Episode 578: War for Talent

Coté (00:00.866)
Now I didn't really understand. I think the technical term is the reporting that I read around this and the reporting on the report the blogging on the reporting. But I think maybe last week I finally got something right, which is if if I read and also had my robot summarize like why the the fable thing happened, it sounds like the clearest explanation as an AI would put it for you, is that the the the Dario guy.
doesn't know how to talk to r to to the administration, which I assume means like some Trump whack jobs. And and then that got that got that all messed up. Now, I think, you know, as I was saying, you know, you got last week I you got you got this you got the AI is gonna kill the civilization, but we're still working on it, people. And then you got you got the Department of War, the people who would rename the Department of Defense to the Department of War. And like, you know, what could go wrong?
Matt (00:51.448)
Yep, can't stop.
Coté (01:00.406)
And so it's it seems like maybe that's what it was. It was wasn't there some I saw some reporting there was some worker who someone was an extreme Democrat or something. And then and then also she was someone liked her who Trump didn't like, and then probably and then and then and then some someone in the administration was probably just like, Yeah, fuck those guys and then just like did the whole thing.
Matt (01:22.008)
Yeah, I I I I I don't know how extreme they were. I think their their the thing that kind of got them in trouble was the fact that they they recognized that the twenty twenty election went off without any interference and and they happened to like vouch for that.
Coté (01:35.907)
Yeah.
That is quite quite the grudge there to to hold for a long time.
Matt (01:40.782)
And and then and then the NSA promptly said, Give us back Mythos, right?
Coté (01:46.208)
Yes, yeah. J just just like all AI people who actually use it. The ones who are actually using it, they're like, I would like to keep using it. That is what I want. And then and then the ones managing it who don't use it and just writing Sharpies and things, like they apparently are not not not cool with it. But that
Matt (01:54.499)
Mm-hmm.
Matt (02:04.301)
They can use the Chinese models.
Brandon (02:05.084)
But but does all this make you more worried, less worried, or neutral on AI? Because it kinda comes back to like all the things we talked about. It's like, well, in the end, it's the people. It's always the people that make everything complicated, right? It's like like this whole thing does seem I mean, all the reporting is right, is it's really just people fighting. It's for whatever reason Amazon reported it, you know, to the government. Feels like they were trying to curry favor. I don't know, who knows why they reported it. It's still not clear.
Coté (02:15.307)
Mm.
Matt (02:23.221)
Yeah.
Brandon (02:34.406)
Then you kind of get this whole thing. And the reason that the administration is so mad is that this person did this, you know, reported on the elections like many years ago. So this is like all past. Like all this is like in the past. And it's just like, so it's just the people fighting, right? It's nothing the technology is not nothing around the technology has like universally changed in the last few months. It's just people and motivations has changed. So does that mean, well, it's just another one of these examples of it's just a a people problem, or like, no, like we have to be like really worried about the, you know, the the potential.
uses of this technology.
Matt (03:05.557)
I I think the people problem aspect is what makes it worse. 'Cause 'cause it would be lovely to to be able to
Think about AI and with with no context. You're like, you know, I've got this technology that doesn't have any biases. wait a second. You know, this technology is from somebody who, you know, I I'm not even gonna talk about SpaceX, but like, you know, you you know that they are built in biases to certain models. You know, you just like, well, you know, this this model won't let me say the word Tiana.
Right. You know, this model won't let me acknowledge that the twenty twenty election wasn't rigged. You know, the this election thinks, you know, people did land on the moon in nineteen sixty nine. And so it's like all of this stuff is seeping into the tools that we are using and and and steering things subtly, but you know
Nevertheless, they are getting steered by human biases on side of inside of LLMs. And then you layer on the top of that, like, well
you can't have access to some because some people don't like them for whatever reasons. Then you have others that are like, Well, this one's been thrown into this, you know, this marketplace because you know, some b biz dev deal needed to happen and we needed you know, we need to make sure that Grok was available in the AWS store despite, you know, nobody wanting access to that. So
Coté (04:39.969)
Yeah, I mean as long as there doesn't come out a model that's like, you know, freedom truth model, then then I think you just steer c steer clear of that.
Matt (04:45.793)
But you know there is. Is isn't that what what Pac Pat from from VMware is is off trying to do? Yeah. I think that was what his play was. He wanted some sort of like you
Brandon (04:53.894)
Gelsingler, he's back. Yeah, no, good question. I don't know. I think he's just back I think he's just back building chips. I think, you know, Gelsinger, the one thing he wants to do is build chips and build fast chips. And like, you know, he's gotta be just so so unbelievably upset that Intel, at least, you know, as of this recording, has had such an incredible comeback. It has to just burn in his in his soul that he didn't, you know, get the get the the credit and live through that one part. So I think he's
So what do you do? Just start another chip company. You're like, we gotta make we gotta do it, gotta fix it. He does. I think he does, right? He likes incredibly complicated strategies and chips. And it's just like, you know, it's like maybe you know, maybe that was go back in time. It's like, hey Pat, why don't you just chill out a little bit? Why don't you just like let's just keep the structure a little bit more sane? And you know what you can do? You can ride this until CPUs are back in in the good graces of the stock market, and you're gonna make tons of money.
Matt (05:25.741)
Well ho Yep. Hop hopefully they enjoyed the last week. Yeah. Yeah.
Coté (05:26.731)
Yeah. The man loves chips.
Brandon (05:52.477)
We don't need seventy-five page memos and four hundred business units. Like we didn't need to do all that. If you just waited, if you had just waited, you would have gotten it. But you just you couldn't leave it alone.
Matt (06:04.169)
Ha ha.
Coté (06:05.404)
That's that's Brandon's grudge is long strategy memos. Nev never never forget.
Brandon (06:08.754)
Well, I like yeah, it's maybe long strategy memos that really aren't much on strategy. They're just much on like a lot of like navel gazing. That would be my that would be my I guess more detailed critique.
Coté (06:24.941)
Well, so okay, let's get some predictions. now that we know, I think I think we have we've got a fuzzy idea of of what was up with the Fable stuff. I I like some people, I got to use Fable for a couple of days. I was very excited because I was gonna get free tokens until like September tenth or something like that. It's good stuff. So, you know, just fable it up. Now I've got two well, two prediction. Well let's let's make one predic let's say it's it's gonna come back at some point, you know.
Brandon (06:40.188)
Yeah, there you go. Mm-hmm.
Matt (06:41.24)
wow.
Coté (06:51.403)
It hasn't been two weeks yet. So there's we've always got the two week window, that that things will happen around here or around there, I should say. and but when it does come back, are they gonna push out the free token period more or is it still gonna be September of whatever it was? Brandon, what do you think?
Brandon (07:08.922)
Yeah, well I think it's it's like the summer. I think Fable went on summer vacation. It's at camp right now, it's having a great time, and it's gonna come back. at least in my world, summer ends like kind of and when I say the end of summer, I mean the end of August, like September. When kids go back to school and everybody in college and high school is like, Hey, we gotta get gotta get these papers going again. We need the best model. It'll come back. And yes, I definitely think I don't know exactly what they but they'll have a a nice promotion. They'll have like, you know, Fable's back, here's the
Matt (07:14.893)
Ha ha.
Coté (07:28.451)
Like that.
Brandon (07:36.134)
the promotion period, maybe that's some free credits or low usage credits for a period of time. So I I pr predict all of that will happen. So yeah, so it comes back in the September and maybe through the fall. There's like a fall promotion of credits. That's what I think is gonna happen.
Matt (07:47.596)
No.
Coté (07:48.897)
Wha what do you what do think, Matt Ray? Are they gonna still cut it off at the same time or are we gonna get it extended?
Matt (07:51.82)
I I I I think I think that like from everything we're kind of seeing out in the marketplace, they're already pressed for resources. You know, they're they're doing deals with with you know the the Memphis data center. They they've got like, you know, partnerships with AWS with you know, everybody that they can to get resources. Why do they need to give away more? You know, at this point
Coté (08:00.855)
Mm-hmm, mm-hmm.
Matt (08:16.205)
And why would they give away something that, you know, may cost f you know, thirty to fifty dollars an hour of just, you know, if toke for per GPU? I mean it's it's insane how I think there's a a lot of pent up demand for it and people would be like, Well, you know, I gotta have the best. I don't need a free ones to to, you know, nudge me that way and then you know, it immediately eats all their tokens and you know.
I I don't think they need to discount it at all. I think, you know, they've they've got more business than they can handle and they're losing money faster than they can handle. So why why accelerate that?
Coté (08:53.695)
Mm. Yeah. Yeah. Yeah. I I like I like I like Brandon's theory that the the AI token promo windows are driven by when people want to cheat on their homework. Or n not
Matt (09:07.201)
But you don't you don't need you don't need fable for that, right? I mean I I see
Coté (09:10.261)
Not may maybe l not only cheat on your homework, but also have a a affordable tutor and pairing for your educational enrichment. I d I don't know if
Matt (09:19.681)
Well that's what I'm saying. I I think I think part of freshman orientation these days is gonna need to be like token management. You know, they're gonna they're gonna need to teach them like, hey, hey, tap the brakes here. If you use, you your Opus four five, it'll it'll get you a it'll get you a solid B plus. And yeah.
Coté (09:26.283)
yeah.
Coté (09:34.774)
Now that would that that would be a good experiment is we're gonna let you use AI as much as you want, just like I think our generation might have been the the last generation that couldn't use calculators all the time. I feel like I feel like my kids are just like calculating it up. Like no no problem. And and maybe the experiment is like, all right, we're gonna let you use AI all the time, just like calculators, but you got a monthly budget. So you gotta manage your budget. And and like once once you run out of tokens
Matt (09:45.718)
yeah.
Brandon (09:49.489)
Yeah.
Matt (10:01.045)
Your calculator budget.
Coté (10:04.247)
Time to get out your pencil because you're gonna be writing and shit, right? Go find go go to the library if you wanna look something up.
Matt (10:10.069)
Yeah, yeah. F find yourself a good local L O
Coté (10:13.635)
That's that's right. And exactly. And if and if you can install your own LLM d on that Chromebook that's ten years old, sure.
Matt (10:18.958)
Yeah. Ask Siri to do your homework.
Brandon (10:25.009)
On that note, this episode is sponsored by Sentry. Applications break in all kinds of ways, crashes, slowdowns, regressions, the stuff you only see once real users hit it. Sentry catches all of it, and that and that's S-E-N-T-R-Y. You get traces, replays, errors, profiles, and the details around them, like stack traces, commits, releases, and the developers who broke it, all in one connected view. So you're not jumping between tools trying to figure out what happened. Sentry shows you how the request moved, what ran, what slowed down, and what the user saw.
Sear, Century's AI debugging agent, takes it from there. It uses all of that century context to tell you the root cost, suggest a fix, and can open the PR. It also reviews your PRs for you and flags breaking changes with a fix ready. Try it for free at century.io. That's s-n-t-ry-dio, and tell them that we sent you. They have a free dev plan and listeners of the show can use the code SDT26. Again, that's SDT26 for $100 in century credits for new users.
And of course, we thank Century for sponsoring our show.
Coté (11:27.619)
So there's a new magic quadrant out. And you know, it's been a long time, I think, since we talked about a magic quadrant. So I'm a little excited about about what's gonna happen here. I I I even had the the robot do a little bit of analysis for me. I thought it was Jermaine. I th I thought it was you know, would go along with this because the robot is of course an AI, and this magic quadrant is the AI governance platform.
Now I didn't actually govern my use of it, but I governed my understanding of what's in this magic platform. Because AR governance platform makes sense. It's a platform when you're using AI that will govern it. you know, do do all these things. So
Brandon (12:11.129)
just want to before you go on, I just want to say I think you coined the a new phrase. You said magic platform in there. And I I I think while that doesn't have any official definition, I want to trademark it immediately. And I feel like that is now a software defined talk trademark. No one else can use it. And we will later on define what exactly that means. And we even I'm not sure, but go on.
Coté (12:16.203)
Yes.
Matt (12:18.689)
Hey wait, wait, wait.
Matt (12:29.773)
Part Yeah, it's part of the the software defined consulting arm.
Brandon (12:33.839)
Yeah, that is absolutely part of the consulting group. And like just like BCG has their their matrix, we have a magic platform. Top that, Gardner, top that, McKinsey, we win.
Coté (12:44.875)
I think I think maybe the the theory of the magic platform is that despite its shortcomings, it magically is successful. And we that maybe that's the kind of we're trying to find unmagical things. Anyways, so here's here's my here in my here's here's here's Opus four point not fable. It's run down. It's it's got four jobs that an IGPD does. Doesn't inventory all the AI stuff running, maybe even the prompts, it's just inventories your stuff.
Asset managed I don't know if you call that asset management, a CMDV, but you know, the first job of any governance is knowing what you have and not only knowing, but like getting a printout of it virtually, I guess. And then you got policy and risk mapping. I'm a little unclear on this, but I guess it's just sort of like you got all your your acronyms and there's probably like some script that runs and it's like yay or nay, up or down.
Like matches that, you're compliant. Matt Ray, you probably know about this stuff. People coming at you with, I bet you got some like wacky Australian shit they're asking about and you know, all sorts of things. Like that's that's right. You mean the snakes I have to clean out of the house? anyways. And then and then there's the runtime enforcement. Now this is where things get a little magical, it seems like. You know, if you see something out of compliance, you would want to do something about it.
Matt (13:46.304)
Absolutely.
Matt (13:51.725)
No, it's just built into the day job.
Coté (14:11.497)
And and I think what these platforms do is I don't know if any of them actually do anything. They call another tool that does something, which is fine, right? So they might trigger some some action happening. And then you've got audit trail and evidence. Everyone's favorite, right? the the the the blame storming. Who did the what? Who authorized it? What was going on there? There's always a missing part of that, which I think should be in enterprise software, is like, and now what? Right. But I get
Matt (14:39.599)
nobody wants that.
Coté (14:41.591)
So now I think that's the basics of what happens. So I'm imagining what these products do. I haven't had a chance to you know kind of top gear all of them, so I'm not quite sure. But I imagine what they do is, you know, one, they are agnostic. Maybe I don't know what the kids say, multi-AI, hybrid AI, AI bursting. We gotta have some kind of situation. Not not just one one type of of of AI. It'll handle multiple of them.
Brandon (15:02.297)
AI curious.
Matt (15:04.107)
Yeah.
Coté (15:08.873)
And I imagine, you know, starting with the first one, it like knows the models you're using. Probably one would hope, like the different MCP servers. I'm always curious how people track skills if they do something. You'd have to go scan disk and and kind of might be kind of weird how you would do that, but whatever. and maybe even down to the person doing something, I'm not sure. and then it just tells you like, are you in either your own rule set, your own compliance that you want, or are you matching other compliance?
And if something's wrong, maybe it's running on, I don't know, a real time loop, a five minute loop, a nightly loop. At some point it needs to what do they call it in the Kubernetes world? Reconcile against ideal state or something. And so you've got to do that at some point and then just like turn things off who are who are the the bad boys and girls doing things. And then of course you send the email of doom to like PWC or someone and and it's just like these guys are fucked.
because of this policy, the zip file that's attached here. I I don't know, did did did y'all g go over what this is? Does that seem to match expectations?
Brandon (16:18.298)
Think it does, but there's actually a great paragraph in your write up here. And I think I I this sort of the meta question is like in and I want you to answer this in just a second. It's like, you know, do we even need a magic a quadrant for this? And and you as your former analysts, why what were the motivations that created that? And to give you some context, so this is what the the AI said is an an AI governance platform is for the person whose job is to be able to answer the questions how many AI systems do we have in production?
What regulations apply to each? Who approved each one? And can you prove to an auditor that none of them are violating policy? Now here's the key statement. That person did not have a budget line item two years ago. Gardner is betting they will have one by 2027. The strategic planning assumption that AI governance becomes a sovereign AI legal requirement worldwide by 2027 is what justifies this bet. And so it goes on to basically say that, you know, if this does exist, this is, you know.
why this magic quadrant exists. And if it goes away, if in it it says if A guy AI governance just gets absorbed into the data and an analytics platform or the some cybersecurity stack or the model hosting clouds, then the whole MQ is a snapshot of a category that briefly existed and then dissolved to adjacent ones. So I just want to jump to the end and be like, yeah, that's what's gonna happen. Like none of this is going to exist. None of this is going to exist. And it I it
Matt (17:35.681)
Yeah.
Matt (17:39.725)
Ha ha ha.
Brandon (17:45.977)
almost angers me like when I see this stuff be put out because and I guess the reason I was thinking about it is that Gardner in theory, right, I think they would say is charter with educating end users, technical buyers, CIOs on what they should be considering in the future. Right. That was will be what they would say. It's like we're out there to make sure. And they would say that, you know, we we get briefings for the vendors and we consume the information and then we package it together.
Matt (18:06.411)
Yeah.
Brandon (18:15.842)
In a way so that that we, if you will, can be the guardians of good information to the IT people out there making purchasing decisions. And I look at this and I'm like, this is so unnecessary to have a magic quadrant for this category that I firmly believe is not going to exist at all in a few years, is sort of like going against it. It's like confusing everyone. And the reason I feel that way is that, you know, as everyone has heard, like I watched all the keynotes, you know, and the one thing that
Coté (18:24.724)
Mm.
Brandon (18:45.648)
All the keynotes talked about for I mean at length, right? And go back to listen to all episodes, AWS, Azure, and Google was all the, if you will, all the governance stuff they had around all of their solutions, right? So I feel like it's clear that's where this is going to live. And you would be crazy, absolutely nuts to go out and be like, you what I'm gonna do? I'm gonna get third party governance system, and I'm gonna try to put that on top of Azure, Google.
And AWS because this stuff's so new, they're trying, they're still trying to figure it out. So so I feel like there's like a real I don't I I guess let me go back to my question for you, Kote, it's like, what is the incentive here by the analysts in creating this magic quadrant? And I'm just am I being too harsh? Like, are they trying I feel like they're doing it just to like almost create more work and make create more confusion, but tell me I'm wrong. What what what's going on here?
Matt (19:19.593)
Yeah.
Coté (19:40.152)
Well, it's important for analysts to create new things. I we would call it a meme nowadays. I don't know if we'd call it a meme. What would you a a category, a meme, an idea, a saying? Yeah, yeah. So so that that's always important. So yes. I mean that's that's that's the first thing that you're going for. And I think I think in this instance, well there's a few things. One, speaking of like you know, memories of elephants, as as
Matt (19:48.439)
You you gotta you gotta thought lead this thing.
Brandon (19:49.498)
Right.
Coté (20:05.611)
You dear listeners may recall the first Magic Quadrant for Infrastructure as a Service, I don't think had Amazon on it. Or like it was it was way down in in somewhere else. I think maybe Verizon ranked higher than Amazon. I I'd have to verify that, but like whatever. So, you know, the first ones are are like, you know, the first magic quadrant is like the first pancake you cook, right? Like you just just throw it out, like and make the second one right away. Unless you really like I mean, I kinda like the first pancake, but whatever.
Matt (20:14.637)
Yeah.
Matt (20:20.535)
Checks out.
Coté (20:34.581)
Gotta you gotta Yeah, and and kinda wiggly and wobbly. Like it's it's it's sort of like a a child's book like problem.
Matt (20:35.085)
Kind of burnt and like no.
Yeah.
Matt (20:44.026)
So so so there we go. We've got our our Burton Wigley magic platform.
Coté (20:47.125)
Now now the other thing is I think this is this is you know, whatever. Disclaimers insert here, blah blah blah. But like this seems like a a category that should really be a set of capabilities that you have. Because as as as you're alluding to, Brandon, like one, you want to define what the capabilities are. Now I I read that list in kind of a snarky way, but it's like, yeah, yes, that's what we need. Enterprise, right? Like
Matt (21:02.743)
Uh-huh.
Coté (21:15.959)
You gotta have that stuff. And I think I think what would probably be valuable for the the the buyer is to know here's these capabilities that we all have. And if we look at the options that you have to gain those capabilities, whether it's the platform that you're running on, it's a multi-AI or whatever one, here's how you can evaluate what's available there, right? And each of the the the major providers, of course, they would have one drawback, and it would be it it.
lock in, it only works with this one thing best, right? Which is exactly like I'm sure, I mean I'm not sure, but I w I I would assume I would be sure, if that's a sentence, that there's probably over the years been some sort of observability magic quadrant. And they probably have like AWS N Azure things and they're like, it works best with AWS Nazure. but you still put it in there as like here are tools to do to do that thing. And that's that feels like
Matt (21:46.75)
It's not cross platform. Yeah.
Matt (22:01.825)
What w what
Matt (22:08.588)
Yeah.
Coté (22:15.073)
what you would want going on with with this. and then also, you know, the other the other thing that's weird is the the the the market size that they give it, which is this very small market size. but but I think I think that also gets to the point of like, well it's probably actually if you if you bring in all of the people who have these capabilities, it probably is a larger market. It's always hard to put like a value on a market of embedded things.
Matt (22:26.495)
Absolutely.
Coté (22:41.879)
Although actually Matt Ray, maybe it's not. If if you're paying public cloud stuff, like people are paying per every unit of use, right?
Matt (22:52.297)
Right, but but when I look at this I'm like, where's the magic quadrant for I am management? Right? You don't have one, right? Every cloud provider is like, here's how you h manage, you know, permissions. You you don't have like some
Coté (22:58.603)
Mm.
Matt (23:05.567)
you know, service now tool that's like, we'll manage your IAM permissions from Azure to AWS to GCP and we'll you know, we'll flatten them and treat them all the same. It's like, no, you go into each tool, you set it up, and you manage them independently. There you don't need a magic quadrant for that. And so like the things that are missing from this magic quadrant are all the people who provide the cloud or who provide the AWI the AI services who already have their own management consoles for permissions.
Coté (23:30.69)
Mm-hmm.
Matt (23:30.763)
You know, like there is an anthropic console. You know, I'm I I've never seen it, but I know it exists, right? There is a bedrock console that manages like probably all the permissions listed here, but you don't list your, you know, your AI providers on this quadrant because none of them are cross-ai provider, but do they need to be? Right?
Coté (23:51.266)
Mm.
Brandon (23:51.952)
Yeah, and I think even going like one step further, right? I think kind of the things you outlined, Kote, make total sense. Like, here are the things you need to be looking for. And now I an article or a report that would be helpful would be like, here's what governance requires, right? Write all that out. Like kind of as you know, you've kind of taken a stab at and how probably if we could read the report, of course we can't read it because it's behind the paywall, maybe they do that some of that there. But then like this is where you want to get to the point of like, hey, when they when they're looking at
Matt (24:15.575)
Somebody look at it.
Brandon (24:21.443)
The CSP provider, magic quadrant, right? These are gonna be features and functions they expect, you know, all of the major clouds to have built in, right? And so then within that, they could say, like, some type of like, you know, report that says, like, hey, here's all this is what it means to do governance, right? And then it says, like, here how all the providers are gonna provide governance. And of course, to Matt's point, of course, the drawback is they only provide it for those things. And then at the end, you could say,
Matt (24:42.497)
Yeah.
Brandon (24:50.649)
Potentially down the road, if it is required, here's possibly a few vendors to kind of watch or about how they would do it across, let's call it multi-AI, right? Okay, that would be like a useful piece of research, right? But like this idea that you're gonna put a magic quadrant out, it's like you know, it's like having a party with no guests. We're not gonna invite the actual platforms into the party that are the major part. And then to list like
I mean, this was so much fun in the software defined talk Slack, so everyone should join. Like, I would say I would challenge anybody, if anyone can name all of these vendors w on this, it would I mean you you you're you've cheated, right? It's impossible. Like most of these vendors you've never heard of. No one's ever heard of them, right? And it's like, what like what is the point of that? Like who's gonna like buy and this is the idea of a magic quadrant is sort of to give you some guidance, like
Matt (25:41.174)
Yeah.
Brandon (25:44.847)
If I buy this, this is what's going on. It's like you should, I mean, you should be wary of every single company here, even to the point they even kind of talk about like it's likely if any of this gets traction, they're all going to be bought up, right? So don't, you know, so don't even if you think you're doing multi-ai stuff, you'd have to almost for sure account for like, well, what happens when of course everyone's favorite what is this, credo.ai, which we're I'm sure we've all we're all using, you know, they're in the what are they in the visionary quadrant. Like, well, what are we gonna do when they get bought up?
Matt (25:56.257)
Well
Matt (26:08.141)
Yeah.
Brandon (26:14.615)
Right, you know, that we're gonna move over to to Montor and all some other obscure name here.
Matt (26:15.703)
Well and and and
And probably the reason SAP ServiceNow and IBM are on here is they already acquired one of these. Like there at no point did did somebody at SAP say, like, we're gonna get ahead of this curve and we're gonna you know, we're gonna be the the thought leaders for this. They're like, Hey, we need to be in that quadrant. You know, we need to be on the map for this, we see it coming, go acquire something, put it you know, and and like I don't know that, but I know that
Brandon (26:24.632)
Right.
Coté (26:45.517)
Well so so there this is something related that that was going on in not only in the AI field, but this idea let me see. I was listening I was listening to kind of like an o overview of of of AI news this week and apparently
I didn't follow this at the time, but there was a lot of fancy AI, high-level people moving around, going to anthrop on through I'm gonna start saying anthropic, going to anthropic from Google, and then Google lost some untold billions of dollars of of share price. Who knows why? But you know, the the these two things happen on the same day, so that's that's how smart financial analysts are. just do doing that. And it got me thinking like like
Especially there was this one person I remember profiled who had like gone back and forth between open AI and some other place. And and it reminded me of three things. One, as Brandon said, very few people are actually there for the mission. They're they're there for the money. Right. And and you can test that out easily by just seeing w how how they move around. And then contrary to that, you know, as someone who's been at the same company for eleven years, it got me thinking like, man, you gotta stay to vest.
Like if you're if you're shipping like this is no good, right? Like like you you're gonna lose all the stuff and then you gotta start over again. And so that's you know, that's not cool. There's there's some advice for young technologists out there is look at your vesting window and just like s stick to that. at least for the three years or however long. But then three, I was also thinking like, I mean
Matt (28:10.829)
Yeah, yeah.
Coté (28:20.469)
Like are things that desperate that you can't just like wait six months for stuff to change? Like why like what is the dynamic of wanting to move? N and then the fourth thing was like, well, if it is true, then I guess the big companies just can't like do stuff fast enough. I I don't know. It's it was a good I I think I think seeing usually I see like famous executive e people move around and just I I don't care.
but in this instance I think it's a good kind of experiment of a lot of our our our kind of theories here of like, can large tech companies do things or do they need to acquire? Like what actually motivates people to to to do their work and why wouldn't you just stay there to to vest? I don't know. So
Brandon (29:04.014)
I think a perfect question to answer that would be another question for you, Kate. Would be something like, now let's go back in time, right? The iPhone has just been launched. Like there's rumors that Google's building something similar, Android. And you're working at either pick tick pick pick your place. You're either at Nokia or you're at BlackBerry. And you're like, you know, it's good. Those are phones are good. We could, you know, I I I I'm on a good vesting schedule here at Blackberry. One point Blackberry like one of the
Coté (29:24.035)
Mm.
Brandon (29:33.625)
you know, high flying stocks. Like I should probably work out my you know, wait the next three years, right? And then get my stock and you know, I'll do some good work. I'll I'll build a good mobile platform. It's not going anywhere, right? Blackberry's been around forever. There's no there's no way it's going out of business. So I mean, because I think that's sort of what I feel like it's worth discussing kind of what we're seeing here, right? Is that it's kind it's really becoming pretty clear, right? That there are really only two places, maybe three.
But really, it seems like anthropic and open AI, they seem to be getting this escape velocity, right? The frontier models they are building are just flat out better. Right now, Google is the obvious one that's sort of in this weird limbo, right? Because they started all this, they have the team of researchers, they seem to have an endless bench, but they just they don't they can't quite like they get close, but they can't quite, you know, get into the the realm. Whereas Microsoft
Coté (30:09.185)
Mm-hmm, mm-hmm.
Brandon (30:31.202)
Seems like they've sort of given up on their own frontier model. I don't know what's going on there. AWS, same thing. Given up on the frontier model, right? They're just they're happy to just, if you will, rent the infrastructure, which is extremely profitable business, right? You've got Facebook, very confusing situation there. I don't I mean, I put some of the articles in there. Even I I don't even know how to discuss it. It's just like everyone's mad. They're firing people. They've spent billions of dollars. So it it's so you just kind of look around. You're like, so it's kind of back to like, well, if I'm a great
mobile phone developer or I want to build, you know, great mobile platforms, there is kind of this point where you're like, I need to go where the action is. And I need to go where the places that are that are actually going to succeed. And it does kind of look like right now that the two places are. And this is why it seems like it's just the pick is either anthropic or open AI. And then on the money side, I think two things. One, I think they always will probably match what you have, right? So they'll make you whole if you come if if you are the
Coté (31:12.589)
Mm-hmm, mm-hmm.
Brandon (31:29.614)
quote unquote research engineer that has these skills. And then two, I think you got to be looking at yourself as like, yeah, what I have here, if I let this vest out, it may not be worth nearly as much as the other thing. Now that's, you know, that's obviously I'm kind of painting a picture that the frontier models is becoming differentiated, right? That these models really are staying ahead. Whereas, you know, people have r consistently said, you know, you know, the open source models aren't gonna catch up. But somebody joked somewhere online was this like, well
the foundation the the open source models, they're always just six months behind. But they're always six months behind. They're never as good as like the next good thing. So I don't know, Matt, w what is your take? Is it escape velocity happening here? Or is it just people you know rushing out too soon?
Matt (32:14.549)
well I mean I can't speak for the the individuals. I I think if if I were one of those folks, I'd be like, get yours. Yeah. Get get paid in cash if if possible. upfront if possible. But I mean I I I think I think it's kind of underselling the fact that, you know, the the you know open models and and you know the rest of the market is six months behind. It's like
Brandon (32:26.605)
Right.
Matt (32:44.257)
You know, just because you know, you can get the latest and greatest and most expensive, most people don't know what to do with that. And and so, you know, you see I I mean, I've seen firsthand, you know, enterprises that are like, you know, we have to have the latest and greatest, and then you you know, you look at what they're doing, you're like, Jesus, you could have done this three years ago with, you know, Chat GPT one. and and so I think that there's like this
There's there's definitely a a tension of like, do you need to pay a premium for you know some model that costs you know sixty billion dollars to train? when in six months you can have the same thing from the Chinese or or from Canada or from you know other countries that are providing open weight LLMs that you know don't have the same capabilities, but you they're literally where we were in January. And like
on on pennies on the dollar in comparison.
Brandon (33:43.151)
So but again, kind of play that. I look at the fable discussion we've been having for the past few weeks, Kote, you know, what's your take? It's like, yeah, I mean, that's true, but like then people then a new model comes out and people are upset they can't use it. And of course there's tremendous demand, you know, that the fact that Claude gets used, you know, has high usage, can't, you know, like you said earlier in the show, right? Can't keep the capacity going, right? So it's like so clearly there are and may maybe it's just a psychological thing. People want to use the newest stuff, right? I don't know. What what do you think?
Coté (34:13.217)
Yeah, we y and I I like I like your your escape velocity thing, Brandon, because that that that that that gets at it is I guess I gu I guess I guess what what kicked off this thinking in my head when I was reading or you know, looking over the stuff is if you have people going back and forth between all these companies, then they also haven't decided which one is gonna be the winner take most all, right? Like because the whole the the whole phone analogy is like, well, there's there's only gonna be one.
pretty much that that you're gonna want to work at. But I so if there if there's two or three, well eventually you've got to decide like these Google people, it's not gonna happen. And maybe, maybe that's maybe that's the analysis that the
Brandon (34:53.58)
And to be fair, the one, I guess the Nobel what is it, Glaureate or Nobel winner? I can't remember which exactly his title was. The did the what the protein folding. I'm getting that wrong. But you know, he he'd been at Google a long time. So it wasn't like he he's not that wasn't a rash decision, right? He so it's clearly like he got to the point where like, I've been here a long time, I won a Nobel prize, I figured out this protein thing, and it's time for me to go to anthropic, right? So and I think that's why to back to your when you said like the stock market went down or why Google's it's like
Coté (35:18.198)
Yeah, yeah.
Brandon (35:23.874)
That is a pretty telling thing, right? It's like 'cause you gotta think Google was like, What do you like I mean, they have infinite money, right? And and you know, it's like they probably were just like, What do you want? And he was like, Well, you just don't have it. But what I want is I wanna be on the latest stuff and I wanna be doing the pre training with like these this group of people that are anthropic that I think are the best. And they just, you know, they can't provide that.
Coté (35:43.873)
I get yeah. So I maybe maybe this is the the model I'm trying to figure out. There's one, you got you got it could be any any sort of like thing like this. You got people moving around all the time. They're just like flighty and they wanna chase fun stuff and then they get compensated and doesn't that's just it. Like they just shift around, right? Or and or there's a model that says
Brandon (35:59.103)
Mm-hmm.
Coté (36:06.665)
These this is how like, you know, nerdy economists would probably think is like we can use the movement of of these high level people between companies to indicate which one is the best company. 'Cause we trust their judgment and they're making some sort of like Nash equilibrium prisoners dilemma marshmallow thing in their head where they're they've they've been like, I've been here ten years. Right, right. And s and s
Brandon (36:30.635)
Right. They're the insiders. The insiders are moving, right? They know in theory.
Coté (36:33.837)
They've made a judgment that this place can't cut it and so I gotta go over here. It's it's the Apple that I'm gonna go to. And and I it just seems like with with as as as fast as as fast, as frequently as things changed over the past two years, like Brandon, you remember Cursor? That that was a big deal. Like and now not.
I mean, I g I guess it's a sixty billion dollar big deal, but like it's not like people talk about cursor all the time, except in the way that I am. and so like it seems like that's what I was thinking of is just like, well, if you wait six months, things will be different. Now it and and so it's difficult to like figure out how to react to to people moving around, but it does seem important or not, or or are they just gonna move back in six months?
Matt (37:10.509)
Exactly.
Brandon (37:10.551)
Yeah, definitely could be. Mm-hmm.
Brandon (37:22.519)
Well, but I to your six months thing, yeah, but to your six months thing, I mean, anything is obviously anything can change, but it does feel like the talent has gravitated to basically two people. 'Cause it it's such a small group of like people that are, if you will, are at the top of this game. And it seems like they're mostly at anthropic and open AI at this point. So it's like
Matt (37:22.765)
I mean
Coté (37:42.253)
So that that that would be that would be the trigger you would be looking for is once once both of these companies have let's say sixty or seventy percent of whoever we agree is the top talent, then that's the BlackBerry to Apple switch. And then once one of them has like forty to fifty percent of the talent, then they are like, you know, that's that's the definite like Apple switch that they're they're the the big ones.
Brandon (37:59.425)
Right.
Brandon (38:08.057)
Or I I think you can couple that with, you know, just the quote unquote products or the models themselves. It's like, okay, take I don't know, just pick any of them, Azure, AWS, put whichever whi ever one you want. It's like, okay, if they have a model that's just as good, it will it should eventually show up in the usage, right? And you know, other people have made this point was talking about, you know, Microsoft Azure had access or has access to the open AI stuff for like, I don't know, almost since the beginning.
at least of of of this like you know if you will since the LLMs have gotten so popular and they've not been able to take that advantage and do anything with it, right? Like nobody thinks Microsoft Copilot is the best, right? In fact, might the people the announcements are like, copilot lets you change the model. You know, they say it's like, you can pick Claude or you can pick you know, OpenAI or whatever you want or Codex or whatever you want. Right. And it's like, so there's some real evidence that like
It's not happening. Same thing with AWS. Like they've built their own models. No one talks about them. You know, they're just they mostly they're on the in the keynotes, they're talked about as like, these are models that we built that you can then pre-train with your own data, but no one talks about them like this is the one I want to use. So so there's clearly something like happening that it isn't just about even if the quote right person is at this company, like there's something else in this, you know, if you will, in the secret sauce. Maybe it's more people or maybe it's the the way they're dedicated to doing it or
You know, and this is where and then this is back to the consolidation. And I I do kind of go back to you know, the Blackberry analogy was I remember I had I think an iPhone three, like the iPhone 3G. I just gotten it and of course I liked it like everyone else. And I was working with someone that was a huge fan of the Blackberry, right? And he handed me like I can't remember what it's called, Blackberry Storm or something. Anyway, he handed me to I was like, let me check it out. And you know, I remember I just swiped up on it and like it had this jittery like the physics.
Matt (39:48.333)
Yeah.
Coté (39:52.215)
Mm.
Brandon (40:00.855)
You know, like in the you know how the bouncy feel of an iPhone, it just like kind of works, and you're like, just and it feels so natural. And it was like, and I remember scrolling up on it and it was like this weird, like the physics weren't right. And I was like, This is never gonna happen. You know, like you could just look at you could hold the phone and be like, It's never good. Like they had the keyboard and stuff, and he was telling me about the keyboard, and I was like, No, it's over. Like it's and you know and and again, it's like you say to myself, Well, why can't they do the physics? Why isn't it like
Matt (40:02.892)
Yeah, yeah.
Brandon (40:27.657)
I don't know. They couldn't do it. They couldn't like anyone at that time that knows what I'm talking about. Like they could not get it right. And it was just the end. And that's kind of what, you know, kind of back to like what people want to use. It's like, I don't know. You know, I d I can never tell you like what's the best agent, but I just know mostly when I ask Claude, like it figures it out and gets it all right and gets it done. Like I'm not looking to move, you know, unless someone's like it needs to be a lot, a lot. Not doesn't need to just be slightly better. It needs to either be like free.
Matt (40:48.598)
Yeah.
Brandon (40:55.105)
And I can use it forever and or it needs to be like somehow ten times better and I still wanna pay. And I don't see any of that happening right now.
Coté (41:02.827)
Now now since we know that you watch all the keynotes, Brandon, the the other the other the other like the other thing to isolate here is that it's not just a let's say you know product management diarrhea in a po positive diarrhea situation where like it it feels like the anthropic in particular just like releasing stuff very frequently. Like all s you know, some of it is like
Brandon (41:06.312)
That's what I'm here for.
Coté (41:32.631)
Goofy and trivial, like like, you know, I I th I think the the the the one that I can remember the most that was funny, and maybe OpenAI did at first, but it's just like we'll let you export memories. And really it's just a prompt and it downloads a markdown file that you upload somewhere, which is like great. So there's a lot of innovations like that and products that are just sort of like markdown files, which is fine, but they're always doing something. And they have a they have I
They don't have a lot of SKUs, so to speak, but they have a lot of functionality, a lot of different things. And OpenAI, you know, they they closed down their side quest or whatever. So they had a lot of stuff going on. But now if we look at the other event, your Microsoft's, your Amazons, other people, do they have an equal number of products and velocity coming out?
And and and I asked this because there there's another re like the model is important, but also like the product around it could be important. And one thing you could say about the Googles and the I don't know if this is true, about the Googles, the Amazons and the Microsoft's and the other one is just like they just can't ship the products. Like maybe their models are great, but the products, they're just not they're not shipping the export your memory thing or something like that. But I don't know. I don't watch all the keynotes. Do they are are they crushing it with velocity of new products coming out?
Brandon (42:51.36)
Well, I would say I think you're hitting on another part there. It's like, yeah, and I think maybe we could use this Amazon announcement. They have, you know, a new version of is it Kiro? K-I-R-O? It's like their new IDE, right? Or not they're new, but I guess it's like the updated IDE, spec driven development. And it that feels like a very traditionally developed product. It was like, there are too many people doing stuff. We need specs and we should have a spec and we should have a UI for it and then
Matt (43:02.336)
Yeah.
Brandon (43:20.554)
It should go through this release process, blah, blah, blah. Right. And it's like the part that I really get a glimpse into a little bit is that the people at Anthropic are using Claude to build everything, right? And they are so close. Like the people conceptualizing the products, building the products and releasing the products are all the same team. There's no six page memo like going up three layers and like getting a review and having like a UI designer. It's like, no, I think that and then, you know, just to like kind of
You know, I kind of read through this stuff for AWS and I'm just like, I don't think anyone wants this. You what this is? This is the plan command in Claude. You just like write you just type slash plan and you just start telling it what it what you want and you just tell it what you can tell it any type of functional specification you want it to create. It'll do it forever. And that's the kind of how I get the sense of like how things are happening in anthropic. They're like, you know what? We need to build something like this. Let's just build it and we'll build it with Claude and we'll learn about how to do it, and then we'll get things that we think are really good and we'll release them really fast. And some of them
Coté (44:14.828)
Right, right, right.
Brandon (44:20.17)
like the plan feature and other things take off very, very fast. And the stuff that doesn't, I think they're just like, yeah, it's to your point. It's like, it's just a markdown file. It doesn't matter. We'll just, you know, we we can just move on. Whereas, you know, Amazon, right, they're gonna have like they're gonna they're gonna have AWS Day and they're gonna have presentations and they're gonna have executives and they're gonna have talking about you know, and it's just a totally different way of doing it. And I think this is back to like why this talent seems to like kind of congeal there.
And and that's everything we know from the outside is that Claude code, right, came out of just one or two people building it for themselves at Anthropic. And that's where it's like when the user can be the builder and can also release the product and it's you know, and it's hitting on a real demand, that's when you're getting that super product market fit. That's what everybody wants, right? But it's like, you know, it's it's it's like a a meteor. It it happens so rarely. And I think, you know, it's and it's not to begrudge Amazon.
Or Azure, because Amazon was this at one point. Go back in time. This is what they were doing, right? They're like, we're building all this stuff for the retail site. What if we just built a version of it that we could, you know, let other people use? And that's why it was so good, right? That's why it was like, S3, like, that's what we need. And Nitro. And it's like they had all those advantages. But now they're they're not that anymore, right? They're completely different. They're like a completely different company. So so these are the things I think that just very maybe almost impossible to replicate. I don't know, Matt, what do you think?
Matt (45:44.27)
this is where I I I become you know, I I I'm I'm skeptical because what I see is like there's definitely excitement about like, this you know, cutting edge stuff. But in that six month went lag of you know, you've got your frontier models, you've got your you know, crazy, you know, gas town clawed, you know, bleeding edge stuff.
Enterprises aren't keeping up. What they do is is six months later, AWS shows up with like, hey, we've got some, you know, we've got a a palatable version of that for you. It's called Kiro. I've seen Kiro in the in the market. Yeah, I've seen it in in customer sites. I've I've seen people using it. And I I'm seeing Cloud too. But this is this is how
This is why so many people are on Azure, right? you have a vendor that you've trusted for years and years, and you know, who knows what's happening with those wacky startups. But if if I go with somebody who I trust I already have an existing enterprise relationship going back years and years, and they can keep up enough, or you know, or they can hit all those check boxes. You know, they they don't you know, the the magic quadrant is lagging six months too, right? So
Brandon (47:00.3)
But what do you think about like, okay, but the one data point that keeps coming back to me, and I there's at least two reports of this, is that there are quote unquote, you know, approval cycles that people can make at both AWS and Google and other places where it's like where developers have gone is like, I want Claude. I do not want to use the tool that you're giving me. The tool that you're using me is not sufficient for my job. Therefore, I'm escalating to my management. I want to get a Claude subscription. And some people like these are thousands of people.
developers, I guess, mostly of either company have done it. And I'm always like, wow, what what a signal. Like what a signal of product market fit where like the people that are working on building something competitive are like, I can't use, I can't use our stuff. I gotta use this other thing. It's like that's pretty extraordinary. Go ahead.
Matt (47:47.896)
But I've also seen customers who are like, yeah, we we had that and we ran we ran out of money. That and and and you know, you're like, you know, you've got this hot shit. Well, you know, your budget did not account for doubling, you know, doubling the the CTO's budget just to handle all these tokens. But you know what? AW AWS showed up with like, you know, hey, for you know, just ten per ten percent more or whatever, we can give you something that was good enough six months ago.
Brandon (47:54.764)
Mm-hmm.
Coté (48:16.746)
Mm.
Matt (48:16.917)
And by the time
Brandon (48:16.981)
This totally reminds me, though. This reminds me of the the IT person that's like, I just got a great deal on all these blackberries. And he's handing them out to everybody. And it's like, and then Kote's walking in with like an iPhone. And then I'm walking you're walking in that with an Android. And then we all have these like, and then people are like, What's that on your desk? Like, I don't know. It's a Blackberry. I've never used it. I just my work gives it to me. And it's just like the world is just moving by and everyone's like, no, the Blackberry's great. In fact, we got a great deal on it. In fact,
Matt (48:23.565)
I I I I I hear you. yeah.
Brandon (48:44.147)
And it's we had great plans. No one's really using the data because no one wants these things. No one's using these things.
Matt (48:48.321)
But but the but the problem is you're paying you're paying the the the six month premium to give it to everybody. And and so you have you have like you know, junior developers who are like, Hey Fable, you know, can can can you start a spreadsheet for me? Right? And it's like, yeah, I can do that. You know, what what do you need? And like, I don't know, some numbers, you know, and and you're like you could just hear the money like being lit on fire in the background and you have you have people
burning money. I mean it it's it's just
Brandon (49:20.395)
But isn't that just a pro I kind of go back to like this just a a slight product tweak? Cause I even when I use a claw, it's just like, what's the other model? What's the cheaper model? I'm trying to look at what I'm using now where it's like it'll just figure it out for you. Like it'll look at your thing and be like, you don't need the it's like, yeah, Sonnet's like, look the it'll look at the request and help you a little bit and be you don't okay. You did you say hello to me? I'm gonna put you on Sonnet. Like, you don't need the fable model. So so I feel like a lot of these things, this is more
Coté (49:28.578)
You got you know you got opus and sonnet.
Matt (49:40.845)
But but but the problem is the problem is at the end of the month they've burned through their year.
Brandon (49:47.872)
Right, but I think I mean to me that's just this is just back to like more evidence that they're doing that anthropic and and open AI, I think, are in the right place. It's like, okay, well, people want it so much that they can't quite afford it. So we gotta help them basically use their their tokens better. It's not that like I want to go use the inferior thing. I mean, again, like we can check back and sit because at some point, like I I guess it's so it's like the Bitcoin argument. Okay, like a year from now, then if this is all true and the quote unquote model, the frontier model, then everyone should be using an open source.
model in a year from now. And it's like, okay, so if you had to bet your net worth a year from now, what's gonna happen? Everyone's using open source models or Claud and it and and or OpenAI are still leading the way. Like what are you gonna bet on?
Matt (50:27.295)
I I I I I will bet you that that OpenAI and Anthropic are sixty billion dollars in the whole farther.
Brandon (50:35.307)
Mm-hmm. But they'll be worth trillions of dollars and they'll have the most users. Mm-hmm.
Matt (50:37.995)
Sure, sure. Or there's going to be, you know, as we're seeing, you know, this week, people are like, wait a second. You know, I'm you know, though these companies spend ten dollars to make one. That math will never add up. And and meanwhile, the tide just keeps rising on these things that, you know, cost you know, cost a dollar to make a dollar. No, no, no, no, no. I'm just saying like
Brandon (50:58.739)
All so what do you say? We put it out there. So you're saying a year from now they don't exist? Like what what what are what what's gonna happen? They're not gonna be public, they're not gonna be super wealthy. What what's not gonna happen? What's the bad thing that's gonna happen?
Matt (51:08.993)
Okay, so so watching SpaceX, we we they're now below water on their IPO price. if you're anthropic and you or you open AI, what do you think happens when you do the same thing? I think there's a a spike and a drop. They're public and they're below IPO price.
Brandon (51:10.827)
Mm-hmm.
Mm-hmm. Right.
Brandon (51:20.831)
Right. So what what do you think? So anthropic a year from now, what do you th like what's it worth?
Brandon (51:27.913)
Okay. Are they worth let's just a trillion dollars. Let's make a trillion. Are they worth a trillion dollars? Okay.
Matt (51:30.229)
Yes. They're that that's what's crazy, is they're worth a trillion and they're losing money, they're just hemorrhaging money, and that stock just continues to decline. And you're like, Come on, Rivian, turn it around.
Brandon (51:37.706)
Yeah.
I do. well Rivian, that's another one I'm hoping for. I don't know. I cheer them on, but who knows? That one I'm I'm worried about. Yeah, Rivian, I don't know what it's saying. But well, I don't know. I this will be a fun thing. You know, this is I almost need to tie it back. I need to add your bets to the over over on the the Enterprise AI show. Brian and I had a little draft. So need to I need to add you in there, Matt, right? And we'll have to like maybe we'll have you on there and you could recast your prediction as well. I don't know. I just you know, my feeling at least this week, and again, who knows, you know, like you said, Kote, things always change just like
Matt (51:44.715)
What are they down? Ninety six percent?
Coté (51:49.152)
Moving on.
Matt (51:58.995)
Ha ha ha.
Brandon (52:08.083)
It feels like these companies are kind of hitting it. They're they're kind of move starting to separate themselves. And I do think, you know, I feel like feel like they have the momentum to like to kind of, if you will, get to to your ultimate point. They feel like they have the momentum to get to that profitability and could be it could be huge wins, right? Just like, you know, Apple ultimately, you know, achieved the escape velocity. And, you know, at least in the mobile world, no one's ever looked back.
Matt (52:15.617)
I just
Coté (52:33.824)
You know, I I think the conclusion is it's it's it always I always shake my head at like how common sense it is. But the thing that's holding back these these these big companies that put as you were kinda going into Brandon, they they put a lot of thinking into it. They're trying to come up with the ideal solution, instead of just some slash command. Like, you know, I'm thinking, maybe that whole magic quadrant will just be replaced by like slash govern in cloud code at some point. And it
Brandon (52:58.207)
What have you had? That's a fun thing. Go look at the stock price of Gardner. Way down, right? Because people w 'cause what do people do? Right? To your point. Go ahead.
Coté (53:02.204)
Mm. Yeah, well that's a whole other story. But I and I mean I I think I think I think the issue with the big companies is when it comes to like pushing out products, they probably suffer from bureaucracy, which is definitely not something we suffer from. In fact, it's a vital part of the show that we have where we engage with the community, we tell people what's going on. Speaking of Brandon, do we have any bureaucracy this episode?
Matt (53:06.38)
Ha ha.
Matt (53:18.625)
We embrace it.
Brandon (53:26.197)
The only practice we have this week is just to remind everyone the summer season is coming, except for Matt, who reminded me before the show, he's living in winter, which I always forget. But for everyone else in the northern hemisphere, you're gonna have your water bottles, you're gonna have all your stuff going to the beach, going to the pool, and you should have put a sticker on that stuff on that water bottle. And so you probably don't have enough stickers and you probably need some stickers for conference season coming up. And what you can do is you can email your postal address to stickers at software defined talk.com. I'll be happy to send you one or more stickers.
Anywhere in the world.
Coté (53:57.389)
Well, there's a lot of conferences to go over. I'm gonna be at We Are Developers Europe July 8th and 10th in Berlin. I'm speaking there. There's also DevOps Days Graz September 4th to 5th. We got the Cloud Foundry Summit, September 21st to 22nd. I'll be speaking there too in Heidelberg. that'll be fun. DevOps Days Rocky, also September 22nd to 23rd. I'm not gonna be able to make it to all of these conferences because they're overlapping, unfortunately. we've got a discount code for that, 26 D O D S W E Def Talk.
Which you can get if you go to software defined talk.com slash five seventy-eight. Then there's also another We Are Developers, September 23rd to 25th, over in San Jose. Is that right? You know, Silicon Valley. and their discount code is Dev Pod 26. But they're also, there's a pool of tickets you can try to get if you go to software defined talk.com slash five seventy-eight. You got DevOps Dallas also in September.
DevOps Ville New also in September at the end of September. Then I'll be at DevOps Days Istanbul October twenty fourth. It's Turkey. So the conferences are on Saturdays, which will be enjoyable. and then there's a VMware user group in Orlando. And then also I'll be at the cloud native Denmark conference November nineteenth, speaking there. and you know
So check out those conferences. If you have any other conferences that that you know of or you you're interested in, you should tell us about Cause I'm always interested. I've I've been, as a side note, in the conference section here, I've been using speaking of our robot friends, I've downloaded years and years of conference abstracts from various conferences to do some analysis of like what actually gets accepted. And in I haven't I haven't done a a a redo of this, but at some point I'm gonna say about two thirds of the way through
The number one speaker across the conferences I looked at was our friend No SSH JJ. He showed up all sorts of places over the years. So I'm gonna do I need to do some leaderboard analysis, but he was quite prolific in over the years at these places. So good good for him and other people you would suspect too showed up quite a bit. well, speaking of quite a bit, what do you have to recommend this episode, Brandon?
Brandon (56:16.498)
recommendation this week is disclosure day and I went and saw it in the theater and I thought it was pretty good. but then there's been a lot of online discourse. So some people did not like it. That's all with my family. So I r really enjoyed it. My son was just like thought it was just okay. My wife was not that excited about it. So I guess I'm offering up three different perspectives there. but I I like a little Steven Spielberg sci-fi like I don't know man you know ET and like
Matt (56:41.632)
It's a lot of Steven Spielberg.
Brandon (56:45.502)
You know, and it's it sounds like some of the stuff I would do. Some of the stuff I saw online, you know, pi people like picking a plot apart different plot things. I'm like, guys, it's a movie. You know, it's a fun movie. Let's not let's not let's not overthink everything. Can't we just enjoy the movie? So I'm just throwing it out there. I really enjoyed it. I think on Rotten Tomatoes they got pretty good store scores. But if you are such a person and sometimes I'm like this, you you like to snark, you wanna like, you know, if you will,
Matt (56:52.683)
Yeah. I I saw it too.
Brandon (57:14.078)
be be more superior to other people that enjoyed the movie. You should go watch the movie and then you can you can like learn about all the plot holes in it. And then you can be at the at the cocktail party at the beach. You can be like, yes I'll disclosure day. That was stupid. I didn't get that that part, right? So that's what the rats that's where either way you can see it and enjoy it. You you can be that. I don't know, maybe maybe we'll make that the after show. Matt, don't reveal your your review. I'll I'll put that in the after show where Matt will give his review of disclosure day and see if he agrees with me or see if he he actually enjoyed it.
Matt (57:24.973)
Let's talk about disclosure day. That's the after show.
Matt (57:35.309)
Ha ha.
Brandon (57:43.316)
We'll see. So stay stay to the after show for that.
Coté (57:43.575)
I I I I like I like your hater recommendation there. It's like when you show up, it's like you should you should watch the movie so that when you can show up you can say, Not only do I not like the movie, I actually watched it and I you know, it's get
Matt (57:57.934)
Ha ha ha That
Brandon (57:59.186)
Right. Well, I mean, for those that really listen along, I mean, this is like why I read the business books, right? Like people like be did you read such and such? Yeah, I did. Let me tell you why it sucks, right? That's always a fun thing to do. So, you know, again, you know, that's why people love to have me at conferences and over for the the cocktail parties to be like, Yeah, yeah, I did you what I did read that book and then I listened to the if books could kill episode, you heard that? You want me to destroy this for you? All right, or should we just go have some beers? You know, whatever you want to do. So
Matt (58:08.789)
Well I watched the keynotes, let me tell you
Matt (58:19.851)
Yeah, so yes. Brandon, educated hater.
Coté (58:22.208)
Yeah. Yeah.
Brandon (58:24.404)
That's right. So that so I have so I I I w I wanna be fully just transparent. I have a little bit of that in me. But in this case, I enjoyed it. Anyway.
Coté (58:32.148)
Al always a catalyst at an executive dinner. That's how about yourself, Matt? What do you have to recommend?
Brandon (58:35.642)
Yeah.
Matt (58:36.023)
Yeah, let me tell you why you're wrong.
So so my pick this week is Izzard, The Tragedy of Hamlet, a a one person performance of Hamlet, Shakespeare's Hamlet, by Eddie, now Suzy Izzard. it sounds like it shouldn't work, but it works really well. like it's playing stage, one spotlight, and twenty three speaking roles. So
Yeah, I I saw it here in Sydney. looks like they're touring Australia, New Zealand right now. But if you know, if Shakespeare's your thing, you will probably enjoy the show. which was funny 'cause 'cause I saw it at the opera house, which pulls a lot of like tourists in town, you know, go catch a show at the opera house and there were a lot of Japanese who left at the intermission 'cause they had no idea what was going on.
Coté (59:34.785)
Well, my recommendation is for if you do want to know what's going on, I came across this article from twenty twenty by Molly Fisher, called called Will the Millennium Aesthetic Ever End? And it's got a it's got a couple of things going for it. One, as as a Gen X person, I love a story of like millenniums. Remember those guys? So it's it's fun to see them just becoming people.
Matt (59:57.067)
yeah.
Coté (01:00:00.736)
in in in the cultural, as Brandon says, discourse, right? Like but it it it it does it does kind of lay out a pretty nice like here's here's kind of what the aesthetic looked like for a while. Lots of posters with big words on them and caps, you know, saying all sorts of stuff. Lots of pinks
I think ochre is a color, terracotta kind of colors there, rounded edges. but it's it's a good kind of overview of like that certain kind of style that I think especially dominated like, you know, Instagram ads and things like that for a while. But what I did, of course, is there was no color palettes or or very few pictures in it. So I ask our friend Claude, can you generate me a clo color palette based out of this? And then you know, of course I wanted Chat GPT to make some example images. And then, you know, the ultimate thing I like to do is just like the
I I like reading this, but can you just give me the bullet points of what's going on? So if you wanna if you go to software defined talk dot com slash five seventy eight, you can go to the the report that it generated with a full color pal palette with colors like not quite white, terra, kata, mustard, sea foam, you know, your usual stuff. And you know, you can start start to bring your millennium aesthetic back.
I actually don't remember if she said it was ever gonna end, but I think six years later, looking at what this aesthetic is, it's still around kind of.
Coté (01:01:27.778)
So the other thing I was gonna recommend I I not recommend, but I was I was I was talking about scanning all these things. If you go to if you go to talks.cote.io, of course you can see all my talks there. But you can find a little link to my my CFP tracker.
And you can see my my nosediving success at getting talks accepted in a SVG chart that will magically resize. and it's kind of fun to see like data you know applied. Yeah, it's it's it's the power of the robot, but it's it's kinda it's
Brandon (01:01:58.152)
Wow, we got a whole odd report here. Geez. Okay.
Matt (01:02:00.245)
Yeah, it's extensive.
Brandon (01:02:04.381)
This is too much actual data. You should protect this from your employers. You don't wanna see you need lines always going up and to the right. So, you know. Okay.
Coté (01:02:08.801)
Yeah.
Well, the year's not over yet, so there there's still some some chances there. But I think I think it's it's you know, it's it's a it's fun to do like this kind of vibe coding exercise over a a corpus that you know very well and see kind of how how you how it works and how you find things and and what works and not. And it's it's it's a it's a good experience of of coding things up. yeah, well that's that's pretty much it for this episode. As always.
You've listened to another fantastic episode of Software Defined Talk. If you have found yourself at this point and you don't know how you got here, what you should do is go to software defined talk.com slash five seven eight, subscribe to the podcast, because you must have liked it.
and you know I don't I I don't remind people this so often, but speaking of checking out someone's BlackBerry with the strange scrolling, when you have that BlackBerry in your hand, find their podcast app and subscribe them to Software Defined Talk. And you know, all of our our sibling podcasts as well. They don't really have to know that you did it nor listen to it. We will appreciate it nonetheless. and with that, we'll see everyone next time. Bye-bye.
Matt (01:03:19.617)
Bye.
Brandon (01:03:24.979)
Coach, what is this in here? Anime WoW for the shorts? What what am what am I am I supposed to be using that? What is that?
Coté (01:03:27.462)
well I have I have one of the one of the one of the number one shorts that I think I must have stolen from our thing here is me talking about the anime wow. Remember that? And then in that short, in I never actually play it in that short, and several commenters are like, you never actually play it. Now my favorite one is a comment to those commenters who are like, I I know what it sounds like. And so I thought I actually have it rigged up here. So I thought I would play.
Brandon (01:03:36.561)
Mm-hmm. Yeah.
Brandon (01:03:45.561)
No.
Coté (01:03:55.021)
Here is what the anime wow sounds like. Are is everyone ready? I think it'll work. It'd be embarrassing if it doesn't.
Matt (01:03:58.645)
yeah.
Brandon (01:03:59.146)
Go ahead.
Coté (01:04:05.324)
Didn't work. Well, you'll have to put that in, Brandon. Or just chop this whole segment out. Just chop it.
Brandon (01:04:05.713)
Does not work. Did not work.
Matt (01:04:08.727)
Get to it, Braina.
Brandon (01:04:08.873)
No, this one probably you'll you know what, that's your task. You have to eventually find a way through all your your beautiful roadcaster how to play the anime wow sound, you know?
Coté (01:04:18.76)
I know why. It's 'cause it's on mute. That's well, maybe I don't know. Let's let me see if what would that hold on, let me do a test first.
Matt (01:04:21.069)
Mm.
Brandon (01:04:22.441)
All right, here we go. We'll be quiet. Give us a countdown. That way we'll be real quiet.
Matt (01:04:23.707)
here we go.
Coté (01:04:31.794)
No. Never mind. I don't I don't know if that's gonna work. right. Well, one day we'll have fun.
Brandon (01:04:38.717)
All right, well, I don't know, maybe if we get to anime maybe that's what we need anime wow in the shorts. But well but listen, I don't want to leave the audience waiting. Okay, Matt, you what was your review of Disclosure Day?
Coté (01:04:43.125)
Uh-huh.
Matt (01:04:51.413)
It was fine. I mean, it was it was a very w what's the Rotten Tomatoes open or IMDB rating? Let me guess. Seven one.
Coté (01:04:51.919)
yeah.
Brandon (01:04:53.723)
Okay, I'll take that. I'll take fine.
Coté (01:04:55.69)
It was on the fine scale.
Brandon (01:05:06.533)
Mm, I think it's I don't know, I think it's like eighty percent with the critics. I don't know.
Matt (01:05:12.479)
It was fine. I mean it was it was you could have told i i it I could i if i went into that movie not knowing who it was, I would have told you it was a Steven Spielberg film. It was just like it felt like a mishmash of of previous things where like it's like, I've seen this scene before. I I've seen that reaction on people, I've seen, you know, this type of car chase. You know, I I you know there's a certain style that he has, and then
Brandon (01:05:14.685)
Yeah, we'll pull it up.
Coté (01:05:15.25)
Yeah.
Brandon (01:05:24.167)
Okay. I think that's fair.
Matt (01:05:42.859)
I just didn't buy the end and I thought some of the like I thought some of the alien stuff looked kinda h hoky. I don't know. It just I I wanted to like it a lot more than I did.
Brandon (01:05:52.701)
Yes, I'll I'll give the official numbers. The tomato meter was eighty percent and the pop go pop commentator seventy one percent. So I don't know, I w I don't know which rating do you use? Seven point one? You mean I don't know which one you're I don't have that that scale.
Matt (01:06:03.401)
I don't know. I I I I it was it was it was enjoyable but forgettable.
Brandon (01:06:10.568)
I think that's fair. I think in I think it is. It's just like, we had nothing. It's summer movie. That's sort of like the like we had the fam we had stuff. We didn't have anything to do. So it's like let's go watch this movie. So
Coté (01:06:10.672)
Mm.
Matt (01:06:14.88)
Yeah.
Matt (01:06:18.687)
Yeah, w yeah, we we had kind of a you know, a conversation about like where it fell in the pantheon of of Spielberg and I was like, it's the bottom half, you know. which, you know, of course the Pantheon of Spielberg I exactly. But but it's like at no point did you think like, ooh, this is a strong contender to, you know, knock off Jaws or something. Like, no, you're like this is this is, you know, above that last Indiana Jones movie, but it's
Brandon (01:06:32.04)
It's a high bar though. I mean he's made some good movies. So, I mean that's yeah.
Brandon (01:06:40.794)
Agreed. I agreed.
Brandon (01:06:45.084)
Yeah. No, it's like the band playing the hits. It's like, yeah, okay. Well, you know, this is, you know, that's good though. But I'd rather this than like him go out and do some like I don't know, some like very niche art movie, you know, where he gets away from his like you know, it's like, yeah, give me a just g just give me something. Just give me what you g
Matt (01:06:49.237)
Yeah. Yeah.
Coté (01:06:57.27)
Something important.
Brandon (01:07:05.296)
well, Shannon Lewis Less was it's always one of those movies. It's a great movie you should see, but it's tough. It's a tough one. Yeah, it's I agree. It's that's the Seinfeld episode. So Well I don't know, Kote, how I has this what does this leave you? Will you see disclosure day after this this long review?
Matt (01:07:12.469)
No, I know, I'm kidding, I'm kidding. But but yeah, it's I think yeah.
Coté (01:07:25.172)
Yeah, it seems nice. Seems fine.
Brandon (01:07:28.018)
Go to watch it on he'll watch it on an iPad on the airplane someday and be Yeah, I saw that. It wasn't bad. It was good.
Matt (01:07:28.203)
It's fine.
Coté (01:07:32.567)
I mean may maybe it's I I was wondering listening to that if it's the case of of there's like s so much movie, TV, et cetera stuff that it's pretty hard for something to be something outstanding to be like outs remarkable and outstanding. Like, you know, you watch you watch all the episodes of like the TV shows that come out and you're kinda like, yeah, I remember watching that at the time. It seems great, fantastic. But it's just like there's just this endless flow of stuff.
And you know, it's
Matt (01:08:02.645)
Yeah. This is in the flow of stuff.
Coté (01:08:06.762)
Yeah, yeah. I mean a you know, in in comparison, right? Like there's there's however many Jurassic Park movies there are now and they're just kind of all fine. Like d the and but but they have the same like you know
Matt (01:08:18.093)
Yeah.
Coté (01:08:23.116)
They're pretty much the same plot as like the first one with even better like even better special effects and dinosaurs. You know, it's like, get there's gonna be an island and the dinosaurs are out of control and there's some humans who want that and also see alien. So, you know, same same sort of thing. But but once you set that bar, it's hard to like I don't know.
Matt (01:08:49.077)
I mean I've had a few things movies I've seen like this year that have, you know, gone above the bar but yeah. I mean and but I haven't seen in the theater lately. That's the problem. I've the things I've seen in the theater have been kinda like, meh, yeah.
Brandon (01:09:04.872)
Well it's interesting, and maybe there's some correlation. We can do we'll do a segment on the on the main show around like I don't know, all these people have access to like lots of money and they know how to make money. But like I think maybe the best movie I've seen this sum in the theater or this year was like Project Hail Mary. I just really liked it and really enjoyed it and it was like and it's like it's again, it's an alien movie, it's space movie and it's just like, well, like all these other people have like the same kind of money and talent and it's just like sometimes it hits. And then, you know, we we went and saw
Matt (01:09:20.29)
Yeah.
Brandon (01:09:33.916)
The start in the Star Wars movie. And that's just like total waste. You know, it's just like that was that kind of just I mean, that was really kind of disappointing. You're just like, when is this going to be over? Right? Kind of thing. So so on that note, when this is going to be over, I want to thank Brett. Haven't seen you in a long time. Brett, good to have you on the chat during the stream today. And I want to thank everyone that watched us on the stream as well as everyone that watched us on the video. YouTube channel is growing, so we really appreciate it. And with that, we will talk to you next time.
Matt (01:09:43.639)
Go see sheep detectives.