[00:00:07] Speaker A: Welcome to the cloud pod, where the forecast is always cloudy. We talk weekly about all things AWs, GCP, and Azure.
[00:00:14] Speaker B: We are your hosts, Justin, Jonathan, Ryan, and Matthew.
[00:00:18] Speaker C: Episode 266 recorded for the week of June 25, 2024. AWS billing finally comes into focus. Good evening, Matt, Jonathan and Ryan. Full house just arrived.
Well, I last week spent a fantastic few days at the Finops X conference, which if you've not been or have been thinking about going, I would definitely recommend going to the PhenopsX conference. It is fun to go to a conference where you don't have to explain to people what phenops is. They just know. Just sort of anytime you finops are like, what? What's that? And you'd explain it to them. But, uh, it was good. It was great to get to see a bunch of people from the show who listened to the show, as well as friends that I have in the Finoptex conference, including Rob Martin and Joe Daley, et cetera. But overall, it was much bigger this year than it was last year still at the beautiful San Diego Marriott Marquis, which is great because it's on the water. And so the weather was beautiful. It was 80 degrees while it was sweltering at my house in the east bay of San Francisco. So it was nice overall, a couple of things I thought were interesting. Focus 1.0 is now officially generally available, and there was a lot of talk about adoption and all the cloud vendors taking advantage of the focus standard. And another big thing we talked a lot about is SaaS companies starting to take on the focus standard as well. So Datadog, new Relic, Snowflake, these guys are all potentially going to be future adopters of focus. And so really focus could become a global standardization in how you do consumption based billing reporting, which I think could be really interesting. So even as a SaaS company where I work for my day job, we may have to eventually support focus just so customers can get unified, standardized billing for consumption services, which is sort of didn't see that happening. But it's kind of interesting.
[00:02:12] Speaker B: I mean, it's a natural progression given how much of our workloads are being know, sent to, you know, SAS providers like snowflake or confluent.
You need to visualize and rationalize those costs the exact same way. So that's pretty cool, but it is interesting.
[00:02:30] Speaker C: Yeah, yeah, I will see how it evolves. I think, you know, I think the focus 1.1 draft is already going through, so, you know, this thing's going to continue to evolve a little bit, but you know, it was fun. One of the sessions I did was a chalk talk where basically you went and they gave you a piece of paper of a printed out charts drawn on the graph paper, and then all of the schema of the focus 1.0. And basically your goal was on a whiteboard was to draw out how you would use the schema to basically recreate the charts that were on the piece of paper.
I literally had a data analytics person in my little group, and she knew what she was doing and she just busted out some beautiful queries. But it was fun to talk about it and how you might use things. And we had a whole debate about there's two fields for build date. So there's the date you consumed it, and then there's the date you were billed for it. Those can be two different dates based on wake cloud providers. So both of those are reported in the focus standard, for example, which is interesting to think about. I can see why that's important for this or that, as well as just thinking about how you start. Differences between the different cloud providers, percentage of discounts, uh, which is some of the charts that we had to do. So, uh, really interesting. Really quite good. Uh, like I said, the conference is bigger this year. It'll be there again next year. I think they will probably grow out of this facility at some point. I don't know when that'll be. They didn't. I think they officially sold out, but they were selling tickets up until last week, and I'm sure some people don't show up and they kind of oversell the conference on purpose. But, uh, yeah, I think they had to probably expand inside of the conference center they're in. Uh, just because it's a little tight. But there is one thing that happens at the Phenops X that I have seen at no other conference, which I think is pretty cool, is that onstage together, talking as one unit is Amazon Azure and Google.
[00:04:21] Speaker A: No more click.
[00:04:22] Speaker B: Are they in like individual, like glass pods or something?
[00:04:25] Speaker C: No, they just put them in normal chairs. And then Junior, who's basically the leader of the Finops foundation, did an interview session where they were talking about focus and doing all the adoption of focus and what they're doing and cloud. It's just fun to see them not be fighting with each other and just onstage all talking about how they're making funops better and adoption of focus and what they're doing. And then the second day, they actually did a whole.
Each vendor basically had a chance to talk about all the thinobs announcements and they all announced something at focus or, sorry, they all announced something around focus as well as announced new features and capabilities at the conference, which was nice to see because they didn't just save it for reinvent or next, they actually had stuff there.
[00:05:05] Speaker A: That's cool.
[00:05:05] Speaker B: So not only is cost the common metric when you're trying to basically do any comparison across anything, but it's also bringing the world together.
[00:05:14] Speaker C: Yeah, bringing the world together. Three cloud providers at a time.
[00:05:16] Speaker D: We all just want your money.
[00:05:17] Speaker C: Yeah. Now my question is, is Oracle going to get on stage with them next year?
That'd be the perfect thing.
Show floor, lots more vendors this year. Still a lot of old, old original gen one, finops tools, cloud abilities out there, cloud health, et cetera. IBM owns like seven companies on the floor at this point. I don't know how many they own. It seems like they own everybody. And then there was a bunch of new vendors as well this year that had different ideas, different takes on things.
It's interesting because there was a really clear distinction that junior made around AI for Finops and finops for Aih. Those are two different things.
And so most of the vendors are some type of, doing some type of finops for AI. But I was talking to one vendor, cloud zero, and they were talking about, you got to set up this very complicated YAML file to do your connectors and stuff like that. My first question is, you guys have AI for that, so I don't have to do as much work on that. And they're like, oh, that's a really good idea, we should do that. I'm like, okay. So I see very quickly my mentality on vendors is going to start changing where I'm thinking about how are you making AI make my life easier with your product? Which is sort of interesting because I hadn't, it was just one of these moments on the show floor. I was like, hmm, okay. And then one of the companies I talked to last year, they were stealth. Last year they were actually out and announced this year was adaptive six. I thought they were really interesting because one of the things I talked to them about last year was finups. Tooling is great, but when you need to make changes to tags, you need to do different things or you don't want them to use GP. Two instances, you want to use gp three and the Finops tool is doing it. There's no way to write that back to terraform or write that back to the enterprise code you're ideally using. So they were showing in their solution basically, it prevented someone from deploying 1000 boxes with gp two versus gp three for rds.
But then when it prompted them that they had done this thing and said, well, would you like us to write you a pull request to fix it in the code? And actually wrote a pull request in GitHub that would basically change it to gp three. Again, baby steps. Very heavily AWS focused. But the potential for gen two in ops tools, I think is really interesting and really has a lot of high potential.
We'll see what they do next year.
[00:07:36] Speaker B: What was the blocking mechanism? Is it permissions or is it like you're executing the command through the.
[00:07:42] Speaker C: They plugged into the terraform flow of the get. Basically it's part of the Jenkins or GitHub actions pipeline. And so basically detected that it violated a rule that you're using outdated infrastructure. And basically that's why I was able to prevent it. Then it gave you remediation, like ignore me, go fix your code yourself. We can write a pull request for you. We could open a Jira ticket, whatever mechanism you wanted to use to remediate.
[00:08:07] Speaker B: Got it.
[00:08:07] Speaker D: It's cool that Finops is becoming security is where it's in the pipeline now, and not just something that someone just looks at afterwards. And it's like, cool. By the way, you just spent $100,000 in three days by accident because you did the wrong instance type. So it's kind of nice you're getting these things in early and actually able to make it be what we want it to be versus then have to retrofit the world.
[00:08:31] Speaker C: Yeah, the shift left of finops was a big topic. How do we get visibility? How do we show people what things are going to cost? How do we make sure that people are aware of what they're doing?
And so I think it's a recognition of that is important and just as important as security, as your cost. And in some ways, security is part of your cost story, because if you bankrupt your company, that's a pretty bad security situation.
Any other questions about phenop sex? Again, recommend going in the future.
[00:09:01] Speaker A: Did you go to the after party?
[00:09:03] Speaker C: They have it on the midway. I did not go because I was invited to a dinner, but they did do the party on the midway again, Goose and with other character from Top Gun, Tom Cruise's character, Maverick. Maverick and Goose are on the deck, so you can take photos of them on the midway. But, yeah, I didn't make it out there this year. I did last year. It's a great. That's a great venue in general. It's a good party, good time, good food and music, et cetera. But I had other obligations.
[00:09:30] Speaker A: That's cool. I'm really happy that phenomena has become this entity that's going to have sway over lots of different things in organizations now, because somebody was always accountable for the cost of things, but not at this level. And I think having people who wear phenops hats who get consulted around the use of services is going to be awesome. Just thinking about AI and how expensive that can be, I think phenops will drive things that you wouldn't have expected. Things like data quality, because you start with better quality data, it'll lower the cost of building models, things like that. So I think we'll just see it driving improvements across the business.
[00:10:07] Speaker C: Yeah, the sustainability as well. It's another big area topic of how you know your cost of things and how it's driving your sustainability story, et cetera. That's a big deal too.
Good. All right, well, let's get into other news that happened during the conference. First up, Digitalocean has announced that they're offering you managed open search, launching a managed overseer, offering a comprehensive solution designed for in depth log analysis, simplified troubleshooting and optimizing application performance. With digitalocean, you can pinpoint, analyze log data with ease, customize log retention, enhance security and scale with your business, and receive forwarded logs from multiple sources, including digital ocean managed services like droplets and managed databases. And I really just put the story here because we don't really care about OpenSearch, but I really put it here because I wanted to comment more about the fact that Opensearch has taken over the world. No one is doing elasticsearch anymore. Everything is becoming open search very quickly. And I'm intrigued to see if that follows suit with hashicorp terraform as well with Opentofu and Open bow. As IBM starts to consume hashicorpenne, you know, probably the end of the year. Kind of interesting, but yeah, great. If you're on digital ocean, another great badge service, including kubernetes, other things they have now with open search.
Cool.
[00:11:19] Speaker B: Yeah, no. Anyone who can orchestrate clusters of. That is awesome. In my book.
[00:11:26] Speaker C: You did it. You just didn't. You lost hair because of it.
[00:11:30] Speaker D: Didn't sleep for a few days. It's fine.
[00:11:33] Speaker A: Not a few days.
[00:11:34] Speaker C: Yeah.
[00:11:35] Speaker A: Years of our life.
[00:11:36] Speaker C: Years. Yeah, there's the. There's the damage that occurred in the moment, and then there's the long term damage that we have yet to suffer.
[00:11:43] Speaker B: Yes, that's true.
[00:11:44] Speaker A: Yeah, yeah. If we could make it like a retroactive claim for therapy.
[00:11:49] Speaker D: Workers comp. Jonathan just submitted workers comp claim.
[00:11:54] Speaker C: I had to help Ryan. Well, I didn't really help. I just read his code. Not even his code, but someone's code that he inherited in Powershell. And it had been a while since I looked at Powershell, and I might need therapy for that, too, because it was spaghetti coat of monster that I. It's so bad.
[00:12:11] Speaker B: So bad. And it's so crucial. So it's. It's always those, right? It's the. The important ones where everything revolves around it and so no one touches it, and so they end up getting the worst spaghetti code. If it's old and tacked. Tacked on. It's always is.
[00:12:28] Speaker C: Yeah. Well, then, like, trying to figure out the sub module mappings and how it pulls in sub modules and the. Trying to chase back.
[00:12:34] Speaker B: Yeah.
[00:12:35] Speaker C: I'm like, I. I took me three.
[00:12:38] Speaker B: Months to understand that code, reverse engineering how it works, and now I've got it pretty down. So, like, when it breaks, I could fix it, but I wouldn't try to refactor it at all.
[00:12:50] Speaker C: Yeah. I think there was a moment where you were like, you posted a snippet of code. You're like, here's the problem. And I'm like, I don't. I don't freaking see that. And then I'm like, you're like, tank in the Matrix right now. All you see the characters walking across the green screen, you know, the green text of the Matrix coming across, and all I see is greenhouse.
And you were like, yeah, that's. That's pretty accurate.
[00:13:10] Speaker B: Sorry.
[00:13:11] Speaker C: Yeah, yeah.
Well, so anyways, Powershell code. Yeah, it's great. Don't. Don't use it. Python for the win or go lang or something.
[00:13:19] Speaker D: In windows, Powershell is actually getting pretty decent.
[00:13:22] Speaker B: It will say, yeah. The reality is that all this code's fault is not really the fault of Powershell.
I'm actually always impressed by actually how. How well this code executes, given how it's written.
And so. And a lot of that has to do with, you know, how powershell hands around objects and conserves resources that way. So it's. It's actually kind of cool.
[00:13:49] Speaker C: Yeah. If you didn't have to support it, it could be beautiful.
So support it still.
[00:13:56] Speaker B: Yeah.
[00:13:57] Speaker C: Well, if you want to have other beautiful code, let's talk about Rust.
For rust developers, you can now store and access your libraries known as crates on AWS code artifact. No additional rust packages beyond the gigabyte stored requests and data transfer will be charged to you. And so you get the ability now to use code artifact like everyone else with your rust packages and card.
[00:14:21] Speaker B: That is. I mean, I'm a fan of never running a software repository again. And so it's the support of these new features and functions that allows that managed service or usage. So I think it's fantastic.
[00:14:34] Speaker C: I'm mostly impressed. It's got a full AWS blog post write up. I'm like, you're just adding rust support to code artifact? And like, no, no, they did 700 words on rust and code examples and how to use it. And so if you're apparently Amazon might be getting big into Rust. I don't know, because this is a lot more than I expected, I think.
[00:14:51] Speaker A: So, because half the blog post is actually selling Rust as a programming language and saying how wonderful it is.
[00:14:56] Speaker B: Yeah, yeah, I've seen it happen. Like, I haven't tried to adopt Rust, but everyone that I know of, I've seen them happen like one by one, like becomes like that cultish Rust is amazing. So I think, you know, on, at some point I'll have to join them.
[00:15:14] Speaker C: But see, I see this happened to the Ruby community too. Everyone got really excited about Erlang and I looked at that and I was like, this is a bad call. Like, abort, abort. And everyone else fell into Erlang and then realized that was terrible. And they ended up in rust too, but they had to take that detour through Erlang. And I feel like Java is kind of the, rust is kind of the Java people all fall to and then, like someone's gonna realize that rust is terrible as well. Then something else will come out of it and then they all end up over there. I don't know if that is yet.
[00:15:40] Speaker B: Yeah, well, Java is terrible, so, I mean, I don't know if that's even, how terrible could rust be to be worse than that.
[00:15:48] Speaker C: I mean, I mean, that's possible is that Java is just so bad that Russ can't, can't possibly, like, I mean, because Erlang was bad and Ruby was not great either, as you guys like to point out to me all the time.
[00:15:58] Speaker D: Here's Peter, just to stick up for Ruby.
[00:16:01] Speaker C: Yeah, he's not here anymore. He left. So, you know, like all Ruby people do, they just leave.
They write a bunch of crappy code, they put it in the code and then they quit.
[00:16:11] Speaker B: And you'll never manage those dependencies ever.
[00:16:14] Speaker C: Exactly. Exactly.
[00:16:15] Speaker D: Ruby gem. Hell, that's all I have to say.
[00:16:19] Speaker C: I was reading about something with like some nvm environment manager for JavaScript. And I was just like, yeah, this reminds me, just like rvm just as bad, just as horrible.
[00:16:31] Speaker A: Anyways, I think a lot of the problem with is the popularity. The perceived popularity of things like rust, though, is a lot of it's come from people who've never used it. It's like that. They're buying into it on the sales points. Oh look, it's memory safe. It must be amazing. I don't know how many people are.
[00:16:48] Speaker C: Using it, but to me, could you imagine? Just like my entire reason for choosing a language is because it's thread safe or memory safe. Okay?
[00:16:56] Speaker A: I mean, there's valuable places you would want to guarantee things like that, but these are perfectly good tools for analyzing code anymore, which will solve that for you. Anyway. So do I think we're going to port the whole Linux kernel to rust? No. Are we going to start writing more device drivers and things in rust? Sure. I think it'll be a hybrid effort, but no one's going to do a mass rewrite of anything.
[00:17:17] Speaker C: So the Q 120 24 red monk programming language ranking has rust in the 90th percentile of popularity on GitHub and the 75th percentile on stack overflow by number of tags that mention rust. But then I get concerned because Kotlin and Scala are higher than it, and Ruby's over here to the right with typescript and Java, et cetera. So yeah, how popular it is or not, it's hard to say. I'm just going to say a swift is past objective c. That's objective c is tough.
[00:17:51] Speaker A: I don't show stack overflow is a great metric, it just shows what people have problems with, not what people.
[00:17:56] Speaker C: Well, that's why it's two dimensions. It's popularity, ranked by GitHub by number of projects that have language, and then popular rank on sitefold by number of tags. So, yeah, I mean, you can challenge roadmap methodology, I'm fine with it.
Well, back to Amazon. Amazon is pleased to announce that anthropic Cloud 3.5 sonnet model is now available to you in bedrock. They're the first one to announce it, but others will announce it today as well. Cloud 3.5 raises the industry bar for intelligence, outperforming other generative AI models on a wide range of evaluations, including Anthropoc's previously most intelligent model, Claude three opus cloud 3.5 sonnet. Key improvements include visual processing and understanding, writing and content generation, custom support and natural language processing, analytics and insights and coding and software development capabilities to help you write better code or a sonnet of code.
I played it and.
[00:18:52] Speaker A: And it wrote some code that ran. First time? What? First time?
Yep.
I just, I was like, written this and I skimmed through it like, yep, that looks good. Copy paste works. First time. It's awesome. I also had it generate. So that was, that was the LDAP group management stuff. I had it right?
And then I also had it write browser, browsers, game tic tac toe in HTML and JavaScript. And it just spat that out like it's done it a thousand times, which it probably has.
That worked fine. And I wanted to tweak it. It's like it's AI that it was playing tic tac toe with. So maybe a new version. And then I played it. It was like, no going back twice.
[00:19:37] Speaker C: Did you win?
[00:19:38] Speaker B: Did I win?
[00:19:39] Speaker A: Yeah.
[00:19:39] Speaker C: Did you win the tic tac toe?
[00:19:40] Speaker A: I did.
[00:19:42] Speaker C: So it's not that smart.
[00:19:43] Speaker A: It always goes second, so.
[00:19:45] Speaker C: Ah, well, that's bad. That's bad. Tic tac toe.
[00:19:51] Speaker A: I'll ask it to change that and see what happens. But no, I was super impressed.
[00:19:55] Speaker C: You should play a chess against it. See how your chess skills are against.
[00:19:58] Speaker A: I could try that. I did ask it about what his golf handicap would be and then told me that it would be better to have a golf hunter gap in the teens so that people didn't think it was too busy on the golf course instead of doing what it's supposed to be doing.
[00:20:11] Speaker C: I mean, apparently while you're vice president of America, you can get your handicapped Mando six. I think that's what he said.
[00:20:19] Speaker B: Yeah, who knows?
[00:20:23] Speaker C: All right.
[00:20:23] Speaker B: And we have a rule against talking about politics.
[00:20:26] Speaker C: I didn't. I just talked about a debate that.
[00:20:28] Speaker B: Happened that might be fair enough. Fair enough.
[00:20:32] Speaker C: Amazon is announcing the general availability of a fully managed ML flow capability on Amazon Sagemaker. ML Flow is a widely used open source tool playing a crucial role in helping machine learning teams manage the entire machine learning lifecycle, which is basically code to your ops team to do it for them. Customers can now effortlessly set up and manage MLFlo tracking servers with just a few steps, streamline the process and boosting productivity. MLFlow can be used to track multiple attempts at training models as runs with experiments, which I don't need to track how many times I've tried and failed. Compare these runs with visualizations and evaluate models, and register the best models to a model registry. There are three core components to ML flow capability on Sage banker, including the ML Flow tracking server, which is a standalone HTTP server serving multiple rest API endpoints for tracking runs and experiments, enabling you to better monitor your ML experiment. Next up is the ML Flow back end metadata store, which is critical part of the MLFlow tracking server, where all metadata related to experiments, runs and artifacts persisted. And finally, the ML Flow Artifact store, which provides a storage location for all artifacts generated during ML experiments, such as trained models, datasets, logs, and plots.
[00:21:37] Speaker A: That's interesting because ML Flow for years has been a competitor to Sagemaker in AWS. People have just deployed that and used it to manage running experiments and building models and the workflows around that. And so for them to bring it in like this is both interesting and kind of cool to have it as a managed service now but no longer a competitor anymore.
[00:21:57] Speaker C: Well, if you can't beat them, join them by making them part of your product.
[00:22:00] Speaker D: Wasn't ML flow built by databricks though, or is it running in there? I thought it was built by like.
[00:22:07] Speaker C: Open source, by, I don't know where ML Flow exactly came from. It might have been databricks and then they open sourced it.
[00:22:13] Speaker A: I don't know.
[00:22:14] Speaker C: But it's an open source product on its own now.
[00:22:16] Speaker A: It was, yeah, was open source, but by them.
There you go.
[00:22:22] Speaker C: I mean, what did Cloud Sonnet say?
Well, I'm just glad to have more options because the more people want to do ML and AI stuff, the more you realize how much tooling you need to merely make it successful. So having options is appreciated.
[00:22:39] Speaker A: Yep.
[00:22:40] Speaker C: So at FinopsX, Amazon announced the AWS compute optimizer is delivering you new recommendations for RDS, MySQL and postgres databases. These recommendations help you detect idle RDS instances and identify the optimal instance types and provisioned IOP settings for your existing RDS DB instance. So you can reduce costs for over provisioned RDS database instances or increase the performance of under provisioned workloads to right size RDS databases. It takes in multiple variables and can be very time consuming. And you need to dedicate engineering resources specializing databases to analyze the cost performance. So the right sizing recognition makes that all easy and machine learning driven. It makes recognition two parts. First, a focus on the DB instance size, and the second will be recognition for your DB instance storage.
[00:23:23] Speaker B: I was really happy to see that they included the IOP's provisioning in this optimization. Just so many database engineers, you know, they live by that, that, that one outage because of, you know, drive performance and all that. All the SQL engines say you need to have this and all these things, and then it just sits idle and it's so much cost.
So this is pretty great to see because I think that, you know, I think this is one area I know that most places I've worked where there's a heavy over provisioning.
[00:23:55] Speaker C: Well, a. Because it takes those really, really expensive people who know what they're doing to do that, which is definitely a factor, but it can be complicated very quickly.
[00:24:07] Speaker A: Yeah, I think next step, though, check the button and make it adjust it for me dynamically. Need more apps. Give me more apps. Just tell me what you're doing. Manage your. For me, it's still kind of like, I know ids isn't supposed to be a serverless service, but it could. It could be a lot closer to being a serverless type service where you don't need to think about that stuff.
[00:24:30] Speaker C: That'd be great. One thing it does annoy me slightly about rds for this is that they still don't support savings plans for RDS.
If it does tell you that, hey, you should move to an M class from a C class instance type, but you had rise on it. Your host.
[00:24:45] Speaker B: Yeah.
[00:24:48] Speaker C: That'S why my one complaint, the storage works at any time because you can change that basically anytime you want to, even on RDS. Ri. But I wish they would bring a savings plan or some form of savings plan that give me more flexibility here. Then this tool will be perfect.
[00:25:02] Speaker A: Yeah. But unlike Google, though, at least they have different classes of instances available when you need them.
[00:25:10] Speaker C: Too soon, John. Too soon.
Don't worry.
[00:25:13] Speaker D: Asher's not far behind.
[00:25:15] Speaker B: Not bitter at all.
[00:25:17] Speaker C: Yeah, we need this insight. We gave it to you. You didn't provision it fast enough.
[00:25:22] Speaker A: Yeah, exactly. Infrastructure is code. As long as you give us six months notice that you want to deploy something. Okay.
[00:25:27] Speaker C: And then when they give you, you wait three weeks, and they're like, here's the server. But you didn't provision it, so we gave it to somebody else. Yeah, because you just told me ten minutes ago, my code doesn't run that fast.
[00:25:39] Speaker A: Yeah, I get that. The quota is not necessarily representative of all the capacity they have available.
Yeah. Yeah.
[00:25:47] Speaker C: The quota thing is annoying, too, because even, you know, even being able to automatically set the quota doesn't give you anything. So you can automatically adjust the quota up to higher levels, and then they still don't have capacity for you.
So then, you know, it's the one time where I finally realized that reservations might be important for doctor.
Because on Amazon, I would never have thought about it, like, oh yeah, I need a reservation for these now. I'm like, yeah, I might need reservations in the doctor site because otherwise that capacity will not be there.
Well, AWS billing and cost management has two new data export types. First of all, you can now export the cost optimization hub recommendations to s three. And if you're excited about that focus 1.0 you in preview, you can export that to s three as well. So all of your Amazon Kerr data can now be in the new focus standard. So you can use focus for all your new reporting.
[00:26:39] Speaker B: That's pretty great because that report is terrible to parse.
[00:26:45] Speaker D: Come on, it's fun.
[00:26:49] Speaker C: As I was telling Matt before the show, just put your building data into bigquery. It solves all your problems. It's trying to run excel on your laptop at 100 million rows does not work well even in 64 bit mode.
[00:26:59] Speaker D: No, it really doesn't. I can tell you that from today's experience.
[00:27:04] Speaker C: Yeah, even my finapps guy, he asked me, he's like, hey, can I get a server just to load excel on? Because it kills my laptop every time I try to load the Google billing data. I was like, yeah, we could probably do that, but you should really learn bigquery.
It'll help you out a lot.
All right, well, apparently, according to a report from Business Insider, Amazon is developing a consumer facing AI chat bot that would complete with OpenAI's chat GPT and can be revealed later this year. New chatbot service codenamed Metis would be accessed via web browser and provided by a new foundational model, which they already have, Titan. So why do they need a new one? The move would join a growing list of technology companies building their own AI assistants, including Google and Microsoft.
Like I said, I assume this will be on top of Titan. They already sort of have a chatbot, which is Q, so I don't. Apparently they decided that's not what this thing's going to be powered by. But Metis, I assume, would become part of Alexa. If I were to be a guessing person.
[00:28:03] Speaker B: Yeah, that would be my guess as well, especially since they announced a recent overhaul of the Alexa service.
[00:28:09] Speaker A: Yeah, they got a child shoot too. They wouldn't give it away for free.
[00:28:12] Speaker C: I ordered maybe $10 a month.
[00:28:14] Speaker A: Yeah, no thanks. Yeah.
[00:28:16] Speaker C: So you want a device that you're gonna put into my house that you're already spying on me with? You want me to charge you $10 to make the spying more relevant? To who, me or you?
[00:28:26] Speaker B: Well, I mean, the promises were, you know, like it would, you could do complex workflows and it would detect workflows patterns and stuff, but, you know, who knows what that's actually going to be in practice.
[00:28:36] Speaker C: Yeah, kind of. I've kind of fallen out of love with my, my dingus devices over the last like six months. I just find I don't use them like I used to. I don't know if that's just because they're just getting worse or I've just moved on with my life.
[00:28:53] Speaker B: They're getting worse. I think at least the echoes that I have around here, because there's lately the support with third party apps too, for home automation that's drying up. So they're changing stuff and then deprecating the APIs, which is no good.
[00:29:13] Speaker D: Yeah, I feel the same way. We used to use them a lot, even for timers and alarms and stuff like that, and we've kind of moved on.
[00:29:20] Speaker C: I feel like, yeah, timer, the kitchen timer is probably the one last place that it still has a lot of value in our house.
And beyond that, the kids maybe play music on theirs in their bedroom sometimes, but that's it.
Well, moving on to GCP. Google regularly improves the cluster auto scaler and they want to shed some light on some small improvements they've been making that add up to big improvements for you. First up, target replica count tracking is a new feature that accelerates scaling when you add several pods simultaneously, which could be for new deployments or large resizes. It also eliminates a previous 32nd delay that affected gpu auto scaling scalability is headed to open source that the entire community can benefit from. The improved Kubernetes performance that the new cluster autoscaler will bring.
Next up is the fast homogenous scale up. If you have numerous identical pods, this optimization sees up scaling the process by efficiently bin packing pods onto nodes.
New less cpu waste, which the cluster autistic now makes decisions faster, which is especially noticeable when you need multiple scale ups across different node pools. And the memory optimization, which is not very visible to most people, but the cloud or sorry, the cluster auto has undergone memory options that contribute to its overall efficiency and improved utilization across your fleet.
[00:30:39] Speaker B: Very cool. I still see dummy workloads for pod scaling and, you know, kind of hackery in order to address this issue with some, some of my spiky workloads. And so like, this is, the more they add these improvements, the more we can sort of stop supporting all this stuff.
[00:31:00] Speaker C: Good.
[00:31:01] Speaker B: Awesome.
[00:31:03] Speaker D: Also, like other, you know, they are making a point to bring it into the mainstream. So, you know, while it's on GKE should move out to the rest of the community. So if you're in other clouds or on prem or anywhere else, you will get these features, which is nice.
[00:31:17] Speaker C: You just have to wait for them to get picked up into Kubernetes core and then you'll get them.
Well, we have a couple of Magic quadrants this week, courtesy of Gartner of course. First up is the magic quadrant for analytics and business intelligence platforms. This one again, I was kind of funny when Google's first one to announce it, but that isn't leading it.
So now on top, Microsoft was the leader, followed by Salesforce. But Google, click and Oracle were all at the same level. In this year's quadrant. AWS has climbed a bit too, which is now leading the challengers box and almost the point where it could cross over to leaders quadrant, but not quite there yet. They highlight strengths for Google, growth, market momentum, excellent composability and a strong development stack. Cautions Limited augmented analytics back penetration, which I know that is weak product execution of visual data preparation capabilities and unlimited native automated insights. Amazon got strengths of competitive pricing, integrated global cloud service and a service architecture that enables scale and performance. But they got cautioned on their limited deployment options because it's a cloud service. Governance capabilities rely on separated licensed Amazon products and gaps in data preparation capabilities and Azure got strengths and strong combination of price, functionality, ecosystem, the copilot brand closely associated with generative AIH and the power portfolio and product ambition of fabric, with our cautions being governance of content creation, publication, unproven interoperability with other services. And Azure is the only deployment option, which again, I hate it when they ding them on. This is a cloud solution that's only available on this cloud.
[00:32:50] Speaker B: Yeah, it's kind of a terrible practice.
[00:32:53] Speaker C: It's coming up more and more though, in the Gartner Magic quadrants I'm noticing.
[00:32:56] Speaker A: What does Copilot have to do with Genii have to do with this thing?
[00:33:01] Speaker C: Because you want copilot and all your analytics so you can ask it to make pretty reports with AI and natural.
[00:33:07] Speaker B: Language, they say, and automatically generate insights based on the data.
[00:33:13] Speaker D: Okay, yeah, I mean in theory I haven't played with it yet that much, but I know there's like Copilot for PowerPoint in Excel and like be able to do okay.
I used it mainly for email and other stuff and like teams catch up. I haven't used it in all the other tools. In all the other Microsoft office tools. Yeah.
[00:33:34] Speaker C: The biggest problem with all the copilots in Microsoft is there's too many of them and they all cost $30 a month.
[00:33:42] Speaker B: Yeah, no, the sprawl on AI is going to be madness because I already, I have access to two in the day job and I compare them frequently.
And GitHub Copilot was slightly better at answering questions about bigquery than Gemini was. But yeah, no, it's like if I, it's only because I'm not really paying for it yet. Is that okay? But like by the time you add up all the things that I actually want AI on, yeah, it's going to be like at least five or six or more. That's kind of a lot.
[00:34:17] Speaker C: Yeah, well, I mean, even just for my own personal, like I'm paying for copilot and I'm paying for Gemini for the cloud pod and I'm like, hmm, what? I really need to pay for both those. And I think, you know, Jonathan's trying paying for some of, some of he likes and then I have also LM studio and I have a bunch I downloaded that I like as well. So like how many AI's do I need?
[00:34:36] Speaker B: Yeah, I mean I'm cheap and miserly so I'm really focused on, you know, the LM studio action. But it's, you know, I don't have the hardware to really do anything fun with it.
[00:34:49] Speaker A: You want to get halves of me, just $15,000 each.
[00:34:54] Speaker D: It's one time though, it's, you know, per month.
[00:34:58] Speaker A: Yeah. Plus the megawatt of power needed.
[00:35:02] Speaker D: Right, but that's Jonathan's problem. You're paying for the hardware and can use it. Jonathan will pay for the power and operate.
[00:35:08] Speaker C: Jonathan has all the solar. He's good.
Well, the second magic quadrant was the data science and machine learning platforms. This one doesn't look that different except for databricks. Microsoft, Google and Amazon are all up in the top right quadrant and a bunch of other vendors I don't care about in other ones. Google got strength of foundational models, the balance portfolio of services and delivery of those services with cautions of data and I governance platform ecosystem and the core data science capability is their focus. Without as much machine learning, Amazon got security compliance as their strength, custom AI infrastructure with sagemaker and training and community as all strengths with precautions being their heavy focus on core data science and not ML, the complexity and the gen AI market sentiment that Amazon is behind. And then Microsoft gets strengths of enterprise gen AI capabilities upskilling research and development and cautions of separate product lines, data exploration and visualization. And copilot integration. So before they got, they got the plus on Copilot here, they got dinged on Copilot.
[00:36:13] Speaker D: I just like how Sagemaker is like a, you know, a positive there for AWS. I'm like Sagemaker has been there for like ten years. I feel like now it's been there for a long time.
[00:36:24] Speaker B: Yeah.
[00:36:25] Speaker D: Before nobody used it. Now everyone's like ooh, Sagemaker. And I'm like it's been there for a long time. Like you forget that.
[00:36:33] Speaker C: And then the ones that don't like Sage Maker, you show them vertex and they're like oh this is so much better. And it's like it's written by the same product manager that Google stole. It's just gen two of sagemaker.
[00:36:45] Speaker D: Don't worry about that.
[00:36:46] Speaker C: Yeah don't worry about that.
Well GCP had five cost management innovations that they announced at Finops x 2024 1st up a bigquery view that shows billing data and guess what? Focus format bigquery view that's available for free.
Speaking in the language of the business, not technology. With the new Gemini cloud assist to help you ask of natural language questions about all of your billing information. The Finops Hub is working through cloud environments for sustainability and carbon with new reporting on both of those.
Number four, a new scenario modeling for cuds with ability to look at different look back periods, eliminate data noise that is hey you're at a Black Friday because that only happens once a year and seeing instant results and collaborating with confidence on all of your CUD decisions. And number five, sending actual alerts for cost analyzer detection with easy email and pub sub capabilities as well as direct integration into slacken.
[00:37:37] Speaker B: Awesome.
[00:37:38] Speaker C: And the modeling for cuts was probably the highlight of the announcements from the vendors disability to do a lot of different windows. Amazon has some look back so thats like 30 or 90 days, a bunch of different carve out periods and then yeah being able to say like look this weekend where that outage happened and we spun up a bajillion Kubernetes clusters that kept failing or hey a Black Friday event or those sort of things. Thats nice. Be able to just eliminate that and say thats just a data anomaly. Count that into my cut analysis as I think through it.
[00:38:07] Speaker B: That is very cool.
[00:38:10] Speaker C: And then they announced cloud 3.5 sonnet on Vertex AI and it's super expensive.
Was bedrock also expensive?
[00:38:20] Speaker A: I didn't check the bedrock pricing, I didn't check bedrock. But on Vertex you pay differently for the input tokens and the output tokens, but effectively, since you wanted to be generating stuff because it's Genaida, $15 per million tokens, which is quite pricey.
Like the smaller model, the Haiku model is only $1.25 per million tokens. So just to put it in perspective, Opus is $75 per million tokens.
Bit of money, not a small model. I'm happy with my $20 a month to anthropic right now. I think.
[00:38:56] Speaker C: Yeah, I think you're a steal of a deal.
[00:38:58] Speaker D: You're winning.
[00:38:59] Speaker A: Yeah, I'd be even happier if I could just download the model.
[00:39:03] Speaker C: Right, well, then you can write it at Alamstudio. Yeah, perfect.
Data stream Google Cloud's serverless change data capture service recently introduced a new feature called Append only mode that streamlines the process of replicating changes from your operational databases to bigquery. This feature offers an efficient and cost effective way to maintain historical records and track changes to operational data over time, and traditional change to capture based replication. When a record in your source database is updated or deleted, the corresponding record in the destination is overwritten, making it difficult to track the history of those changes. Append only mode addresses this challenge by preserving every change as a new role on your target. Bigquery table and each row includes metadata that captures the type of change. Insert, update, or delete, a unique identifier, a timestamp, and other relevant information which can be used to order and filter the data as needed. Append only mode is particularly beneficial in scenarios where you need to maintain a historical record of changes. Some common use cases. It could include auditing and compliance, trend analysis customer 360 degree reviews, analyzing embedding, drift, and time traveling.
[00:40:04] Speaker B: Interesting.
[00:40:05] Speaker C: Yeah, I was like, this is cool. I could use this in some auditing use cases for sure.
[00:40:09] Speaker B: Time travel would be very interesting to me because that's one of the common things that I come across, which is, did this exist? Because it doesn't exist now. When did it stop existing?
[00:40:21] Speaker A: Yeah, it's just so much safer as well. The last thing you want to do is accidentally replicate deleting a bunch of data to someplace like Bigquery and have to repopulate it from somewhere very expensively. So I think time travel is a great feature.
[00:40:36] Speaker C: Yeah, I shared it with some people internally at the day job. This is an interesting way to maybe replace some of our auditing capabilities.
[00:40:44] Speaker A: Yeah, and I guess hopefully with well crafted queries.
Although you pay more for storage, it doesn't impact performance.
[00:40:53] Speaker B: Too much mongo right now.
[00:40:55] Speaker C: So.
[00:41:00] Speaker A: Google only takes cash, not souls. So.
[00:41:06] Speaker D: One day they'll figure out how to monetize your soul and then they'll take your soul.
[00:41:10] Speaker C: Yeah, well, Azure continues to be on summer vacation, but they sent their poor Finops people to Finops X instead of going on vacation. They don't get a vacation, they don't get one.
I have several FinOPs X announcements from Azure that they presented. First up is the Azure is adding an estimate total cost before deploying from the console. When you select ten of the server type, it'll now tell you the cost in the console, which please all other cloud providers. Add that this is my biggest complaint in AWS all time is like you're using the wizard. I've chosen to use the launch wizard because I'm click opting my way through this. You can at least give me the price what I want to do. You get new, deeper insights into Kubernetes costs with detailed AKs breakdowns the new advisor workbooks to help optimize your costs new and updated phenOps tools, including Azure optimization engine, cost optimization workbook and PhenOps Framework 2024 updates. You can export multiple data sets from billing, including focus 1.0 compliant data. So now everyone has it. And self service analytics via fabric. For those of you paying through the nose for fabric, yeah, they had to.
[00:42:15] Speaker D: Slide the fabric in there.
[00:42:19] Speaker B: I'm guessing they spent like a boatload of money getting that developed. And so because they're pushing that hard.
[00:42:24] Speaker D: Oh yeah, everywhere.
[00:42:26] Speaker C: If you're excited about all of those things, you're going to have to wait for the monthly finops update blog post because these are all from screenshots that I took of the presentation at the conference that they said on slide. I've yet to release blog posts for, so sorry about that. But hopefully in the monthly finops update they will have all of these for you to look at, which should be coming out next week because it's end of the month.
[00:42:49] Speaker D: Yeah, that's one of those announcements actually do look at because there's normally some good tidbits be buried somewhere in those announcements. They just, Azure really does like to announce the things that they're doing.
[00:43:01] Speaker C: I feel like they're really bad at announcing what they're doing. And the things they do announce are just so egregious. You're just like, why?
I mean, there's been stuff that they've talked about a lot of AI announcements that had nothing in them, like AI for this market segment. I'm like, I don't care. So we haven't been talking about those but even their updates, typically we're deprecating a thing or there's a new version of a thing that has barely any improvements because updated security updates, which isn't helpful either to talk about on the show. We cut through the noise on that.
[00:43:31] Speaker D: Or it's like it's in private preview, then preview and then public preview and then something else. Then there's Ga. So by the time you get there you're like I'm not interested because I'm not going to use it until it's ga'd or production workload.
[00:43:45] Speaker C: And then when they ga it, they don't actually re announce what it is. So they just give you like this thing is Ga. And I had to go back and look at what the preview was, actually figure out what the thing it's really like. They need a Jeff Barr at Azure. Like their dev rel is bad.
[00:43:59] Speaker D: Yeah, it's, it's a pain point for me on my day job because I try to keep up and keep track of things and there's just times I'm like, I don't know what you've actually released because all it says is this thing has been ga'd or we've added support for this thing in this place. I'm like cool. Anything real here?
[00:44:17] Speaker C: Well I was like as always refer back like well for those of us been with us since the early adopter version, we've made these significant changes and you're like cool. What was it before? So you had to, it takes a lot of work to go through those announcements, which I do for you so you're welcome listeners and me.
[00:44:35] Speaker D: And they definitely steal Justin's work.
[00:44:39] Speaker C: You're welcome. Thank you Oracle. Again, also at the Finops X they had announced something and they're announcing the launch of Phenops Hub Phenops spirit. By being on the floor and announcing their phenops hub product to all of us. Businesses contract cloud resource costs more closely than ever. However, within Ops hub helps practitioners to answer several questions including our budgets on track. Do we have enough credits left so we don't pay overages or get sued by Oracle? What resource groups have the greatest impact on cost? What can I optimize in my costs and how can I prevent unauthorized spending? Because I don't want the lawyers to attack me.
[00:45:16] Speaker B: I mean I like this but it is by all my temptation is just to make fun of Oracle as usual.
[00:45:24] Speaker D: I mean like that they like try to like actually make it, you know, wherever else like has it. They're like are we on track, like trying to make it actually be an actual item for a person versus like here's your budget. Or like here's a deeper dive into it. Like that feels like they went with the action oriented approach. Like am I over budget? Am I going to have to pay more money? It's my CFO going to yell at me type thing, you know, versus like here's integration to Kubernetes or whatever else. So I feel like they're, they went the action route of the business route which I think its interesting.
[00:45:56] Speaker C: I do like that angle. It is a nice change of pace. Heres why were trying to help you answer when you go google for this later. My CEO is yelling at me how can I save money? Ill be like heres a tool you can use thatll help you save money. So it does help in the SEO side as well.
[00:46:13] Speaker A: Well, budgets are so invisible to most people in the organization now, I think. And so if they're building tooling which makes it put in front of people and shows you what the budget is and shows you where you are on track or not on track, it's just a useful tool. It's a nice perspective.
[00:46:30] Speaker C: I can't relate to you, Jonathan, because I live in my budget.
Not being aware of your budget. What? That's a weird, weird thing. I don't know what that's like.
[00:46:45] Speaker B: We don't let you design anything though.
[00:46:47] Speaker C: That's true. Yeah. I just yell at you when you spend too much money for my budget.
Well, that's another fantastic week here in the cloud and we're heading into the 4 July holiday.
So hopefully for those of you who are going off of town, this podcast will find you well on the 4 July holiday.
[00:47:05] Speaker A: See you later guys.
[00:47:07] Speaker C: Everybody.
And that is the week in cloud. Check out our website, the home of the cloud pod where you can join our newsletter slack team. Send feedback or ask
[email protected] or tweet us with the hashtag hash poundtheclab pod.
All right, well, I have all four of you here so I've been saving an aftershow topic because we are evenly slipped between and Android users and iPhone users. Oh wow. And WWDC just happened.
[00:47:48] Speaker D: Jonathan, help me out.
[00:47:49] Speaker C: This is the time of year that I like to hear. You know, we're talking about what they announced and then Jonathan can say, yeah, Android's had that for five years. Yeah.
So they, so the highlights are the 13 highlights of it and there's a lot of other features that come out too. But first up is they have their own version of AI they're calling Apple intelligence. So it's going to be apparently in iPhone, iPad and Macintosh computers, which I found funny because they have Apple intelligence, which is going to have a bunch of its own compute servers that they're going to use to do off device AI computing. But if their AI cannot answer your question, you will have the option to send the data to OpenAI and then OpenAI chat GPT can answer the question for you. Which then Elon Musk basically got on Twitter and was like, if you put OpenAI into the iPhones, we'll ban iPhones and Apple devices on our campus. Which I thought was just. Okay, Elon, you're just mad because no one chooses you for Grok.
[00:48:46] Speaker B: Yeah.
[00:48:49] Speaker D: Is this supported by the vendor in any of the cloud vendors natively?
I don't think so.
[00:48:56] Speaker C: I don't know. It's not in bedrock. It's not anything that I'm aware of.
I don't think it's really even available other than in Twitter. And what's available in Twitter is nothing great, but I'm not paying for it either. So maybe that's my problem. It's maybe the key, but, yeah, the nice thing about this. So they are doing chat, they said first, but they are open to the idea of having Gemini. They also said they're going to be building out their own LM model. So that one, I think you guys have us on this one, Android, I think you guys are already integrating into chat GPT and Gemini, I assume. I already have chat GPT on my iPhone. I have an app, but having it natively in the operating system is kind of nice.
[00:49:33] Speaker A: Yep. Or have Gemini don't use it. It doesn't work properly.
[00:49:36] Speaker B: Perfect.
[00:49:38] Speaker D: Don't worry, it won't work on iPhone either.
[00:49:40] Speaker A: Yeah, I'm like, just doing basic things, like the kitchen time you mentioned earlier, like, hey, cool, start the time for five minutes. And so it doesn't open it in the clock app where you think it would do with a stopwatch. That persists, like if you turn your phone off or switch to a different browser. No, it opens up in a browser on a page. Really? Oh, yeah. It's crap. Really?
[00:50:04] Speaker D: Yeah.
It's actually the only feature I use the. Hey, Google feature for on my phone, which is to.
[00:50:11] Speaker A: Is that using Gemini, though? Is that using a Gemini interface or not? Because.
[00:50:15] Speaker D: No, I think I actually. No, I think you're right. I think I never authorized it to steal all my data.
[00:50:20] Speaker A: Yeah, well done. Bother.
[00:50:24] Speaker D: They already have Jonathan's got it.
[00:50:28] Speaker C: Well, that sounds riveting. I'm sold on the Android phone off that conversation. Yeah, Siri is getting an AI boost, so they're basically going to be again that chat GPT linkage from Siri, which anything to make Siri better is a win in my book because Siri is the worst.
And so I'm hoping when they get their own LLM that I don't have to have the chat GPT part of Siri, but that's gonna be nice. So I guess on the home assistants, which I have one that's in a box that I've never taken out of in my drawer. Does it have access to Gemini?
[00:50:59] Speaker A: Don't know, but I tend not to use those things anymore.
Like Ryan said, put them in a drawer and don't use them anymore. I've been playing with some free stuff like Mycroft and home assistant to try and get away from the cloud connected services.
[00:51:16] Speaker D: You know you do a cloud pod every week?
[00:51:19] Speaker A: Yeah.
[00:51:20] Speaker D: Or is that why.
[00:51:22] Speaker C: Yeah, one of the two we already talked about, chat GPT and Siri. So that one married. We're getting all these AI features and mail messages and photos. And the most important feature of this one, I think, is that we're getting custom emojis that you can put common text into and it'll create you a custom emoji of whatever text you put into the Apple assistant, uh, thing there. So you guys have that on Android?
[00:51:47] Speaker D: Yes.
Yeah, it's like, if I write a text message, like, how are you doing it? Like, pops up and gives you options, like different, like bubbles and bubble letters and other things like that.
[00:51:59] Speaker C: So apples do that too, where if you're in a text message and you write in like, chicken, it'll pop up. Like, would you like to replace the word chicken with the chicken emoji?
[00:52:06] Speaker D: Oh, it does that too. But this is just like any text. Like you like a, like a small phrase, it will give you the option to, like, write it in some fun letters and colors.
Clearly fun like that.
[00:52:20] Speaker A: I turned off those features for the messaging gap because, I don't know, I just, it wasn't send to be a message. I want to send my message the way I want it to somebody. So. But, yeah, I think pixel at least have the AI integrated things. I'm not sure if they do custom emojis yet, but they did, like custom backgrounds and various other custom things that you could, you could do.
[00:52:41] Speaker C: Nice, nice.
This one I know Android's had for a long time. We're getting more customizations for iOS home screens. So we don't have to have just sheets of apps anymore. We can actually put, we've been able to put widgets in place for now, but we're actually able to create. This is a clean space on my photo that I don't want obstructed by icons now as well as we're going to be able to do multiple home screens for different things. Like, I want a gaming home screen and I want a productivity and I want to work one. And I know you guys have had that for a while. Can we do that now in iPads? You can do sidecar, which is sort of similar, but this is like sidecar on steroids.
[00:53:16] Speaker B: Yeah, I can do a different home screen depending on my focus that you can do.
[00:53:20] Speaker C: But this is like actually saying, I only want. This home screen is only going to have five icons.
You can also decide that you don't want custom color, you don't want custom icons anymore from every. You can make them all the same color. You can theme them different ways. There's a lot of customization to the look and feel of your iPhone coming in iOS 18.
[00:53:40] Speaker D: The placement of apps was like one of the things that you drove me absolutely crazy when I had an iPhone was like, it had to go in the row and across and I was like, what? What if I want like it vertically? No, no, you can't do that. Like, you don't be creative like that. That's how I always felt it. Always. That was like the thing that drove me absolutely crazy with iPhones and to an absurd level. I don't know why.
[00:54:06] Speaker B: Yeah, no, I get it. Right. Because the minute you start playing around with widgets, they're also constricted in size and placement. And so it's like, well, I don't want those four apps. That's the only thing I have left for. I would rather those just be on their other page. But I can't do it. Yeah.
[00:54:23] Speaker A: Can you put them in folders?
Can I drag all into a folder?
[00:54:27] Speaker B: But I'd rather just that space be blank.
[00:54:32] Speaker A: Oh, wow. I can sell an app that has no icon, a transparent icon on the Apple store just for people to put on the home screen to black out certain spaces. Hell if I've known that.
Yeah. Three invisible app.
[00:54:49] Speaker C: Next up is RCs support, which Android has had for a while. This is supposedly going to make the green versus blue fight that we all have. Why we don't have a cloud pod chat team Jonathan and Matt in there. That's just terrible.
But apparently that's going to be coming out. We're rolling out that. I haven't heard of RCS. It's still going to be green text boxes. It is. I assume it's going to be.
[00:55:12] Speaker B: They did announce that it'd still be green, but it's going to support much more.
[00:55:17] Speaker C: You can do edits, you can do tap backs, you can do all kinds, all the things RCS supports. Basically you can do which. That's all I really wanted. Thank you Apple, actually. Thank you eU, for making Apple do this.
[00:55:28] Speaker D: Yeah, it's the sending of photos and like videos that were, that's like, you know, I have a kid is like from last week when I show up on time, you know, and, uh, you know, like, I have a group friend chat where, like, we're going to the farmers market, wherever somebody sends a picture. And like, if I'm in this thread with, you know, three of like, our other friend, that's a really good couple. I can't get the picture. It's crap. Or the video is like 160 by 160 pixels. I'm like, come on, guys.
[00:56:00] Speaker C: Well, get a group chat. Let you know in October.
Apple tv adds insights, which is the same thing as Amazon Prime. X ray basically gives you real time information on who is on your tv the scene is about, et cetera. Apparently Apple's going to be using machine learning to do this, which I think is a terrible choice. Amazon uses an army of indian people or Amazon Turks to basically specify the stuff for you. So you don't have to do this and we'll see which one works out better. I never have used x ray, I will honestly say so. I don't think I'll ever use insights, but it's there if I want it.
[00:56:36] Speaker B: Can you not use it? When you pause, it automatically comes up, right?
[00:56:40] Speaker C: I mean, do I read it? No. Yes, it pops up. And also, I don't opt that many prime shows.
[00:56:47] Speaker A: My tv has those features. Is it LG? I think LG has a feature where they look at what's on screen and then try and send you targeted ads or other things of interest and stuff like that. I turned all that crap off.
[00:56:58] Speaker C: I don't even care about that much. Passwords app. So I already had to move off the last password because I got hacked many times. And now one password. And now when one password gets hacked someday I will now have a passwords app on my iPhone or Mac devices that will all be synchronized so I can have my own standalone app. Yes, it's already been there. I know this is a standalone app, just like your one password or your lastpass, so you don't have to go into the keychain anymore to find your password, which was a terrible experience.
[00:57:24] Speaker B: That's true.
[00:57:26] Speaker A: So they changed something kind of like fundamental at the operating system level to support this. Now, assuming that the password app will have to be able to fill in forms in other apps.
[00:57:34] Speaker C: Yeah, they've had that in safari forever. And they even have a plugin for Chrome that talks to the passwords that are in your keychain. But if you wanted to go edit them or you had a desktop app that you need to get a password for, going into keychain to get your password was very painful.
[00:57:50] Speaker B: And the standalone apps have had that access on Apple to input into an application the password for a while now.
[00:58:00] Speaker C: Ipados is adding a calculator app, which for those of you who are on Android tablets are like, what? You didn't have a calculator? Your iPads. Yes. It's been a long suffering thing where you had to go buy a third party calculator app on iPads forever.
[00:58:14] Speaker B: Really?
[00:58:15] Speaker C: Yes, really. This has been.
[00:58:16] Speaker B: It's on iPhones?
[00:58:18] Speaker C: IPads, yes. There is a native iPhone app for calculator that has never been ported to iPhone or iPad until now.
And they've added pencil support. I don't know why.
And honestly the only reason why it's there is they added pencil support and there's ability to. You can go now, write out formulas with the pencil in the notes app or in the calculator app and it'll actually do the math equation now. And so this is just a way to demo. It is a cool feature. This is just a way to demo that pencil support. And so they're like, well, port the calculator app over, then we'll add this capability. And so now we have it finally on iPads, which is great, and we get a pencil support to drawn out equations. So I don't know if you guys have anything on the Android side that does this. Exactly. I'm sure a third party app does. I could have done it with a third party app on iPad as well, but this one looks pretty neat for iPad users.
[00:59:06] Speaker A: Native? No, I don't think so. At least not, not on stock Android. I know Samsung has a whole bunch of extra bloatware that they ship when they have pencil support and so do Motorola. I don't know about the calculator app though, but I tend to use like the Wolfram Alpha search, which would do the same thing like take pictures of equations and solve it for you for your homework if you need to.
[00:59:27] Speaker B: I didn't know it. I didn't know. I do pictures.
[00:59:32] Speaker C: Mac OS 15 will let you mirror your iPhone to your Mac OS. So you can control your Mac, your iPhone from your computer directly, which is kind of cool.
Which also is supposed to fix a bunch of the audio problems with continuity between the iPhone and the Mac. So this one you guys definitely don't have, but I appreciate this one.
[00:59:50] Speaker D: That is a cool feature though.
[00:59:52] Speaker C: Yeah, I said if you're a Mac developer or iPhone developer, I think that's something because right now the only way to do that on a Mac is you have to do an emulator inside of Xcode. And so, uh, yeah, that's a slow emulated experience, even though they're all arm processors now, you think it'd be fine. So slow and painful, and you could connect to your third party device, do it that way too. But now you'll be able to actually mirror it to your computer. So I think it's, I don't know how practical it's going to be. I like the idea of it, but we'll see.
[01:00:20] Speaker B: I can't think of a single use of that I would use because you're.
[01:00:23] Speaker C: Not an iPhone developer.
[01:00:25] Speaker B: That's true.
[01:00:27] Speaker C: Other than you do, is if you have your work laptop. If you have a Mac and your work laptop, you could connect messages to your work laptop, which is probably not the best idea because your work laptop is being spied on by your corporate overlord all the time and all your text messages could be sucked out of your Macintosh. So if you're just mirroring the iPhone, you'd be able to do text messaging through the mirror.
[01:00:47] Speaker B: So you're saying that the, the sharing won't be tied to your icloud account because I'm sure we tied to our iCloud account. Bad news.
[01:00:54] Speaker C: I'm sure it will be. But what you decide to sync to your device is a different conversation. I don't know. We'll see. This one. I don't know if I'll use it either, to be honest. But it's new Apple Watch get service widgets, which is this dumb thing on iPhones where basically we built an island, which you guys make fun of us all the time with our iPhone forehead island, that basically there's this thing that Apple developers have done where they basically, like, if you're going on a united flight, for example, it'll pop up in that dead space above the island. It'll tell you how long until your flight boards and all these things when you look at the home screen. So you get this live widget information that's in a very convenient place and that didn't exist on the watch and now it'll be on the watch too.
I don't know androids have watches but I don't know what they do or anything about them because I dont know anybody else. One.
[01:01:42] Speaker A: So whats it doing exactly? Its just like providing context, contextual information.
[01:01:47] Speaker C: Its like a real time widget that just lives. It isnt a widget it just embeds into the screen basically.
[01:01:52] Speaker B: So if I set a timer and then I close my phone then the timer still says visible on the top of my screen or when im listening if youre listening to Spotify or something this song information will be there and its dynamic and it sort of stays with the open apps but sort of doesn't so kind of this is sort of a helpful thing. I kind of like it.
[01:02:16] Speaker A: I mean we've got lock screen widgets so if you set a timer it'll, it'll send locks.
[01:02:20] Speaker C: Do you have to add that widget to your thing or does it.
[01:02:22] Speaker A: No no it just appears based on, based on content.
[01:02:25] Speaker C: Basically the feature you guys have had wherever.
[01:02:27] Speaker A: Yeah so like if you've got a flight in your calendar it will, it pops up contextual information about when to leave for the airport and whether it's late or not, that kind of thing.
But Google have been reading our emails for 20 years now.
[01:02:43] Speaker C: For those of you who have a vision OS two, which is not me because I cannot justify $3,400 for it, it'll support spatial photos and ultra wide Mac display. So if you're using your Mac inside your vision OS two, you can have a bigger screen. It's a weird one. And then AirPods Pro will let you quietly respond to Siri. So instead of being a douchebag in a party trying to respond to your AirPods, you can now just whisper to.
[01:03:09] Speaker B: I kind of like that right because that's one of my biggest gripes about all your voice activated anything is that it's like socially like you'll be I want to know something and then I have to be like you know stop a conversation and ask some robot something and it just, it feels very awkward. So I kind of like the idea of being able to do that a little quieter.
[01:03:30] Speaker C: There you go. So that's it. So yeah I think we basically confirmed that nothing we got was not already an Android other than maybe some weird quality of life improvements that you guys don't need.
[01:03:42] Speaker A: I mean, the mirroring thing is useful because Android has had mirroring for a long time, but it's only one way. So you can cast a screen to something else, but you couldn't.
There's no backward path to click on buttons and that kind of thing, unless you use the Android debugger. But then that's a whole. That's not a consumer product.
That is kind of cool.
[01:04:03] Speaker C: Wow, that was. All right. Cool.
[01:04:06] Speaker A: Yeah.
[01:04:07] Speaker C: Yeah.
Well, that's it. That's all I got.
You guys can bring to me the Android party. When you guys released version of Android, we don't even know which dessert are you guys on these days in Android world? Is that still a thing? The last time I had an Android was eclair, I'm pretty sure.
[01:04:22] Speaker B: Yeah, that's about the same time I had one.
That's a long time ago.
[01:04:26] Speaker A: I can't remember. They stopped making the.
[01:04:28] Speaker D: Well, they stopped pastry, sweets, whatever.
[01:04:32] Speaker C: They probably did because Google killed all the fun.
[01:04:35] Speaker D: But, yeah, it was upside down. 14 is upside down cake, 13 just for Justin and Jonathan and Ryan. Here is tiramisu.
[01:04:48] Speaker C: That's an inside joke there that no one will understand.
[01:04:50] Speaker B: No one will understand, but we will laugh about that. Yeah.
[01:04:53] Speaker D: And fifth geed, which is coming out, I guess third beta just came out, so I guess that's soon getting released is vanilla ice cream.
Like, they had a stretch with some of these ones.
[01:05:07] Speaker C: Yeah. Well, in the. In the Mac world, you know, we named macOS after places in California. So right now we're on Sonoma, but the new version will be sequoia, which I actually remembered that one. Sequoia was something I heard, like most of the names they choose, like Catalina and, uh, Monterey. Monterey. I never remember that. Most of the time, that's how that works. But Sequoia, I remember that one. I remembered it every time. It's come up like. Well, yeah, the new backwards. Oh, yeah. Sequoia.
[01:05:31] Speaker B: Is Sequoia a place?
[01:05:33] Speaker C: I don't know.
[01:05:35] Speaker B: Yeah. I'm just wondering if it's a departure. I mean, I'm sure there's a sequoia.
[01:05:40] Speaker D: Sequoia National Park.
[01:05:41] Speaker C: Sequoia National park. Yeah, that's a thing. Yeah.
Yeah, it's a place.
I mean, because they also had. What was the big rock? El capitan in Yosemite was a thing at one time.
[01:05:54] Speaker B: Oh, yeah. No, you're right. Yeah, that would be. It sort of fits in that theme. Yeah.
[01:05:58] Speaker C: Yeah.
[01:05:58] Speaker B: Okay.
[01:05:58] Speaker C: Yeah. So it's a California landmark. That's all you need to know.
I'm surprised they haven't used Golden Gate or anything.
Anyways, that's it. We'll see you guys next week.
[01:06:08] Speaker A: All right, see you later. Bye.
