice database. Technology is a wonderful thing. I know. Or like the fridges that have cameras so that you can see what's in the fridge. It's almost just like, yeah, I'm still waiting on that fridge that can order my milk for me automatically. Yeah.

That's impossible though. With as much milk as my family goes through. They can get hacked and put mimes on. Yeah. Sounds like you got that reference. You should have. Mimes? Yeah. I don't get that. Suck it, Jin Yang. Oh, is that, what's that show? I forgot the name of the show. That's why. Silicon Valley? Yeah, there you go.

Anyway, so in this scenario, Matthew McConaughey is not playing a character. He's playing himself. So the story revolves around him making a reservation. Now, Matthew McConaughey goes to a restaurant. And they don't cater to his every needs. They don't go, oh, hey, it's pouring outside. Let me get you a table right inside over here. Second of all, Matthew McConaughey dining alone? Doubt it.

And then the premise of the ad, which I didn't really think about until I read this article. In fact, let me pull up this article so I can read it the way he said it, because it was it really. It really spoke to me. Let's see. Let me scroll down. Okay. He says in this title, what AI was meant to be. Given what we've seen here, it is clear that AI is meant to be.

is clearly what ai is meant to be ai is meant to be ignored discarded rejected in favor of the shelter of community and the warm embrace of a friend and that's in reference to the end of the of the ad where uh woody harrelson calls out to him and says, hey, come over here. What are you doing since he's sitting there surrounded by people, a well-lit restaurant and everything, and McConaughey's over there in a dark restaurant being rained on. And it wasn't AI that made...

that that brought him over and and you know solve the day it was people and so the argument here is that the ad is basically saying people and community are the ones that that really make this special that make the dining out experience special not the ai that knew what to order you, what type of food to order you and where to put you seating wise. Yeah. It was just, I don't think it worked. No, it's horrible. Yeah. It just, it just shows a big disconnect between.

what's Salesforce marketing is doing and, and what the software is doing as well. And I don't know how, if I said that correctly. Yeah, I know. I know what you mean, I think, but that, and also that ad's not new. That ad's been on, I've seen that. I'm probably on sports. Cause that's the thing. I think that's the only thing I watch on TV anymore. Yeah.

For a couple months, maybe? Yeah, like I said, I think I've seen that ad before. So they've had plenty of time to get feedback on it. I mean, it must... I don't know. Maybe you and I are in a minority. I think people just saw it and ignored it. They didn't even really understand it. Could be.

I mean, aside from the Salesforce logo that appears somewhere on it, does that say Salesforce ad? Do you think Salesforce? No, but this goes back to the... And I don't know. I still can't tell if this is a joke or not. My Twitter, I feel like, is full of this. Like, jokes about what is...

Like around the fact that like no one knows what Salesforce even does. Yeah. So I wonder if it just kind of falls into that trap. Like people kind of ignore it because they don't understand Salesforce anyway. Yeah. As, as prevalent as it is, um, I still get a lot of people that don't know. what it is and the ones that do use it they're not in awe of it it's not like you look at the consulting world the the uh implementation partners and you

And you see how they react at Dreamforce and everything else. And you think Salesforce was this God's gift to the software world. You know, it's just celebrated. It's lauded. It's God's gift. It brought me out of poverty. Yeah, I think it's, oh yeah, all that stuff. But then you talk to real people doing real things and they're like, yeah, that's right. I got to use it. My company makes me use it.

A lot of them will have complaints. It's slow. It's clicky. I just think there's just so many bad implementations. And the bigger the company or the organization that's using it, the... the more likely it's a crap implementation because that means that you can only pretty much hire like an Accenture or a Deloitte or one of these other garbage companies. Well, it suffers from the fact that it is really flexible.

and that you can do quite a bit with it. It's up to you. So you end up doing things that you probably shouldn't be doing with it. Like in our case, because we have licenses for everybody, it becomes natural to want to put things into it.

Oh, you've got this thing you need to start tracking. Let's put it in Salesforce. You've got this system you need to interact with. It doesn't exist, right? If you've got this system you want to interact with, but you don't want to buy more licenses for more users, then just integrate it with Salesforce and make it sync. And then before you know it, you're syncing with...

10 different systems. Troubleshooting anything is getting really complicated. It's getting slower. It's getting more complex. It's still a very clicky interface. So it's just... I think it just sometimes suffers from its own success. It could be, and I don't know what you do about that, but because it does so much of it, it just boils down to how well the implementation was conceived of.

That's a big part of it. And implemented. Yeah. So. But even the technology in itself makes it really difficult to maintain anything big. Yeah. I mean, the whole DevOps process or story around Salesforce is just so bad. I'm spending so much time trying to solve it, trying to figure it out, trying to come up with ways to make it work. And it's, it's almost insurmountable sometimes because the tools get you like almost there.

Like they started adding the scratch orgs and snapshots and things like that. And, and then I've always find something that just makes it impossible or makes it so difficult. I'm like, this is turning into a crap show. I don't know what to do from here.

I couldn't create a snapshot. I told you that because there was some setting on that prevent some, some setting with connected apps. Like I think it's a setting that when you check it, connected apps can only be used by users. If you physically give them permissions at the profile level. to that connected app. Yeah. But if that's on and you create a, what do they call it? Not a snapshot. Is it a snapshot? No. Org shape thing? Org shape. Yeah. Then you can't create any scratch orgs off of that.

off of that shape so you have to turn off that setting create your shape then turn your setting back on and then you can start creating there's a lot of stuff and this is like one of these cross cross feature conflict things as well like Salesforce has so many products and and that have so many features that you start getting weird bugs that are the result of one

product or feature interacting with another product or feature in a way that they just never implemented right yeah and you see that a lot in metadata the metadata api and the whole deployment devops process i guess we're calling that now um you really start to see these things conflict

And you just, you can, yeah, you can, you can kind of just bounce from one brick wall to another. And you're just finding, trying to find a way out through this and you just keep hitting different brick walls. Yeah. Yeah. It makes it really tough to try to streamline things and improve it so that.

So the barrier for doing development on Salesforce is less. And somehow Salesforce is not going to hire any more software engineers to fix these things. Yeah, that's sad. We'll get to that. But since you mentioned, you're talking about connected apps? Yeah. So I ran into this a couple weeks ago. I've been trying to like migrate to at least, I guess there's like the replacement for connected apps, external client apps. Have you seen these?

In fact, now if you go to create a new connected app, it says, oh, are you sure you didn't want to create an external client app instead? Oh, yeah. I think so. I think I know what you're talking about. I forget right now what the benefits of those are.

Yeah, I'm not going to. It's a little bit newer. I mean, I will say that the UI for creating them and dealing with them is a little bit better. But I discovered the hard way that no matter what you do, like you can't enable this. They don't allow the username password flow. Right.

You're talking about named credentials and external credentials, right? No. Okay. Maybe, maybe connected apps gets built on top of that. Cause that sounds very similar to the changes that were made for named credentials. Cause it used to have. a username password flow for the name credential on the classic, but then they switched over to the new paradigm, which is the, it's two parts. So you have your named credential and then the actual authentication portion.

is handled by the external credential portion of it. And that has certain limitations. Yeah, Salesforce has gotten, they've really gotten opinionated on this. They started out by making... even for disconnected apps, making username and password flow disabled by default in orgs. So you have to go find that setting because you'll be getting authentication errors and you won't be able to figure out why.

And then you finally will figure out that you've got to that's disabled in your organ. You have to turn it on. So, yeah, learn that lesson. And now they've taken this next step of. The future of how. OAuth and authentication is going to work, it's going to have a username and password completely blocked. You can't go in and enable it or overwrite it. It's just kind of weird because they're all about...

Client credentials flow, which doesn't even require username and password. It simply requires the client secret. I mean, at least username and password flow requires the client secret and an additional set of credentials for the user.

The client credentials flow doesn't even require the credentials for the user. You just say which user you want it to be. So that, to me, and I think I'm just like... philosophically not getting the point maybe of client credentials but to me client credentials seems less secure than username because it requires less a smaller number of credentials

So I don't see why username and password is so bad that we have to completely block it now. Yet client credentials are just fine. Okay. Someone needs to explain that to me. Because I feel like there's something I'm just not getting. Because I do feel like the... authentication world in general. And if you read even stuff out of an author, whoever, forget what organization that falls under, but yeah, I mean, username and password is considered like, you know, something you should avoid.

Yeah, and it's to a detriment, I think. Yeah, well, I know what they mean. What they mean is they don't want... Your third party app taking they don't want a user providing their username and password to your third party app. In order to log into the app that has the resources you want.

You know what I mean? And I get that. But that's not what I'm doing. That's not how I'm using. That's not how a lot of people are using username and password flow. Well, that's kind of what I mean. I mean, I think it tries to solve one problem, but it ends up creating a problem for it.

for a lot, for the main use case of it. Yeah. And I can see that. I mean, I can definitely see like, Hey, we think it's a bad idea for you to let third parties collect your, your users, Salesforce username and password. And I agree, I wouldn't like that either. The whole point is, you let the third party bounce them to Salesforce to then... to authorize that third party app, that client app. Right. Yeah.

But I'm dealing with non-interactive things. I'm system-to-system communications. We're just securely embedding credentials. It doesn't matter whether it's client ID and secret or client ID, secret, and username and password. Like, they're strong credentials. They're stored securely. This username password flow in this case is no more or less secure, in my opinion, than client credentials flow. But again, I know what they're saying.

You say username and password flow, and people in their mind think, oh, you're collecting people's username and password for Salesforce, and you're just collecting that yourself and storing that, which, yeah, that's a bad practice. But anyway, but I mean, that's kind of how things, a lot of things work. The world economic forum had released a future of jobs thing.

I didn't actually... It was very fancy. It was very long. I really didn't read through it or even read much of the coverage of this. But I'm guessing it's more of the... Jobs are changing. A lot of these jobs are going away. You know, you're going to have basic income and eat bugs and won't own anything, won't have a car, you know. Yeah, hear those. Stir-fried grasshoppers are pretty good. Put a little chili sauce on them. Let me play this because I have a related clip.

To put World Economic Forum into context here. Let's see if this is going to play. I wonder if bugs are in the same family as shellfish. No, right? I don't think so. Well, I don't know. I mean, is a lobster a bug? Maybe. I think they're bugs of the sea.

Yeah, I know. I think it's the protein in the shells that, because I'm sensitive and told that I'm on the verge of having a full-blown shellfish allergy. Yeah. I don't know what it is. I don't know if it's a protein. So if we're all eating bugs, man, I'm screwed. you might be eating cabbages yeah exactly all right let me try this he did a i mean this guy's got you know just crazy energy who malay trump

Zoomed into Davos and just dunked on them for 45 minutes. He did? Yes. This is the all-in podcast, which has got, I think, really big over the course of the past year or so. And he just... destroyed everybody at their dying conference in their irrelevance as they genuflected, begging him to come next year. Did you see the photo of the half-empty conference room?

yes you know most of the events were were just empty chairs it's crazy it's irrelevant i mean so just pause for a second to put this in context so the both of those guys talking so i feel like all in it's kind of kind of a balance like politically like One of the two other guys seem like they're maybe on the conservative side these days, although didn't used to be. And then two of them, I think, are still fairly on the liberal side. But I feel like none of them are.

extreme in their views. This is just based on me listening to kind of an occasional episode from them over the past year or so. These are the two guys that are kind of more on the, I feel like, on the liberal side. Which I think is kind of meaningful, because if you had just some conservative types or some Trumpy types saying this, it would not be as meaningful. Just considering that economic form, I think, has always been viewed as a very progressive.

leading organization. Sure. Gotcha. These people are irrelevant now. Thomas, you guys remember when Davos used to be a really big deal? Is it dead? I think it's worse, Chamath. I think it's a counter indicator now. It is. And there was a good conversation yesterday. Graham Allison, who spoke at our summit, was on this panel, and they basically said, we lost, they won.

Our group here at Davos lost. It's over. Yeah. And they're no longer, to your point, on the right side of history. They said, we had a point of view on the future. We believe we were going in the right direction. And everyone... told us in this last couple of months that we were wrong. The gentleman, by the way, that spoke after Graham Allison was also very good. This is incredible. This is the greatest comeback in political history.

of a politician. And then, therefore, he thinks he can do anything. We need to also factor in not only who's won, which is Trump, but who's lost, which is to say us. And then he has a whole, the guy has a whole long monologue on why they lost and what they lost, which I think is pretty relevant that that, you know, kind of elite, if you will, has been replaced with a populist vote and a populist leader.

I thought they were talking about Davos, though. They were, I guess. And I didn't watch it. Maybe I should go back and watch it. But I said that Trump, you know, flew in and gave like a 40-minute speech at Davos.

Yeah, but the way they're describing it, it almost seems like the goal of Davos, in terms of whether or not they're successful or not... is indicative of whether or not they were able to have a certain party in power in the united states that's what it seemed like i know and i uh you know and that's not davos's mission it's not a political i mean okay

Loosely, it's not intended to be a political organization. It's meant to be more of a globalist economic forum. Yeah, I mean, I view them as they kind of, their view of themselves was rising above. politics, almost replacing politics. In some cases, they had some pretty anti-democratic efforts and activities. I think as certain people got up to speak, they were projecting some of their own personal views.

And unfortunately, it probably got entrenched into the to what the original mission was. Yeah. I think a lot of these organizations start out with good intentions and wanting to explore ideas and bring in. people from all over the world to kind of contribute to these ideas but then then they become i don't want to say corrupt but they just the ideas and stuff become to take on a life of their own and before they know it they're they're in this bubble yeah

And they lose touch with what the original mission was and what, what everyone else is experiencing in the world. I just want to know, are we still in the fourth industrial revolution or is this the fifth now? I don't know.

I don't hear anyone talking about it. But these are the same guys that are saying, like you said, that we're going to need universal income because AI is going to take over and everything. And I do think, to a certain extent, AI and computing and robotics and everything else is going to... and has been changed the way we work and certain, some of our jobs are not going to exist, but I don't think it's as doom and gloom as everyone thinks.

I try not to think that way. It might be doom and gloom for me personally, but I think general, if you just speak about general in the world, not everything revolves around software and computers, despite what Wall Street wants you to think.

The company's saying every company's a software company. Every company's not a software company. I had a conversation, so I had a daddy-daughter date with my daughter yesterday, and she was asking me about some of the things I do, and I found myself having to explain to her what a database was. You know, her interest is journalism because that's what she wants to be in journalism. And she had no and she was forced. She she tells me how she was really upset that she was forced and goaded by.

the education system to try to learn computers and try to code. She knew about them. I've exposed it to her. She was not interested, but everyone around her kept telling her to get us to go into STEM, to do this, to do that. And she didn't want that. But case in point, you know, not everyone cares about computers as much as we do. In fact, someone made the argument that computers are so good and so smart that people are losing the ability to actually...

understand them in the same way you can't fix your car anymore. Yeah. You really can't fix your computer anymore because they've gotten so complex and so automated and so computerized that there's nothing for you to do. You can't.

really change your oil or something. Yeah. Do you know that's a secret code to your car when you change your oil to turn your light off? Oh, the... Well, you can... Yeah, I mean, you have to... Yeah, I mean, there's just all these little things about your car that you have to know. Yeah. It's not as simple as it used to be.

You still put the dipstick in. Yeah, got oil. I'm good. Now you got to do the secret code, the brake, brake, gas, turn, turn the key knob, do a little jig in your car, just turn the light off. Well, they plugged the little thing into the OBD two-part to do that. Did you know my car does not even have a dipstick? You cannot check the oil level on it. Really? Yes.

It's insane. It's right up there with a fridge with a giant screen the size of your fridge with a camera so that it can show you what's inside your fridge. It's ludicrous. Why not just transparent fridge then? If you care that much, but I don't want to see what, I mean, it just, I can open the door and look, it's fine. Yeah. I think we did. I think you used the term over-index. I feel like WEF over-indexed on a lot of stuff that made them sound a little crazy to a lot of people.

I do think that the shift and I don't think it's like, again, I don't think there's like a necessarily a left versus right. It's on it. It's a different axis. Was it one of those guys that said, you know, it was like a people have chosen like pop populism, I guess, maybe, maybe that's. Hmm. I'm not sure. I'm not smart enough for all that, John. I just want to know which revolution we're in and if it's going to be good for me or not. The AI revolution, I guess. Right. Yeah.

Oh, what was the heading? Oh, this was a great heading about AI. Where was it? Oh, I think it was in this article that I mentioned before. where somebody had mentioned about AI being upset because Chinese AI plagiarized our U.S. Plagiarism AI. Don't plagiarize our plagiarism. Yeah. Yeah, there's a lot of that going on, I'm sure.

Okay. So I have actually a couple more clips. We can just go through some clips if you want. I like clips. They're fun. Yeah. Easy. This one is about AI. Builds on what? This is the all in. I had actually several kind of clips on the all-in, didn't you? I did, actually. That's exactly what I did. Builds on what Pincus was saying, which is, I think that this is an interesting window into the future of AI.

in the sense that I think what companies are internalizing slowly, that the first place that AI disintermediates is actually middle management. You thought that it was the customer support person. Maybe you thought it was the engineer. Maybe you thought it was the designer. Maybe you thought it was the product manager. I think those are less true. I think it's the middle manager. It's the functionary.

That basically is acting as essentially cartilage inside of this organization has less and less to do in a place where... AI enabled systems are making a lot of decisions on behalf of businesses. Pause that there. I feel like that's not the main narrative that you hear that much, but it.

If you think about it and more of his explanation, I can see that. I need to hear it one more time, but it sounded like he was saying something about the middleman. No, middlemanagement. He's saying AI is going to take out middlemanagement primarily. I'll keep going. And if you take a step back, how did that evolve? Well, it started because of what I mentioned a couple weeks ago.

What the software industrial complex will really be known for after it is dismantled and gone? So he's talked about this before. That's why he kind of breezes past it. But he believes that the software industrial complex is what he calls this. is really going to be completely taken out and reshaped, I guess. Is that it created...

All of these really dysfunctional org charts in a company. Meaning, when you are a system of record that sells something, let's make it up. Hey, I'm the best general ledger software. Okay, and a CEO says, great, implement that. And then all of a sudden, it's like, well, you're going to need a CFO. Okay, but then the CFO comes to you and says, well, you actually need a head of FP&A, ahead of this, ahead of that, ahead of the other thing.

Then the CEO says, okay. Then those people hire. And what happens is the software sprawl is really what created those jobs. Somebody else says, here's a customer relationship management thing. Okay, great. So now you hire a CMO and a CRO. Those people build infrastructure. So my premise about this is if you look inside of any org chart of any company, you can map those jobs.

to some clunky old piece of software that was sold to them. Maybe it was 10 years ago. Maybe it was 20 years ago. But that's why orgs are this bulky. And now when you have all of these new next-gen businesses that are ripping all of that software out, The middle management layers that used to manage that software are no longer necessary. That's why you're not hiring MBAs. And I think that this trend is only going to grow. So I don't think...

It's funny because, you know, I think we're so biased. We always just, you and I, we think about, you know, software developers and engineers and, you know, testers and QA and like all the stuff, but we really haven't focused that much on. Yeah, the management class that forms around these giant systems that companies have. I think about it now these days. Do you? Yeah, because I see it. Yeah.

There's a lot of management in between, let's say, what the business needs to be able to do their job before it gets to me as the engineer to actually build it. Yeah. And within that, everyone's making some kind of decision. It's like the telephone game. You know, I need this and then it changes to this and it changes to that. By the time it gets to me, it's 10 times bigger. Again, I hate to go back to this conversation I had yesterday because I was trying to explain it to my daughter.

How the bureaucracy is so heavy in these large organizations. Let's say I just wanted to add a field to capture a website on an account record. That's how all the database conversation got started. But in my case, and again, this is an extreme example, a very simple example that would take two weeks to put that field on the screen and get it into production because of the process has to have a user story.

It has to have tech design. It has to have QA test plans. Then it has to be built. Then it has to be tested. Then it has to be UAT'd. Our release cycles are every two weeks. So then it's got to sit there and wait for two weeks until, that's assuming it happens really quick.

Then it's got to wait till the next deployment cycle. Two weeks. Now, for much larger projects, that process makes sense. There's a lot of things to account for. There's a lot of things to coordinate. There's a lot of things to test. But when you talk about the day-to-day of, I just want one field. Yeah. You can't do that. Yeah. Less than two weeks. Yeah. And there's, man, there's, I feel like there's a lot of different things you can blame that on. Yeah. And, and everything you mentioned.

All of that bureaucratic process and those mechanisms, I understand why they exist. But at some point, organizations fail to, they cease the ability to... Be, like, pragmatic and practical about things. Like, when your organization takes literally a 60-second task... And turns it into a two week thing that requires, you know, 18 different teams, gates, processes, cards, issues, bugs, whatever it is. And that's. Again.

I'm sure you can make a case for that whole process and all that stuff. But I'm telling you, you can make that case all the way to your grave. If that's the way everything is run. Yeah. Well, even trying to, and here's where the bureaucracy still has thorns on it that'll snag you when you try to bypass it. You know, let's envision that whole process as a thorny bush and you're trying to bypass it. It's got these thorns and it'll snag you.

So we did try to bypass that process. So we could create a field in a day or whatever. Well, what happened is we had to have delegated admins. We had to have a place for them to make the change. that field needs to go into our version control system. So we still had to find some way of getting that information into version control. It still needed to be propagated to all of our lower environments. So we had to have a process for that.

So even though we could technically create the field and move it on, the machine had to be run in the back end. I get that, you know, because if we're going to have sources or the... Code is the source of truth or infrastructure is code. I don't know if they really use that term in the Salesforce space very much, but we kind of, it's kind of, in a way, it's infrastructure is code. Certainly metadata is code in a way. I mean.

You have to have that. I mean, if you go just went and created the field in production, depending on what your deployment technology is, it may when you have your next deployment, I see that it might see that rogue field in production and just nuke it. Oh, that shouldn't be there. And it just gets gone, you know, because it's not.

It wasn't part of the source of truth. Right. But when you look at back to what Shemath was talking about, you know, we've, we've clearly been in a, and are increasingly headed into a layoff cycle here. And. When you look at a lot of the just and this is just anecdotally, just who the people I've seen on LinkedIn and different places that have announced that, you know, they they got laid off. It is a lot of these. It's a lot of like this middle management type. It's not.

It's not as many software engineers and writers and graphic artists as I would expect to be. It's a different mix. And I'm not sure what to make of that other than maybe Chamath is kind of right about this. Well, I think the title ignores some aspects of what that person was doing and where they're at in their career. I look at my own career.

And I'm being pushed towards the management side of things, towards the team lead side, eventually to management. And I'm just going to be managing, what, a group of people? And at some point, my role is not needed. it's just like this natural progression where you get pushed out until you like hit the cliff and you drop off. Yeah. It just depends. I mean, if there is, if there's actually a need there and you're.

providing that with a lot of value, then absolutely. I mean, it makes sense, but no, I guess the point I'm trying to make is that in corporate America, your value and your pay, thus your pay is key to your role. And so in order to advance your career and to get paid more, you move up the ladder. You go from engineering level one to two to three or whatever the schema is. And then you reach the peak of that and you're like, well, okay, I guess I got to be a team lead now.

So I get, I move into that. I would rather be a coder, but I need to make some money. I've got a family and I've got responsibilities. So now I'm a team lead. Okay, I do pretty well at that. Let's move on to the management. So it's these people that want to be the developers, want to be the engineers, want to be the writers, but they're pushed forward as they advance in their career to be able to justify a higher pay.

The system is not designed to reward good engineers to stay good engineers. They're pushed into management, and then they're told... you're just a paper pusher yeah go away you're just costing us extra money there's not a lot of companies that have that full career path for like scientists and engineer types

Usually the ones that do, they'll, they'll have like the name, maybe like, like technical fellow or something. That's like when you've hit, like, I mean, you're like, that is the peak of that. And it's not, and it's quite high and like respected in the organization. Um, there's just not a lot of companies that have that. I mean, I suspect, you know, I know some microsystems use, I suspect that, you know, Apple, Google, um, maybe meta, um, maybe these newer companies like open AI maybe have.

They have so like, cause they're, they're so truly technology focused companies that is such a core part of their, of their organization and the, the, the core competency, I guess, of the organization that. it's it's a it is actually highly valued when someone sticks in that builds on what i think i just lost my spot i don't know what i said um anyway you know what i'm saying yeah

Yeah, I have to find out where we were here. Is that it created all of these really dysfunctional org charts in a company. Meaning. When you are a system of record that sells something, let's make it up. Hey, I'm the best. You actually need a head of FP&A, a head of this, a head of that, a head of the other thing. Then the CEO says, okay. Then those people hire.

And what happens is the software sprawl is really what created those jobs. Somebody else says, here's a customer relationship management thing. Okay, great. So now you hire a CMO and a CRO. Those people build infrastructure. My premise about this is if you look inside of any org chart of any company, you can map those jobs to some clunky old piece of software that was sold to them. Maybe it was 10 years ago. Maybe it was 20 years ago.

But that's why orgs are this bulky. And now when you have all of these new next-gen businesses that are ripping all of that software out, The middle management layers that used to manage that software are no longer necessary. That's why you're not hiring MBAs. And I think that this trend is only going to grow. So I don't think that this is a commentary on the people. The people are probably quite smart.

But I do think that this is a warning sign that people should not go and pursue these degrees because I think, as Pink has said, you're not taking the trade of least volatility. You're actually taking on a lot more volatility than you probably thought you... shouldn't be taking on by going to a place like Harvard or Stanford. Freebert, any thoughts as we wrap here about this MBA chart? I think the collapse of the MBA program could be the beginning of the unwinding of the higher education market.

why similar i think this is the easiest one to discard first in an age of ai and automation and self-learning and i think that I mean, I've been thinking a lot about, I've, I've young kids, right? So my kid, my oldest kid is seven years old. I had to think about that for a sec. Um, so many kids now. And, um,

Isn't it great, by the way? Yeah, now they start talking about their kids. I go through this all the time. I don't know. Lab when I was there and doing other stuff. But those were like on-the-job experiences. I do think that the social experience and the... Practical experience can be gotten outside of the college infrastructure, this educational infrastructure. Oh, I cannot tell you guys how much time I just had a great idea.

I just had the best idea ever. The all in MBA. Let's start our own MBA program. I think that's the whole point is people should get back to kind of an apprenticeship type model or a self-learning model. Everyone's got their own personal tutor now. That's what I think people aren't realizing. And it is crazy how good these LLMs are at tutoring. I mean, I think that's a profession that. Wow. I mean, I know there's not like a huge industry tutoring, but it's, you know, it's a thing.

These LOMs are outstanding tutors, especially now as they have the ability to just like to look, you know, they can just look at a geometry problem or something and walk you through. It breaks it, you know, helps you identify what type of problem it is.

And then explains to you like which of the principles or proofs or theorems apply to it. And then walks you through step by step, how to work it. And it's just, and that's just, you know, one subject. I mean, there's just, it seems to be, it seems that they seem to really are excelling at just.

tutoring at all kinds of subjects. It's crazy. Yeah. But yeah, the whole like MBAs and just in higher ed in general, like how much when you can really self-learn a ton and it's just doesn't. Yeah. Is that, is that the. That's interesting. Interesting. I don't know. I mean, because I think, you know, this experience of a university is it's more than just just the education you get. It's something that goes into it. And so I'm real. I struggle with this. I think a lot of companies put it.

have recently been putting a lot of priority around having more experienced people on staff because they don't want to have to train a lot of people. You know, they have a need to fill, they want to move fast, and so they want that experienced person.

And I think for people just coming out of college who their only experience is in the education side of things, it's been difficult for them. Yeah, I mean, we apparently have a really big backlog of STEM graduates. Like, there was a professor the other day. one of these top kind of computer science programs saying even his top graduates are having problems just getting interviews. Yeah. But, you know, we're, I think something like 40% of...

of STEM graduates from college are not able to go into STEM. Speaking of over-indexing, I feel like we really, I mean, the whole learn to code, which was just such a... but dismissive thing to say to someone anyway. Um, yeah, that was that, that aged like milk. Mm. Let me see if they have anything else here. Is anything you want to learn with AI? Have you guys heard about what Johnny Ives has done with his kids? I'm pretty sure they skipped college and they just went to work.

directly with him and his design firm. Exactly. That's because their dad's Johnny Ive. I don't count. I'm taking him out to lunch because I want to interview him. Exactly. What better professor than Johnny? I know. That's what I'm saying. Industrial design. Of course, that's what his kids did. But yeah. Interesting. By the way, your daughters and my daughters want to start a podcast together.

I think they just continue rambling after that. We don't do that either. No, we don't. There's a very professional produced show here, John. I do think that's an, that's an interesting, uh, I don't know the right words. I'm struggling with words today, but I think that's an interesting way of looking at the future of things is that we probably should go back to more internships and things like that where apprenticeships.

where you can just learn firsthand. I think it's a much better way of learning. I think companies will get much better value out of it than sending these kids off to go into debt for so much money. At least on top of that, what would that do to our economy? If you go in and you're getting, maybe you're still getting, maybe you're getting paid peanuts, but you're learning a skill and you're growing in the advance in that, but you're advancing without debt.

You know, you're getting paid peanuts, but you don't have the debt. You don't have the $100,000, $200,000 in debt that you have to crawl out of. Your life can get started sooner. Yeah, I think I was just about to forget exactly what I was going to say. I was going to make a very similar comment. I think I was going to say something more along the lines of how much, how the cost of college education has inflated so much, like far, far more than almost anything else in the economy. It's like...

Something like 10x inflation or something ridiculous. It's way too expensive. I don't know what the problem is with these universities. I don't know either. It seemed like we've pushed more kids into college than we ever have, and yet it got more expensive. Well, we also federally funded all the financing, and that just creates a ton of demand. And what happens when demand spikes up? Prices go up.

I've also seen charts. Again, I don't know. I've never done a deep dive or how accurate these are, but let's show how the growth of administrators versus educators in college. That's a problem. Yeah. I mean.

assume those are somewhat accurate but i don't know yeah you see so much nowadays especially on you know on all the social networks it's it's hard to know what what's real you have to really You have to really check a source if it's something you're going to actually form opinions on and make sure it's from a source that is legit or someone that you know is like someone you trust that has had a track record of not spreading.

wrong stuff and not getting caught up easily caught up in other people spreading wrong stuff. Some people get kind of sucked up by the hype of something or the urgency of it, you know, and they kind of, they forget to. simmer down and start in and maybe check facts first yeah i've been caught up in that before yeah oh it's happened to everyone yeah some of it's not even related to the thing i think sometimes you just

You're frustrated with so many other things you want to find a win. You just rally against something. You pick something and you try to make it your win. Okay, I have one more clip. This is from Coder Radio. And they, I think we, did we talk about this in the last show? Mark Minioff, you know, when Ken, I mentioned on that one podcast he went on about they're not going to hire software engineers. Yeah, I think we touched on it. Kota Radio covered it. And.

They talked about it for like 10 or 15 minutes. I just think, I think I just got the first five minutes cause they really went into, it was long and went into all kinds of other stuff about Salesforce. Yeah, exactly. That's gonna be a title. We don't do that. Okay, so Salesforce's CEO made some headlines since the last episode when he was quoted saying he's hiring no more software engineers in 2025. They've had such big boosts from AI. This is such an audacious statement that I...

I actually thought it'd be better if you heard it yourself, because then you hear it from the horse's mouth. So I'll play just a bit of Salesforce's CEO, and then we'll talk about it on the other side. Number one, we're doing our business plan for next year, for 2025 right now. And we're not adding any more software engineers next year because we have increased the productivity this year with AgentForce and with other AI technology.

we're using for our engineering teams by more than 30% to the point where our engineering velocity is incredible. I can't believe what we're achieving in engineering. And then... And this is... I think he's a little sincere question here, but does that mean that we're going to start like some of these things that have been deficiencies or bugs or just outright broken stuff for so long that.

those will actually start to get fixed because they're having such a massive engineering boost from AI. I would think so. Yeah, I would think so too. I mean, I'm, I wonder, I mean like, you know, spring, spring releases already happened, but you know, let's watch the summer release notes and see.

See what's happening. It just explodes. Yeah. With all these things that they just haven't been able to have resources to allocate for these things. Right. That'd be interesting. We will have. I would totally welcome that. Support engineers next year because we have an agentic layer. We'll have more salespeople next year because we need to really explain to everybody exactly the value that we can achieve with AI. So we will add another probably 1,000 to 2,000 salespeople.

the short term. So going to hire 2,000 salespeople, but no software engineers, Mike. Yeah, there's a lot here. I'm going to take the simple take first. This is the kind of thing you say publicly so the investors and the analysts hear it in Booster Stock Price. The non-cynical take is, you know, we touched on it.

a little, I think a little awkwardly last week, the plight of junior developers looking for jobs and not getting them. It is. I, I, that's again, that's, I think that includes those, you know, new CS grads. Or software engineering grads. Yeah. It's going to be a tough path for them, I think. Again, maybe this, was it Chamath or the other guy, Friedberg, I think his name is.

Maybe we do go back to more of an apprentice model. I don't want to call it intern because intern is a different thing, but a true apprentice model. Otherwise, I think it's going to be too difficult for people to get started. In 12 months or even six months, we'll probably have enough data to prove this right or wrong.

There is going to be a less developer hiring or coder hiring, whatever you want to call it. And that's going to be mostly junior devs that take the hit because, I mean, we don't need to rehash last episode, but... Junior devs are very expensive, tend to leave once they're useful, right? And tend to cause problems, right? We've all, I'm sure everybody's had.

I'm sure all of us have been the junior dev for the new guy on the team who blows something up. So, yeah, I don't deny Mark Benioff's point that they're... is going to be less demand. I think he's making such a dramatic statement that we're not hiring any coders at all for a year for the benefit of our good friends at Goldman Sachs and Bank of America doing their work.

Yeah. And I think to be fair, I mean, I think that that was just bluster, like, you know, almost just like corporate propaganda a little bit. It's not that they're I mean, I know for a fact they're hiring software engineers. So we know Benioff. And they have, it's not like he doesn't have a track record of kind of fudging the truth a bit. Yeah. And maybe the first, first app built on lightning. Oh yeah. You know, that kind of stuff.

I will say, I don't think they plan on hiring a lot of software engineers in the United States. That's pretty clear. I don't know that there's a lot of Salesforce developers out there. to be honest. Salesforce is really not hiring a lot of Salesforce developers. They're hiring software engineers who can build in Java and Scala and whatever Salesforce is doing nowadays.

Probably a ton of Python and AI stuff and whatever. I mean, they're really not, they're not that much. I mean, except for like their professional services and customer support and things like that, obviously. Yeah. Yeah. Now I got caught up in my own view of that. But yeah, yeah, you're absolutely right. It's, it's a different technology. So analysis is different. Yeah.

I got a semi-hot take on this one, and that's just because I've been watching Salesforce as they adapt to the AI train and try to get on board. And Salesforce came in a little late, right? They came in after Microsoft. Microsoft and open AI. We got a lot of attention. They came in after Google and they came in super strong. Oh, that's not true at all. I mean, Google's some of the first champion championing AI with Einstein.

I don't think people consider that AI, though. They might not, but Salesforce did. Whereas Google and Microsoft have actually had real, I mean, large AI efforts. And Salesforce has had... They've had teams on this, but nothing close to the scale of Google, like their DeepMind group. Yeah, but these are nothing near close to what Microsoft had as a major investor in open AI. Right. I mean, nothing even remotely close to like that. Sure. But in terms of just applied application of the technology.

A general purpose AI is a little less valuable than say something purposely built. You know what we forget about in this space now that was supposed to be, it was supposed to take this whole space and just run with it. What? IBM Watson. What happened to Watson? Turns out it was a big freaking lie. Kind of like Microsoft's Xbox lie. I don't know about that.

They did a tech demo of when the Xbox came out, and they had the camera system of a character. I forgot what his name was, but a character that you could interact with. And everyone was really excited about this because you think about the different type of gameplay you could have with something like that.

But you would walk in the room, it would recognize you, you could have a conversation with it and everything, and it was all smoke and mirrors. I mean, so many demos are. I see people all the time doing demos. Well, I think... It gets exposed during the demo that it's actually Figma that they're demoing. I think there was a certain amount of that with Salesforce as well, with Einstein and just the way they marketed and packaged the concepts of what it could actually do vis-a-vis the...

Matthew McConaughey commercials. Yeah, I mean, I think on these big software companies, they do that. They're going to start generating demand ahead of the actual availability of the product. But, you know, Watson, I feel like, and I think, it seems like they're still pushing it.

Well, they just they keep they keep reinventing it or trying to like when they when they get up stage, they're like, oh, yeah, well, we're going to do that, too. Now we're going to use LLMs. Now we're going to do this. We're going to do we're going to have operator an agent mode for, you know, it's like they just keep.

I don't, does anyone believe them or do they have any credibility? Well, I want to correct two things I said. One is I don't, I don't think general purpose AI is less valuable than purpose built. industry specific ai i just meant that the early days of the ai were just these general models that people were talking about they didn't really have any utility no one knew what to do with them but salesforce was

advertising and championing it early on. They weren't late to it. In fact, I think they were too early on it because it was vaporware at that point. So that's the only clarification I want to make with what I said. Yeah, it was really the, you know, the transformer aspect of LLMs. And then I think it was open AIs, their kind of post-training regime they have.

That allowed them to build, you know, what we now know as chat GPT. When those two things kind of happened, it seemed like it went fast. And then all of a sudden the demos come out and they make it available to people. And that was the oh shit moment for every enterprise software leader around the world. It makes everyone's efforts up to that point look silly. and pointless and how do we incorporate this new technology instantly quickly into our products sure i agree okay

We're compensating for the fact that they were late on this and their CEO. I played it on live streams months ago. Their CEO just was saying some ridiculous stuff. I guess he's talking about Salesforce still. Just to really try to position Salesforce as just overcompensatingly ahead. We're so far ahead, everybody else. It's crazy, bro. And this leans into that narrative while also... avoiding the fact that they must be suffering financially or something.

Because whenever you're over-investing in sales and under-investing in building the product, that's not a sign that the product's doing particularly well. And you're trying to hire staff that eventually will pay for themselves ideally because you don't have revenue. So they're hiring salespeople. And we don't know what the conditions are on those hires. We don't know if it's, you know.

Commission-based whatever. And I think Salesforce is trying to use this AI excuse, this cover, for why they won't be doing any hiring. And I'll give you an analogy. It would be like if I... If I shut down the Coda radio program, you and I were like, we're done. And what we're doing is we're leaning into agentic podcasting. And every week we're going to have a curated AI podcast that plays all of the developer related and business. Thank you.

and one in a mid-roll, and one at the very end. And we're going to make a lot more money off that because it's going to have zero cost for us to work on it day-to-day once it's set up. The dynamic ad insertion takes care of itself. It's a bid-based marketplace. And we don't have to sit here and spend our time doing it.

It sounds to the people that are super excited about AI like, oh, look at this product you can create. But what it really is is a desperate company that can no longer afford to maintain its product line and is transitioning to AI-generated slot. It puts a bow on it. That's interesting. Salesforce's growth was never based on their internal product because they grew by acquisition.

And they grew fast by acquisition. It's only recently that they've been integrating and creating what they call the core products and taking all these add-ons and making them core. I mean, I kind of, I mean, I definitely, I mean, you're, I think you're right to a large degree, but I mean, they've, they've had core growth in their core clouds. They've had prior to acquisitions for a long time now. I mean, their growth has slowed some and they have, I think, have used acquisitions as a way to.

to add more growth, but they've, I mean, they've definitely grown organically a lot. I just think for people that don't follow Salesforce space real closely, they don't realize how sales and marketing driven a company Salesforce is. And they've always had a big sales department. They've always had a big marketing, like bigger than its peers. And that's one of the criticisms from financial analysts and some of these activist investors.

And just in general, people say, oh, Salesforce is so great. Why is it so expensive to sell in all this? It's that same argument. But this is... For us that have been in the space, we're just used to that. I mean, Salesforce is just a very sales and marketing-driven company. It's the biggest part of their company. It has probably the most number of employees. It's definitely the most expensive.

And so for people that are outside the space to kind of just swoop in and make it seem like this is something new or something, this is par for the course. I mean, they've done pretty well with that model. I guess that's a critique of what he was saying there. It's like, of course they're going to hire more salespeople. That's what Salesforce does. They're going to sell this hard. I think I was trying to say the same thing.

But he alluded to, I don't know if he heard this. I don't know exactly where it was. I'm not going to find it. But, you know, it'd be like, oh, yeah, he said it'd be like if we decided to close Coder Radio down and replace it with like this eight-hole, you know, AI podcast. Well, last week, I think it was last week, Coda Radio announced that they are no more. Not a joke? Not a joke. Rest in peace, Coda Radio. He did say, so there was just like a three-minute episode that I listened to.

where he just announced it and the the guy the co-host guy the other one there he said they were working on a deal for he was gonna he's gonna sell the podcast to that guy and he'll run it on his own outside of because i think um They have a podcast network. It's a network. I mean, it's not a network. It's a company that owns like a half a dozen podcasts. Is that so-called a podcast network? I think so. Okay.

And so I think he's going to try to sell the podcast from the network to that co-host and he'll carry it on. That was the idea. I don't know if that'll happen or not. AI's already coming for the podcast, John. AI killed the podcast star. Did you see that OpenAI, even, so they have this, that was a couple months ago now, they announced this pro plan. So they had the plus, which is 20 bucks a month.

And they came out with a pro, which gives you access to more usage plus this new model. I think it's the reasoning model. What's that one called? I forget now. Anyway. This is the 03. But even at $200 a month, they're losing money on most users. Because it takes so much compute at inference time. to run these things that even which is like typical usage at $200 a month, they're losing money on you. Wow. But then we have deep seek, right? The Chinese.

An organization that released, what is it, was it R1 was their latest? That also kind of, I mean, dude, stocks dropped. You know, panic. Because they supposedly. did for like five million dollars what it cost open ai like 500 million dollars to do like orders of magnitude cheaper because they found smarter ways supposedly and then as the days and weeks went on

I think people have learned that that's probably not the case. I don't know why people didn't already take it with a grain of sand. I mean, it's coming out of China. I think a lot of people were taking it with a grain of salt, but there was also still just there was a lot of panic. It just shows how brittle that market is. I think it's very brittle. I heard someone say that, I don't get the exact quote right, but LLMs are the quickest devaluing assets in the world right now.

You know, you spend several hundred miles, several hundred million dollars to develop one, and it's just got such a short shelf life. And in the way you, not only that, but the way. But you've built it, like all the hardware you've invested into, that may be that brute force way of just throwing insane amounts of hardware. By the way, energy too. This makes like...

this makes Bitcoin miners look like, you know, eco alliance, you know, partners, net zero alliance partners or whatever. I mean, um, yeah, it's like a bad joke. I'm like, I'll see. But yeah, and so it turns out that right now, yeah, it takes all this hardware and all this electricity and so much, right, just to give you a basic answer to some question. And I think everyone said, and maybe this has already happened.

I mean, that's what DeepSeq was kind of telling us, but maybe there's just way better ways that are much, much cheaper to build equivalent or even better LLMs or whatever the next. I'm kind of waiting. We're like, what's after LLMs? I mean, LLMs are just...

Listen, when you throw enough computer at it and it, when you, when you get it to regurgitate its output, you know, and go into like reasoning mode or whatever, it can do some cool stuff that looks smart, but it's still, you got to watch it though. Cause it's, it'll just.

fail fantastically either making stuff up or just convincing itself it's right about things that are not true so whether you're having it generate your website for you or you know come up write a research document for you that kind of stuff you really You can't trust it because they still hallucinate. They still hallucinate. They still have these fundamental flaws of LM technology. And maybe they can keep smoothing those out again with like the post-training steps.

the reinforcement learning type stuff. When that technology, as that improves, maybe they can get the accuracy percentage so high that it doesn't really matter that it's... Still, at its fundamental core, it's statistically putting one token after another. I think it's natural to see the technology improve.

energy usage and hardware requirements to go down i mean that's just the way computers have been going for forever that's true but we're not talking about just the cost of hardware going on we're talking about the the need for the hardware to dramatically drop To achieve the same results. And that's why, you know, NVIDIA investors were very scared. It's like, oh crap, what if they don't need all these H100s anymore? It's going to happen. They're not going to need them.

There will be a lot of kids with some badass graphics computers, gaming computers, if that happens, if all those things hit the open market. I don't know if those are actually, can you use those for computer graphics, those purpose built? do you know nvidia machines i don't even know if you can i mean i guess are they i guess there's still technically gpus but there's still quite a few bit of hardcore gamers that really want to turn up the settings high on you know high quality high dps but

I've learned after buying my son his computer, he doesn't care about the graphics. He wants to play the game. Some of the games he plays are these like Roblox style games that... graphically are trash, but they're enjoying them, they're having fun, they don't care about the graphics. Well, let me tell you my tale of two children. I have one kid that's just like your kid. He has a nice machine, but...

95% of the time he's watching YouTube videos. Now he does tend to watch very educational YouTube videos. He's very much a nerd. Oh, not mine. He's watching these jerks that sit there and scream at the screen all day and he's laughing it up. that you watch. He watches those sometimes, especially if he's watching them play a game. Yeah, they're playing games, but they're over-exaggerating these game plays. But other than that, so it's that, and it's, yeah, it's like Roblox and these just...

low-fidelity graphics games. My other kid, on the other hand, He is so incredibly sensitive to performance. He is constantly trying to tweak or getting upset at something because something's not performing quite right. He's always wanting to upgrade his hardware. It's never enough. Different personality types. Did you want to... I know we have SLDS 2, which is in beta now. I think with Spring... Are we Spring 25? Yeah, Spring 25.

I think new orgs and maybe new professional orgs, whatever they are, already have the new UI. They've had it. Really? Yeah. Well, let's go over... what this is for people because and i haven't paid much attention to it i assume a lot of people have not paid a lot of attention to it um oh it's now available to orgs that opt into the cosmos ui theme never even heard of that that's i'm out of touch

So what's new? You can build adaptable user interfaces with themes ready for future updates. I'm sure they would have said the same thing about SLDS 1, though. Use the SLDS validator to ensure your LWC components comply with SLDS 2 guidelines. I wonder if Salesforce will use that on their own products. No. Yeah.

Number three, there's a validator that suggests fixes for a smoother transition from one to two. Okay, that's cool. Key changes for developers are, and by the way, this is, I'm not going to say this guy's name. Never mind. I can't say it.

keychains for developers. SLDS2 updates base components with, or it updates their DOM structures. So if you've accessed, if you've broken the the gold pierce the veil yeah and and reach down in there um you gotta you gotta watch out you're probably gonna have some broken stuff um so how to get started you enable slds2 validation in vs code Interesting. I'm sure Illuminated Cloud has its equivalent of this. And then you review your annotations, revisit any SLDS, validate, ignore.

validate, ignore annotations, and bring them into compliance where possible. So why upgrade? Well, it feature-proofs your design. It ensures seamless compatibility with the new Cosmos theme and avoids surprises as Salesforce rolls out updates. Okay. Have you looked at Cosmos?

Yeah, it's more bubbly. At least it gets rid of that. Oh, I did see that. I saw that. Yeah. Yeah. It's somewhat subtle, though, it seems like to me. Well, they got rid of that blue background, so you're left with a standard solid color background.

Yeah. Oh, it doesn't have the trailhead background. Yeah. Yeah, the topology map. That is a little distracting. Yeah. Yeah. I mean, I get it. It's the Salesforce's theme, but everything is still trending to... simplicity i mean it's nice if they have a new theme that's a little bubbly and rounded i mean that's fine but can i get a setup that's consistent i know and that's why it's so iframe to hell that it's broken i can't like

link out the way I want to or use things. Profiles is broken for me. Anyways, no rants. I like it. It's cleaner. It's simpler. It looks like it's a little bit more pastel-y. in some of the color schemes, which I think breaks some of the contrast guidelines for accessibility. So there's some things about it that I'm...

that seem like it's a step backwards in terms of like accessibility and stuff, but maybe they'll have like a high contrast theme or something like that, that will cover it. And maybe that's their plan, but I think it does break some of those rules. Like accessibility rules? That would be surprising because Salesforce is... No, look at that. I think it's on one of your links that has the screenshots with SLDS1 and SLDL2.

Look at the difference in the high contrast in the path bar and the button, that dark blue, and then look at the contrast between that and the text and everything else that's on the new version. And that will tell you what you need to know in terms of the contrast. After. Yeah. So after is on the right. Yeah. I mean, I think getting rid of that topology map is probably the biggest win.

Yeah, that was distracting. It's a little unnecessary. I mean, there's more contrast on the menu bar. I think it's... No, because the, you don't have that gray background anymore. So you've basically left with the line and then the text is colored. I think I'm looking at something different than you. Oh, I see. Actually, I see, I do see things that are less contrast, but that's not what I'm looking at. Oh. Yeah.

A little bit less contrast. Like they're using like dark grays more than blacks and stuff. So a dark gray is going to be less contrast in a light gray background than black on a light gray background. But it still looks like it's plenty of contrast to me. I mean, I don't have side issues. I mean, I like the look, and I think from my eyes, it's not bad. I like it, too. It looks just like a nice, subtle refresh to me.

But there was a reason they changed some things from the original SLDS 1. Because when it came out, it wasn't as compliant with accessibility. And they did adjust the contrast on a lot of things for that reason. But this looks like they took a step back from that. And I also want to know, like, I don't think they're going to say this, but like what truly what screens is this going to apply to? All of them.

All of the UI will look like this. Set up? Hopefully. No, I don't think so, John. I mean, I think it's just like the primary. My guess would be the primary. Well, the primary like on platform apps. So sales cloud, maybe service cloud. I see what you mean, the cloud. I mean, the other ones, all the marketing things are on different. They're not going to get this right. Those are totally different paradigms altogether.

A lot of them have adopted the design system. So I'm assuming that they'll be asked by Salesforce to update to the SLDS too. I mean, I think that's huge for Salesforce if they're able to pull that off because that's, you know, it is.

I mean, because they talk about their five criteria. I'm looking at a blog post they did on this. It needs to be seamless. No awkward transitions, interruptions, or disparities, and it's connected and unified. Well, that means you have to do it across all your products, right?

You know, responsive. It needs to be quick to respond or react. It gets smart with user input and adapts to diverse personas, languages, cultures, and environments. Okay, I'm not sure what all that means. It needs to be intuitive. Okay, sure. Trusted. Okay. And delightful. It needs to spark joy, it says. It doesn't need to spark joy. Very Marie Kondo. I like that software sparks joy in terms of just visuals, but I know it gets in the way of productivity. You think so? Yeah.

Do you think that just the nice, tight, compact grid of input elements is more productive? I'm a minimalist in a lot of things, and I think... But the way people... because that's the way customers want it. They want everything on the screen because they think they need everything on the screen. Yeah. And it just becomes way too much. Yep. That's true. So you end up compacting things to fit things more onto the screen. And it's just. I also saw a, this might be a little duplicative, but.

Peter Chittum did a blog or whatever this is called, an article for Salesforce bin. And it was six Salesforce spring 25 updates for developers. First one was. Local dev for lightning experience, desktop and mobile. So I think there's some probably caveats on this. I forget where those were. I was reading about them somewhere.

Um, do you, are you doing much of that? Have you done, have you done this local dev? Cause this, this is not, I feel like everything's freezing up here. Oh yeah. My Salesforce bin tab is just dying. Come on, Ben. Let me try to reload this. Okay, well, we can skip that. Number two, SLDS 2 is coming. Yeah, the next UI theme, Cosmos. I guess Cosmos is beta. Okay. Okay. Three. Apex gets zippy. Wow. Apex can zip and unzip files, John. Yeah.

consider yourself lucky and privileged, I guess. I know. Okay. Wow. That was one out of the five or six. Okay. Number four, you can evaluate dynamic formulas in Apex. Again, nothing new. I don't know if it's just that it's GA now. Yeah, you can have a, just a, you know, and I guess define a formula. just in a string, in Apex, and can evaluate it at runtime. That could be useful in certain situations, yeah. I think so. Although, I have to think of a really good use case.

Why wouldn't I just use Apex? For dynamic things, like maybe a dynamic mapping or formula editor for something custom. Yeah, that's true. Like an app where you let... I think product rules or something. You users define formulas or something, and then you can, yeah, that's, okay, that makes sense. I like that. Thank you. Number five, Apex Guru is generally available.

this tool provides code analysis that uses machine learning to identify potentially problematic code. Is that part of their, how they've kind of repackaged, what is it? Cloud or ChatGPT, I forget one of those, to provide... Yeah, it's like a co-pilot for Salesforce developers. What do they call that? Einstein for developers. Is that part of that?

I don't know. Okay. I'm not using that. I'm using Copilot. API versions 21 through 30 are going away. Let's see. Is that like rest and soap APIs, that type of thing? Oh, yeah, there's a big push on the API stuff, the versioning. We've been doing some analysis on our end. We're finding a lot of managed packages that are still referencing older stuff. Good luck with those. It's fine. We're finding that we're probably not even using them. They're just installed. Also, let's see.

Oh, say goodbye to instance-based domains. So everyone has like, you know, mycompany.my.salesforce.com now. But I guess some people still have, they log in, it's just like na1.salesforce.com or whatever. And then Salesforce functions as being completely retired. It's gone. Yep. What could have been, um, I also had a, I did a summary of spring 25. Um,

with the help of chat GPT. So all your LWCs are going to have to specify an API version. What? I'm surprised I didn't already. I mean, doesn't all metadata require API version anyway? I mean... You can't create an Apex class that is not pegged to an API version. Yeah. Okay. I thought I already did, though. Maybe not. Again, the real-time preview for Lightning Apps. That's cool. Yep, zip files, dynamic formulas, scale.

Concurrent long-running requests. The number of concurrent APEX requests your org can run scales with the number of licenses you have, helping large orgs handle more simultaneous long-running jobs. Oh, that's it. Is that music to yours? Yeah. Okay, good. You can pause and resume scheduled jobs. So instead of having to kill them and then later reschedule them, you can just put it on pause. That's nice. You can reparent.

restrictions for master reparenting restrictions for master detail so enforcement ensures that if you have master detailed relationships with restricted parenting and available your apex code must handle any reparent attempts properly or throw exceptions Is that restriction they're talking about just the fact that you can say you can't reparent this? That's what I was thinking. Yeah. I'm not sure. There's a new database access debug log category.

Track database access in more detail to troubleshoot performance or identify bottlenecks and queries. Okay. Could be useful if you're having a certain kind of problem. Salesforce is retiring. My domain login URL enforcement. Salesforce is retiring older approaches in favor of my domain-based logins. This doesn't mean you can't log in at login.salesforce.com anymore. You have to go to your...

Again, you have to go to your My Domain to log in. Bulk API version 2 query platform events are available, so you can receive platform events when Bulk API v2 query jobs reach certain stages. Those must be some big query jobs if you need events to let you know when it's progressed from like one step of a query to another step of a query. I can't imagine that, but I'm sure someone has that kind of problem.

uh let's see open api documents for s object rest i didn't make you do this you can generate open api spec which is open api is like the new name for swagger um yeah you can create open a or swagger specs for your org for your s objects and stuff interesting so people that for platforms that can consume swagger yeah specs you know yeah that is didn't know that

Yep, modernized SLDS 2. Ongoing updates to the CLI and Visual Studio Code extensions that improve project setup, debugging, and deployment workflows. Data Mask now supports Einstein-generated custom libraries to mask sensitive data more effectively. Okay. You can also configure it to run automatically after every sandbox refresh, which makes a lot of sense. I got to look into that. Yeah. And event relays, and you can now package event relays. I've never done event relays. Okay.

And then agent force for developer use cases. The new generative AI framework, formerly Einstein Copilot, has dev-focused enhancements for building and testing AI-driven assistance. If you plan on integrating or extending agent for us in your custom apps, you'll find new tools for batch testing, agent versions, and direct invocation from Apex or Flows. All right. I did. We were asked to give a shout out to. So, yeah, for the for a guy named Zach Frazier, Zachary Frazier. So he says that he.

I was looking into this open. Remember the open force thing? Okay. Yeah. Looked into it. He's like, Hey, it looks like this is not being maintained. Didn't get an answer or whatever. So he built a new thing called foxygenic.com. And it looks like, yeah, it's just like a kind of a website of open source Salesforce related stuff. Looks nice.

Anyway, also Zach's buddy Eric helped him do it as well. He helped clean up the UI and bootstrap the AWS part of it. That's pretty cool. I poked around in there. Oh, did you? Okay, cool. And we'll have to try to remember to put that in the show notes. Foxy jam. Um, yeah, John, you've got, I see you added some topics on. I've just been blown through mine. That's fine. I'm ready to wrap it up. You always say that. And this is what you do to me.

Oh, I already talked about a lot of them. Okay. I mean, there's a few that we touched on. That's true. Yeah. Okay. Um, I didn't want to say, it sounds like we're going to probably try to do it. A GDS meetup at TDX. Oh, cool. I'm guessing it'll be on the last day again. That's usually what we do. So I would just say check Slack. I think we usually put it in the conference. conversations channel. Yeah, I'm guessing the last day. There's a new place that took over the McKellar bar. Very similar.

And it was pretty, I liked it. So maybe we'll do it there. I know there's a few places around there that we have been using recently, but that one worked out really well. But yeah, just check Slack. If you're on the Slack. then you need to go to www.gooddaysirpodcast.com. Don't forget the www because we have infrastructure issues. And you click on, I think community, right? Yeah. Community.

And you can join or you can get anyone else to add you that's already in. But yeah, join the Slack. Stay in touch. And to that, I say good day, sir. Oh my gosh. You get nothing! You lose! Good day, sir!