Screaming in the Cloud

Richard Seroter, Director of Outbound Product Management at Google, joins Corey on Screaming in the Cloud to discuss what’s new at Google. Corey and Richard discuss how AI can move from a novelty to truly providing value, as well as the importance of people maintaining their skills and abilities rather than using AI as a black box solution. Richard also discusses how he views the DevRel function, and why he feels it’s so critical to communicate expectations for product launches with customers. 

About Richard

Richard Seroter is Director of Outbound Product Management at Google Cloud. He’s also an instructor at Pluralsight, a frequent public speaker, and the author of multiple books on software design and development. Richard maintains a regularly updated blog (seroter.com) on topics of architecture and solution design and can be found on Twitter as @rseroter. 

Links Referenced:

• Google Cloud: https://cloud.google.com
• Personal website: https://seroter.com
• Twitter: https://twitter.com/rseroter
• LinkedIn: https://www.linkedin.com/in/seroter/

Transcript

Announcer: Hello, and welcome to Screaming in the Cloud with your host, Chief Cloud Economist at The Duckbill Group, Corey Quinn. This weekly show features conversations with people doing interesting work in the world of cloud, thoughtful commentary on the state of the technical world, and ridiculous titles for which Corey refuses to apologize. This is Screaming in the Cloud.

Corey: Human-scale teams use Tailscale to build trusted networks. Tailscale Funnel is a great way to share a local service with your team for collaboration, testing, and experimentation.  Funnel securely exposes your dev environment at a stable URL, complete with auto-provisioned TLS certificates. Use it from the command line or the new VS Code extensions. In a few keystrokes, you can securely expose a local port to the internet, right from the IDE.

I did this in a talk I gave at Tailscale Up, their first inaugural developer conference. I used it to present my slides and only revealed that that’s what I was doing at the end of it. It’s awesome, it works! Check it out!

Their free plan now includes 3 users & 100 devices. Try it at snark.cloud/tailscalescream 

Corey: Welcome to Screaming in the Cloud, I’m Corey Quinn. We have returning guest Richard Seroter here who has apparently been collecting words to add to his job title over the years that we’ve been talking to him. Richard, you are now the Director of Product Management and Developer Relations at Google Cloud. Do I have all those words in the correct order and I haven’t forgotten any along the way?

Richard: I think that’s all right. I think my first job was at Anderson Consulting as an analyst, so my goal is to really just add more words to whatever these titles—

Corey: It’s an adjective collection, really. That’s what a career turns into. It’s really the length of a career and success is measured not by accomplishments but by word count on your resume.

Richard: If your business card requires a comma, success.

Corey: So, it’s been about a year or so since we last chatted here. What have you been up to?

Richard: Yeah, plenty of things here, still, at Google Cloud as we took on developer relations. And, but you know, Google Cloud proper, I think AI has—I don’t know if you’ve noticed, AI has kind of taken off with some folks who’s spending a lot the last year… juicing up services and getting things ready there. And you know, myself and the team kind of remaking DevRel for a 2023 sort of worldview. So, yeah we spent the last year just scaling and growing and in covering some new areas like AI, which has been fun.

Corey: You became profitable, which is awesome. I imagined at some point, someone wound up, like, basically realizing that you need to, like, patch the hole in the pipe and suddenly the water bill is no longer $8 billion a quarter. And hey, that works super well. Like, wow, that explains our utility bill and a few other things as well. I imagine the actual cause is slightly more complex than that, but I am a simple creature.

Richard: Yeah. I think we made more than YouTube last quarter, which was a good milestone when you think of—I don’t think anybody who says Google Cloud is a fun side project of Google is talking seriously anymore.

Corey: I misunderstood you at first. I thought you said that you’re pretty sure you made more than I did last year. It’s like, well, yes, if a multi-billion dollar company’s hyperscale cloud doesn’t make more than I personally do, then I have many questions. And if I make more than that, I have a bunch of different questions, all of which could be terrifying to someone.

Richard: You’re killing it. Yeah.

Corey: I’m working on it. So, over the last year, another trend that’s emerged has been a pivot away—thankfully—from all of the Web3 nonsense and instead embracing the sprinkle some AI on it. And I’m not—people are about to listen to this and think, wait a minute, is he subtweeting my company? No, I’m subtweeting everyone’s company because it seems to be a universal phenomenon. What’s your take on it?

Richard: I mean, it’s countercultural now to not start every conversation with let me tell you about our AI story. And hopefully, we’re going to get past this cycle. I think the AI stuff is here to stay. This does not feel like a hype trend to me overall. Like, this is legit tech with real user interest. I think that’s awesome.

I don’t think a year from now, we’re going to be competing over who has the biggest model anymore. Nobody cares. I don’t know if we’re going to hopefully lead with AI the same way as much as, what is it doing for me? What is my experience? Is it better? Can I do this job better? Did you eliminate this complex piece of toil from my day two stuff? That’s what we should be talking about. But right now it’s new and it’s interesting. So, we all have to rub some AI on it.

Corey: I think that there is also a bit of a passing of the buck going on when it comes to AI where I’ve talked to companies that are super excited about how they have this new AI story that’s going to be great. And, “Well, what does it do?” “It lets you query our interface to get an answer.” Okay, is this just cover for being bad UX?

Richard: [laugh]. That can be true in some cases. In other cases, this will fix UXes that will always be hard. Like, do we need to keep changing… I don’t know, I’m sure if you and I go to our favorite cloud providers and go through their documentation, it’s hard to have docs for 200 services and millions of pages. Maybe AI will fix some of that and make it easier to discover stuff.

So in some cases, UIs are just hard at scale. But yes, I think in some cases, this papers over other things not happening by just rubbing some AI on it. Hopefully, for most everybody else, it’s actually interesting, new value. But yeah, that’s a… every week it’s a new press release from somebody saying they’re about to launch some AI stuff. I don’t know how any normal human is keeping up with it.

Corey: I certainly don’t know. I’m curious to see what happens but it’s kind of wild, too, because there you’re right. There is something real there where you ask it to draw you a picture of a pony or something and it does, or give me a bunch of random analysis of this. I asked one recently to go ahead and rank the US presidents by absorbency and with a straight face, it did it, which is kind of amazing. I feel like there’s a lack of imagination in the way that people talk about these things and a certain lack of awareness that you can make this a lot of fun, and in some ways, make that a better showcase of the business value than trying to do the straight-laced thing of having it explain Microsoft Excel to you.

Richard: I think that’s fair. I don’t know how much sometimes whimsy and enterprise mix. Sometimes that can be a tricky part of the value prop. But I’m with you this some of this is hopefully returns to some more creativity of things. I mean, I personally use things like Bard or what have you that, “Hey, I’m trying to think of this idea. Can you give me some suggestions?” Or—I just did a couple weeks ago—“I need sample data for my app.”

I could spend the next ten minutes coming up with Seinfeld and Bob’s Burgers characters, or just give me the list in two seconds in JSON. Like that’s great. So, I’m hoping we get to use this for more fun stuff. I’ll be fascinated to see if when I write the keynote for—I’m working on the keynote for Next, if I can really inject something completely off the wall. I guess you’re challenging me and I respect that.

Corey: Oh, I absolutely am. And one of the things that I believe firmly is that we lose sight of the fact that people are inherently multifaceted. Just because you are a C-level executive at an enterprise does not mean that you’re not also a human being with a sense of creativity and a bit of whimsy as well. Everyone is going to compete to wind up boring you to death with PowerPoint. Find something that sparks the imagination and sparks joy.

Because yes, you’re going to find the boring business case on your own without too much in the way of prodding for that, but isn’t it great to imagine what if? What if we could have fun with some of these things? At least to me, that’s always been the goal is to get people’s attention. Humor has been my path, but there are others.

Richard: I’m with you. I think there’s a lot to that. And the question will be… yeah, I mean, again, to me, you and I talked about this before we started recording, this is the first trend for me in a while that feels purely organic where our customers, now—and I’ll tell our internal folks—our customers have much better ideas than we do. And it’s because they’re doing all kinds of wild things. They’re trying new scenarios, they’re building apps purely based on prompts, and they’re trying to, you know, do this.

And it’s better than what we just come up with, which is awesome. That’s how it should be, versus just some vendor-led hype initiative where it is just boring corporate stuff. So, I like the fact that this isn’t just us talking; it’s the whole industry talking. It’s people talking to my non-technical family members, giving me ideas for what they’re using this stuff for. I think that’s awesome. So yeah, but I’m with you, I think companies can also look for more creative angles than just what’s another way to left-align something in a cell.

Corey: I mean, some of the expressions on this are wild to me. The Photoshop beta with its generative AI play has just been phenomenal. Because it’s weird stuff, like, things that, yeah, I’m never going to be a great artist, let’s be clear, but being able to say remove this person from the background, and it does it, as best I can tell, seamlessly is stuff where yeah, that would have taken me ages to find someone who knows what the hell they’re doing on the internet somewhere and then pay them to do it. Or basically stumble my way through it for two hours and it somehow looks worse afterwards than before I started. It’s the baseline stuff of, I’m never going to be able to have it—to my understanding—go ahead just build me a whole banner ad that does this and hit these tones and the rest, but it is going to help me refine something in that direction, until I can then, you know, hand it to a professional who can take it from my chicken scratching into something real.

Richard: If it will. I think that’s my only concern personally with some of this is I don’t want this to erase expertise or us to think we can just get lazy. I think that I get nervous, like, can I just tell it to do stuff and I don’t even check the output, or I don’t do whatever. So, I think that’s when you go back to, again, enterprise use cases. If this is generating code or instructions or documentation or what have you, I need to trust that output in some way.

Or more importantly, I still need to retain the skills necessary to check it. So, I’m hoping people like you and me and all our —every—all the users out there of this stuff, don’t just offload responsibility to the machine. Like, just always treat it like a kind of slightly drunk friend sitting next to you with good advice and always check it out.

Corey: It’s critical. I think that there’s a lot of concern—and I’m not saying that people are wrong on this—but that people are now going to let it take over their jobs, it’s going to wind up destroying industries. No, I think it’s going to continue to automate things that previously required human intervention. But this has been true since the Industrial Revolution, where opportunities arise and old jobs that used to be critical are no longer centered in quite the same way. The one aspect that does concern me is not that kids are going to be used to cheat on essays like, okay, great, whatever. That seems to be floated mostly by academics who are concerned about the appropriate structure of academia.

For me, the problem is, is there’s a reason that we have people go through 12 years of English class in the United States and that is, it’s not to dissect of the work of long-dead authors. It’s to understand how to write and how to tell us a story and how to frame ideas cohesively. And, “The computer will do that for me,” I feel like that potentially might not serve people particularly well. But as a counterpoint, I was told when I was going to school my entire life that you’re never going to have a calculator in your pocket all the time that you need one. No, but I can also speak now to the open air, ask it any math problem I can imagine, and get a correct answer spoken back to me. That also wasn’t really in the bingo card that I had back then either, so I am a hesitant to try and predict the future.

Richard: Yeah, that’s fair. I think it’s still important for a kid that I know how to make change or do certain things. I don’t want to just offload to calculators or—I want to be able to understand, as you say, literature or things, not just ever print me out a book report. But that happens with us professionals, too, right? Like, I don’t want to just atrophy all of my programming skills because all I’m doing is accepting suggestions from the machine, or that it’s writing my emails for me. Like, that still weirds me out a little bit. I like to write an email or send a tweet or do a summary. To me, I enjoy those things still. I don’t want to—that’s not toil to me. So, I’m hoping that we just use this to make ourselves better and we don’t just use it to make ourselves lazier.

Corey: You mentioned a few minutes ago that you are currently working on writing your keynote for Next, so I’m going to pretend, through a vicious character attack here, that this is—you know, it’s 11 o’clock at night, the day before the Next keynote and you found new and exciting ways to procrastinate, like recording a podcast episode with me. My question for you is, how is this Next going to be different than previous Nexts?

Richard: Hmm. Yeah, I mean, for the first time in a while it’s in person, which is wonderful. So, we’ll have a bunch of folks at Moscone in San Francisco, which is tremendous. And I [unintelligible 00:11:56] it, too, I definitely have online events fatigue. So—because absolutely no one has ever just watched the screen entirely for a 15 or 30 or 60-minute keynote. We’re all tabbing over to something else and multitasking. And at least when I’m in the room, I can at least pretend I’ll be paying attention the whole time. The medium is different. So, first off, I’m just excited—

Corey: Right. It feels a lot ruder to get up and walk out of the front row in the middle of someone’s talk. Now, don’t get me wrong, I’ll still do it because I’m a jerk, but I’ll feel bad about it as I do. I kid, I kid. But yeah, a tab away is always a thing. And we seem to have taken the same structure that works in those events and tried to force it into more or less a non-interactive Zoom call, and I feel like that is just very hard to distinguish.

I will say that Google did a phenomenal job of online events, given the constraints it was operating under. Production value is great, the fact that you took advantage of being in different facilities was awesome. But yeah, it’ll be good to be back in person again. I will be there with bells on in Moscone myself, mostly yelling at people, but you know, that’s what I do.

Richard: It’s what you do. But we missed that hallway track. You missed this sort of bump into people. Do hands-on labs, purposely have nothing to do where you just walk around the show floor. Like we have been missing, I think, society-wise, a little bit of just that intentional boredom. And so, sometimes you need at conference events, too, where you’re like, “I’m going to skip that next talk and just see what’s going on around here.” That’s awesome. You should do that more often.

So, we’re going to have a lot of spaces for just, like, go—like, 6000 square feet of even just going and looking at demos or doing hands-on stuff or talking with other people. Like that’s just the fun, awesome part. And yeah, you’re going to hear a lot about AI, but plenty about other stuff, too. Tons of announcements. But the key is that to me, community stuff, learn from each other stuff, that energy in person, you can’t replicate that online.

Corey: So, an area that you have expanded into has been DevRel, where you’ve always been involved with it, let’s be clear, but it’s becoming a bit more pronounced. And as an outsider, I look at Google Cloud’s DevRel presence and I don’t see as much of it as your staffing levels would indicate, to the naive approach. And let’s be clear, that means from my perspective, all public-facing humorous, probably performative content in different ways, where you have zany music videos that, you know, maybe, I don’t know, parody popular songs do celebrate some exec’s birthday they didn’t know was coming—[fake coughing]. Or creative nonsense on social media. And the the lack of seeing a lot of that could in part be explained by the fact that social media is wildly fracturing into a bunch of different islands which, on balance, is probably a good thing for the internet, but I also suspect it comes down to a common misunderstanding of what DevRel actually is.

It turns out that, contrary to what many people wanted to believe in the before times, it is not getting paid as much as an engineer, spending three times that amount of money on travel expenses every year to travel to exotic places, get on stage, party with your friends, and then give a 45-minute talk that spends two minutes mentioning where you work and 45 minutes talking about, I don’t know, how to pick the right standing desk. That has, in many cases, been the perception of DevRel and I don’t think that’s particularly defensible in our current macroeconomic climate. So, what are all those DevRel people doing?

Richard: [laugh]. That’s such a good loaded question.

Corey: It’s always good to be given a question where the answers are very clear there are right answers and wrong answers, and oh, wow. It’s a fun minefield. Have fun. Go catch.

Richard: Yeah. No, that’s terrific. Yeah, and your first part, we do have a pretty well-distributed team globally, who does a lot of things. Our YouTube channel has, you know, we just crossed a million subscribers who are getting this stuff regularly. It’s more than Amazon and Azure combined on YouTube. So, in terms of like that, audience—

Corey: Counterpoint, you definitionally are YouTube. But that’s neither here nor there, either. I don’t believe you’re juicing the stats, but it’s also somehow… not as awesome if, say, I were to do it, which I’m working on it, but I have a face for radio and it shows.

Richard: [laugh]. Yeah, but a lot of this has been… the quality and quantity. Like, you look at the quantity of video, it overwhelms everyone else because we spend a lot of time, we have a specific media team within my DevRel team that does the studio work, that does the production, that does all that stuff. And it’s a concerted effort. That team’s amazing. They do really awesome work.

But, you know, a lot of DevRel as you say, [sigh] I don’t know about you, I don’t think I’ve ever truly believed in the sort of halo effect of if super smart person works at X company, even if they don’t even talk about that company, that somehow presents good vibes and business benefits to that company. I don’t think we’ve ever proven that’s really true. Maybe you’ve seen counterpoints, where [crosstalk 00:16:34]—

Corey: I can think of anecdata examples of it. Often though, on some level, for me at least, it’s been okay someone I tremendously respect to the industry has gone to work at a company that I’ve never heard of. I will be paying attention to what that company does as a direct result. Conversely, when someone who is super well known, and has been working at a company for a while leaves and then either trashes the company on the way out or doesn’t talk about it, it’s a question of, what’s going on? Did something horrible happen there? Should we no longer like that company? Are we not friends anymore? It’s—and I don’t know if that’s necessarily constructive, either, but it also, on some level, feels like it can shorthand to oh, to be working DevRel, you have to be an influencer, which frankly, I find terrifying.

Richard: Yeah. Yeah. I just—the modern DevRel, hopefully, is doing a little more of product-led growth style work. They’re focusing specifically on how are we helping developers discover, engage, scale, become advocates themselves in the platform, increasing that flywheel through usage, but that has very discreet metrics, it has very specific ownership. Again, personally, I don’t even think DevRel should do as much with sales teams because sales teams have hundreds and sometimes thousands of sales engineers and sales reps. It’s amazing. They have exactly what they need.

I don’t think DevRel is a drop in the bucket to that team. I’d rather talk directly to developers, focus on people who are self-service signups, people who are developers in those big accounts. So, I think the modern DevRel team is doing more in that respect. But when I look at—I just look, Corey, this morning at what my team did last week—so the average DevRel team, I look at what advocacy does, teams writing code labs, they’re building tutorials. Yes, they’re doing some in person events. They wrote some blog posts, published some videos, shipped a couple open-source projects that they contribute to in, like gaming sector, we ship—we have a couple projects there.

They’re actually usually customer zero in the product. They use the product before it ships, provides bugs and feedback to the team, we run DORA workshops—because again, we’re the DevOps Research and Assessment gang—we actually run the tutorial and Docs platform for Google Cloud. We have people who write code samples and reference apps. So, sometimes you see things publicly, but you don’t see the 20,000 code samples in the docs, many written by our team. So, a lot of the times, DevRel is doing work to just enable on some of these different properties, whether that’s blogs or docs, whether that’s guest articles or event series, but all of this should be in service of having that credible relationship to help devs use the platform easier. And I love watching this team do that.

But I think there’s more to it now than years ago, where maybe it was just, let’s do some amazing work and try to have some second, third-order effect. I think DevRel teams that can have very discrete metrics around leading indicators of long-term cloud consumption. And if you can’t measure that successfully, you’ve probably got to rethink the team.

[midroll 00:19:20]

Corey: That’s probably fair. I think that there’s a tremendous series of… I want to call it thankless work. Like having done some of those ridiculous parody videos myself, people look at it and they chuckle and they wind up, that was clever and funny, and they move on to the next one. And they don’t see the fact that, you know, behind the scenes for that three-minute video, there was a five-figure budget to pull all that together with a lot of people doing a bunch of disparate work. Done right, a lot of this stuff looks like it was easy or that there was no work at all.

I mean, at some level, I’m as guilty of that as anyone. We’re recording a podcast now that is going to be handed over to the folks at HumblePod. They are going to produce this into something that sounds coherent, they’re going to fix audio issues, all kinds of other stuff across the board, a full transcript, and the rest. And all of that is invisible to me. It’s like AI; it’s the magic box I drop a file into and get podcast out the other side.

And that does a disservice to those people who are actively working in that space to make things better. Because the good stuff that they do never gets attention, but then the company makes an interesting blunder in some way or another and suddenly, everyone’s out there screaming and wondering why these people aren’t responding on Twitter in 20 seconds when they’re finding out about this stuff for the first time.

Richard: Mm-hm. Yeah, that’s fair. You know, different internal, external expectations of even DevRel. We’ve recently launched—I don’t know if you caught it—something called Jump Start Solutions, which were executable reference architectures. You can come into the Google Cloud Console or hit one of our pages and go, “Hey, I want to do a multi-tier web app.” “Hey, I want to do a data processing pipeline.” Like, use cases.

One click, we blow out the entire thing in the platform, use it, mess around with it, turn it off with one click. Most of those are built by DevRel. Like, my engineers have gone and built that. Tons of work behind the scenes. Really, like, production-grade quality type architectures, really, really great work. There’s going to be—there’s a dozen of these. We’ll GA them at Next—but really, really cool work. That’s DevRel. Now, that’s behind-the-scenes work, but as engineering work.

That can be some of the thankless work of setting up projects, deployment architectures, Terraform, all of them also dropped into GitHub, ton of work documenting those. But yeah, that looks like behind-the-scenes work. But that’s what—I mean, most of DevRel is engineers. These are folks often just building the things that then devs can use to learn the platforms. Is it the flashy work? No. Is it the most important work? Probably.

Corey: I do have a question I’d be remiss not to ask. Since the last time we spoke, relatively recently from this recording, Google—well, I’d say ‘Google announced,’ but they kind of didn’t—Squarespace announced that they’d be taking over Google domains. And there was a lot of silence, which I interpret, to be clear, as people at Google being caught by surprise, by large companies, communication is challenging. And that’s fine, but I don’t think it was anything necessarily nefarious.

And then it came out further in time with an FAQ that Google published on their site, that Google Cloud domains was a part of this as well. And that took a lot of people aback, in the sense—not that it’s hard to migrate a domain from one provider to another, but it brought up the old question of, if you’re building something in cloud, how do you pick what to trust? And I want to be clear before you answer that, I know you work there. I know that there are constraints on what you can or cannot say.

And for people who are wondering why I’m not hitting you harder on this, I want to be very explicit, I can ask you a whole bunch of questions that I already know the answer to, and that answer is that you can’t comment. That’s not constructive or creative. So, I don’t want people to think that I’m not intentionally asking the hard questions, but I also know that I’m not going to get an answer and all I’ll do is make you uncomfortable. But I think it’s fair to ask, how do you evaluate what services or providers or other resources you’re using when you’re building in cloud that are going to be around, that you can trust building on top of?

Richard: It’s a fair question. Not everyone’s on… let’s update our software on a weekly basis and I can just swap things in left. You know, there’s a reason that even Red Hat is so popular with Linux because as a government employee, I can use that Linux and know it’s backwards compatible for 15 years. And they sell that. Like, that’s the value, that this thing works forever.

And Microsoft does the same with a lot of their server products. Like, you know, for better or for worse, [laugh] they will always kind of work with a component you wrote 15 years ago in SharePoint and somehow it runs today. I don’t even know how that’s possible. Love it. That’s impressive.

Now, there’s a cost to that. There’s a giant tax in the vendor space to make that work. But yeah, there’s certain times where even with us, look, we are trying to get better and better at things like comms. And last year we announced—I checked them recently—you know, we have 185 Cloud products in our enterprise APIs. Meaning they have a very, very tight way we would deprecate with very, very long notice, they’ve got certain expectations on guarantees of how long you can use them, quality of service, all the SLAs.

And so, for me, like, I would bank on, first off, for every cloud provider, whether they’re anchor services. Build on those right? You know, S3 is not going anywhere from Amazon. Rock solid service. BigQuery Goodness gracious, it’s the center of Google Cloud.

And you look at a lot of services: what can you bet on that are the anchors? And then you can take bets on things that sit around it. There’s times to be edgy and say, “Hey, I’ll use Service Weaver,” which we open-sourced earlier this year. It’s kind of a cool framework for building apps and we’ll deconstruct it into microservices at deploy time. That’s cool.

Would I literally build my whole business on it? No, I don’t think so. It’s early stuff. Now, would I maybe use it also with some really boring VMs and boring API Gateway and boring storage? Totally. Those are going to be around forever.

I think for me, personally, I try to think of how do I isolate things that have some variability to them. Now, to your point, sometimes you don’t know there’s variability. You would have just thought that service might be around forever. So, how are you supposed to know that that thing could go away at some point? And that’s totally fair. I get that.

Which is why we have to keep being better at comms, making sure more things are in our enterprise APIs, which is almost everything. So, you have some assurances, when I build this thing, I’ve got a multi-year runway if anything ever changes. Nothing’s going to stay the same forever, but nothing should change tomorrow on a dime. We need more trust than that.

Corey: Absolutely. And I agree. And the problem, too, is hidden dependencies. Let’s say what is something very simple. I want to log in to [unintelligible 00:25:34] brand new AWS account and spin of a single EC2 instance. The end. Well, I can trust that EC2 is going to be there. Great. That’s not one service you need to go through that critical path. It is a bare minimum six, possibly as many as twelve, depending upon what it is exactly you’re doing.

And it’s the, you find out after the fact that oh, there was that hidden dependency in there that I wasn’t fully aware of. That is a tricky and delicate balance to strike. And, again, no one is going to ever congratulate you—at all—on the decision to maintain a service that is internally painful and engineering-ly expensive to keep going, but as soon as you kill something, even it’s for this thing doesn’t have any customers, the narrative becomes, “They’re screwing over their customers.” It’s—they just said that it didn’t have any. What’s the concern here?

It’s a messaging problem; it is a reputation problem. Conversely, everyone knows that Amazon does not kill AWS services. Full stop. Yeah, that turns out everyone’s wrong. By my count, they’ve killed ten, full-on AWS services and counting at the moment. But that is not the reputation that they have.

Conversely, I think that the reputation that Google is going to kill everything that it touches is probably not accurate, though I don’t know that I’d want to have them over to babysit either. So, I don’t know. But it is something that it feels like you’re swimming uphill on in many respects, just due to not even deprecation decisions, historically, so much as poor communication around them.

Richard: Mm-hm. I mean, communication can always get better, you know. And that’s, it’s not our customers’ problem to make sure that they can track every weird thing we feel like doing. It’s not their challenge. If our business model changes or our strategy changes, that’s not technically the customer’s problem. So, it’s always our job to make this as easy as possible. Anytime we don’t, we have made a mistake.

So, you know, even DevRel, hey, look, it puts teams in a tough spot. We want our customers to trust us. We have to earn that; you will never just give it to us. At the same time, as you say, “Hey, we’re profitable. It’s great. We’re growing like weeds,” it’s amazing to see how many people are using this platform. I mean, even services, you don’t talk about having—I mean, doing really, really well. But I got to earn that. And you got to earn, more importantly, the scale. I don’t want you to just kick the tires on Google Cloud; I want you to bet on it. But we’re only going to earn that with really good support, really good price, stability, really good feeling like these services are rock solid. Have we totally earned that? We’re getting there, but not as mature as we’d like to get yet, but I like where we’re going.

Corey: I agree. And reputations are tricky. I mean, recently InfluxDB deprecated two regions and wound up turning them off and deleting data. And they wound up getting massive blowback for this, which, to their credit, their co-founder and CTO, Paul Dix—who has been on the show before—wound up talking about and saying, “Yeah, that was us. We’re taking ownership of this.”

But the public announcement said that they had—that data in AWS was not recoverable and they’re reaching out to see if the data in GCP was still available. At which point, I took the wrong impression from this. Like, whoa, whoa, whoa. Hang on. Hold the phone here. Does that mean that data that I delete from a Google Cloud account isn’t really deleted?

Because I have a whole bunch of regulators that would like a word if so. And Paul jumped onto that with, “No, no, no, no, no. I want to be clear, we have a backup system internally that we were using that has that set up. And we deleted the backups on the AWS side; we don’t believe we did on the Google Cloud side. It’s purely us, not a cloud provider problem.” It’s like, “Okay, first, sorry for causing a fire drill.” Secondly, “Okay, that’s great.” But the reason I jumped in that direction was just because it becomes so easy when a narrative gets out there to believe the worst about companies that you don’t even realize you’re doing it.

Richard: No, I understand. It’s reflexive. And I get it. And look, B2B is not B2C, you know? In B2B, it’s not, “Build it and they will come.” I think we have the best cloud infrastructure, the best security posture, and the most sophisticated managed services. I believe that I use all the clouds. I think that’s true. But it doesn’t matter unless you also do the things around it, around support, security, you know, usability, trust, you have to go sell these things and bring them to people. You can’t just sit back and say, “It’s amazing. Everyone’s going to use it.” You’ve got to earn that. And so, that’s something that we’re still on the journey of, but our foundation is terrific. We just got to do a better job on some of these intangibles around it.

Corey: I agree with you, when you s—I think there’s a spirited debate you could have on any of those things you said that you believe that Google Cloud is the best at, with the exception of security, where I think that is unquestionably. I think that is a lot less variable than the others. The others are more or less, “Who has the best cloud infrastructure?” Well, depends on who had what for breakfast today. But the simplicity and the approach you take to security is head and shoulders above the competition.

And I want to make sure I give credit where due: it is because of that simplicity and default posturing that customers wind up better for it as a result. Otherwise, you wind up in this hell of, “You must have at least this much security training to responsibly secure your environment.” And that is never going to happen. People read far less than we wish they would. I want to make very clear that Google deserves the credit for that security posture.

Richard: Yeah, and the other thing, look, I’ll say that, from my observation, where we do something that feels a little special and different is we do think in platforms, we think in both how we build and how we operate and how the console is built by a platform team, you—singularly. How—[is 00:30:51] we’re doing Duet AI that we’ve pre-announced at I/O and are shipping. That is a full platform experience covering a dozen services. That is really hard to do if you have a lot of isolation. So, we’ve done a really cool job thinking in platforms and giving that simplicity at that platform level. Hard to do, but again, we have to bring people to it. You’re not going to discover it by accident.

Corey: Richard, I will let you get back to your tear-filled late-night writing of tomorrow’s Next keynote, but if people want to learn more—once the dust settles—where’s the best place for them to find you?

Richard: Yeah, hopefully, they continue to hang out at cloud.google.com and using all the free stuff, which is great. You can always find me at seroter.com. I read a bunch every day and then I’ve read a blog post every day about what I read, so if you ever want to tune in on that, just see what wacky things I’m checking out in tech, that is good. And I still hang out on different social networks, Twitter at @rseroter and LinkedIn and things like that. But yeah, join in and yell at me about anything I said.

Corey: I did not realize you had a daily reading list of what you put up there. That is news to me and I will definitely track in, and then of course, yell at you from the cheap seats when I disagree with anything that you’ve chosen to include. Thank you so much for taking the time to speak with me and suffer the uncomfortable questions.

Richard: Hey, I love it. If people aren’t talking about us, then we don’t matter, so I would much rather we’d be yelling about us than the opposite there.

Corey: [laugh]. As always, it’s been a pleasure. Richard Seroter, Director of Product Management and Developer Relations at Google Cloud. I’m Cloud Economist Corey Quinn, and this is Screaming in the Cloud. If you’ve enjoyed this podcast, please leave a five-star review on your podcast platform of choice, whereas if you’ve hated this podcast, please leave a five-star review on your podcast platform of choice, along with an angry comment that you had an AI system write for you because you never learned how to structure a sentence.

Corey: If your AWS bill keeps rising and your blood pressure is doing the same, then you need The Duckbill Group. We help companies fix their AWS bill by making it smaller and less horrifying. The Duckbill Group works for you, not AWS. We tailor recommendations to your business and we get to the point. Visit duckbillgroup.com to get started.