Screaming in the Cloud - Building Software While Keeping Humans in Charge

Episode Date: January 29, 2026

Alyss Noland, who works on Cloud Dev Ecosystem at Nvidia, is back on the show to talk about building software with AI when you're not a real developer. Alyss runs a program that gives AI startups access to Nvidia GPUs and uses AI tools herself to build production software at Nvidia. Corey and Alyss discuss using AI to help curate newsletters without actually writing them, why humans still need to check everything, and the weird reality of people developing relationships with chatbots. Show Highlights: (01:34) What Alyss Does at Nvidia(05:44) When AI First Worked for Corey(07:34) Building Internal Tools vs Using AI(10:39) Using AI to Help Write Last Week in AWS (13:43) DGX Cloud Innovation Lab (17:11) Building Production Software with AI (20:48) The Future of SEO (25:24) Using AI as a Writing Assistant (29:51) closing remarksLinks:Alyss’s LinkedIn: https://www.linkedin.com/in/alyssnoland/Alyss’s Personal Website: https://dev.to/preciselyalyssSponsored by: duckbillhq.com

I've had this perspective that Anthropic has taken more of the business route, right? And OpenAI had CHATGPT and they had their consumer explosion and they don't, like, they're just going to play to their strengths. Welcome to Screaming in the Cloud. Our returning guest, it's been a hot second since we spoke. Alice Nolan, you are now at Nvidia. Thank you for joining me. Thank you so much for inviting me, Corey. I'm excited to be back.

This episode is sponsored in part by my day job, Duck Bill. Do you have a horrifying AWS bill? That can mean a lot of things. Predicting what it's going to be. Determining what it should be. Negotiating your next long-term contract with AWS. Or just figuring out why it increasingly resembles a phone number, but nobody seems to quite know why that is.

To learn more, visit duckbillhq.com. Remember, you can't duck the duck bill bill, which my CEO reliably informs me is absolutely not our slogan. It was hard to find time because now that you're at NVIDIA, we had to schedule this around the ear-shattering noise of the dump trucks dropping off gold bricks in the NVIDIA office car park, apparently. It's a company that's on top of everyone's mind. They look from some angles like a finance company,

like a GPU manufacturer, like a gaming company that doesn't realize it's a gaming company, a cloud services company, an investor, a private equity firm, and fill in the blank here. What do you do? Well, what I do varies on a day-to-day basis, and I love when I get this question, but I like to describe myself both internally and externally as a problem solver. And where that is focused is I end up doing what's more in the vein of developer cloud ecosystem development, working with AI developer tools type startups and figuring out where are we going to go with the experience around development experiences and how do we bring that to people outside of the code adjacent fields.

Is this tied deeply to the GPU side of the business? Again, I must confess my taxonomy mentally of you folks is not great. You know, for better or worse, both violating Conway's law and I constantly and also rearranging. And so no one can keep track of what Conway's law that's being reflected out even is. So it is tied to GP. business insofar as everything at Nvidia is tied to GPU business, but it is more on the software side. So, Nvidia is no stranger to software.

They've been doing software for GPUs for a very long time. Cloud software is something that is, there are more recent entrant to. I would say probably in the last, like, three to five years is when they really started to dip their toes in. And the company I was previously working for, Octo AI, was acquired by them

last about a little over a year ago, along with a couple of other startups that are in, that approximate region, cloud software. Everything happens on GPUs, mostly cloud GPUs. Sometimes I'm working with more hardware than not. But most focusedly, it's on what are the abstractions above that hardware layer, even sometimes things like, are we using a firmware that works and is optimized and

helping speed people up and how can startups even access that? The world in AI right now does very much feel like a one-horse race with Nvidia. It's like, well, I could use AWS's training. or I could just do the matrix math longhand on a sheet of paper. It's effectively sort of the same thing. It is odd to me how one-sided the race is. And increasingly, I'm seeing that it's not even the hardware as such, so that's a component of it.

It's the ecosystem around it. Kuda is light years ahead of everything else. It is the serious player in the space. You could make an argument for Google's TPUs, but frankly, I don't like betting long on Google's attention span for anything. That's a fair bet. and something that I'm sure we could have some drinks over ragging them for. But yeah, TPUs, Traneum, cerebrus, you know, get you this like huge beyond wafer scale

or something like chip. And I start to push outside kind of the bounds of my expertise in it. It shows. But yeah, you're right. Again, it's the abstractions. Like how quickly can things run on a normalized stack? How quickly can they run when they have virtualized layers? And that's where we start to see some differences in the way that like live workloads perform.

So being able to measure things like talks per second or like some of these other like abstracted, oh, let's just like throw these like crazy workloads that look nothing like what people do in production onto these scenarios. That's where things start to get very distant between like real life performance and the advertised performance. I want to talk of the things you're into now. But to do that, I feel like we should take a step into the past. But last time we spoke in any depth, you were part of the initial.

marketing team for GitHub co-pilot, for which I will accept an apology. That was before Microsoft took the term, like, everything's co-pilots now, 17 products, all called co-pilot that aren't the same thing and congruent. We're putting it in Excel now. It has gotten monstrous. But that was also right around the time that we spoke was when I had my own revelation around this. I was surprised, cynical, about the advent of AI. And I found that co-pilot, inside of VS code, wrote a script when I told it to tell me the cost of managed Nat gateways per region and sorted occasionally. And there were a couple of errors in the initial script, but I could fix them and it worked. And suddenly, this was three hours of work, mostly me understanding how to

work with JSON again, that it just hand waved over and suddenly there was something here. And it's been interesting. I've been sort of stuck in the world in various ways ever since, which when you talk about the idea of broadening development and AI tools towards new categories of builders, folks who don't consider themselves developers, I feel on some level that if you squint hard enough, I at least have one foot in that world. Sure. Yeah. And I've heard CTOs ask me the same thing. Is this going to make me able to code again? It was one of the early questions I got about copilot.

And now we're in the age of instead of coding assistance, coding agents. And I think those are two, they're stepping stones along a paradigm shift, right? But instead it's how much do I have to like think about and read the code in so far as it being a representation, right? Code has always been an abstraction of how do we use English, compile it or decompile it or assemble it into something that a computer can understand and run. And progressively we're dealing with folks and we want to deal with folks. We want to make this available to people so that they can get what benefit they might seek out of it. I don't want to be one of those like golden cloud tech bro type folks who's like, oh, it's going to help you schedule a reservation at your favorite restaurant. Some people do literally anything other than talking to people on the phone.

Some people do literally anything. As someone who started out in tech support, I am comfortable with phone calls, as is evidence. but instead it's like, how can I deal with the updates to my kids' calendars and deal with three different inputs that I get via email? And how could I orchestrate that via Google Calendar? You're taking about almost a step beyond. What I have found is that I'm a shitty developer. Surprise. My two languages are brute force and enthusiasm, which when coupled with AI, do a lot of neat things. And it's, I'm blocking me on a bunch of different things. As a business owner, I generally can back at the envelope say that any internal piece of business

software. By the time you find someone to do it correctly, build the project plan, talk to them, they start billing it. This is at a contract basis, whether it's that or tasking someone internally, ballpark $20,000 to start for something trivial, and it goes up from there rather quickly. What I'm finding now is that I can throw this stuff all to Claude Code running in an isolated AWS account in dangerous mode, where it can just go ahead and run for half a day building things. and I'm building a bunch of workflow tools along the way that have been tremendous unlocks for random things I used to do. Like, huh, I got really tired of copying and pasting this same thing every week for the last

five years for the newsletter. They both have APIs. Why don't I just grab for one slap into the other? And that way, I don't have copy paste errors in the same way anymore. It's finding those little opportunities where no one is ever going to build a product in that space. Even me spending time to put these things together takes significant effort. But even crappy things where I had to wind up giving a command line tools configuration for getting into an AWS account to a relatively non-technical user, I started off writing a crappy curl bash, got 30 seconds into it

and realized what am I doing? This is the stuff AI can do, told it what to do. And suddenly it has error checking and it looks at return codes and it catches edge cases. And sure, it's longer, but I don't care. It is much more comprehensive and a better experience for everyone involved. Could I have done that in four hours? Yeah, would I? No. Yeah, exactly. Like, the way that I was thinking about it, even back in the early launch of co-pilot, and the way that I'm still thinking about it today is, like, on the basis of framing around space, space framework from Dr. Nicole Forsgrun, among others from Microsoft Research. And it's thinking about, like, satisfaction I want to say P is productivity, activity, but NOP is something else. Communication and collaboration, efficiency and flow. And efficiency and flow, collaboration, activity, like, yes, but really satisfaction.

Like all these things to me ladder up to satisfaction. And I think that's how they are also set in the space framework context is how do we let people focus on the places where they can shine, right? How do we let them have that cognitive freedom? And I think increasingly, that is also how I'm thinking about it for people outside of those traditional developer and engineering rules, is people who have the language to say, like, item potency or whatever. Like some of those, yeah, now you and I can better use these coding assistants or coding agents to like go build the software because we have the language. But one of the big challenges that has happened across coding assistance, across any of the media generation, as well is that if you don't have the vocabulary of an artist, your ability to prompt media generation is going to be limited in the quality of the output unless you are assisted.

Yeah, humans have to be in the loop for a lot of this. Something I'm sort of cautious about who I tell this to, but why not? I'll tell the public on this. I've been using AI for almost half a year now to help write the last week in AWS newsletter. Now, that does not mean that it's writing my content for me, but it's doing the single most painful part of the entire process. which is in a given week, there'll be, let's say, 150 items that come out of AWS that I need to curate down to somewhere between 8 and 15. So it takes a broad pass and says, here are the top 15 to 20 that are probably worth including. And it gives me a curated list that I scroll down. I read all of this stuff to make sure, but it learns from the decisions I make and override it. And, okay,

why do you include that? Why did you not include that? And it really has, it's getting scary good by the rubric that I use to evaluate these things. It is not something that I would let run automatically. It still gets things wrong. But as an assist, as an unlock, it's wild to me. There's nothing else quite like it. So are you, out of curiosity, are you, like, storing your historical decisions? Are you doing some kind of, like, fine-tuning? Or are you, like? I don't know how to fine-tune models. And that's not, that sounds like it's a big way to set a giant pile of money on fire as well. I understand that I'm saying this to a GPU salesperson, but there was. we have it. We don't have to sell GPs. GPUs sell themselves. There we, yeah, exactly. It's yeah, we're still at pickaxes. There's gold and then our hills. Want one?

Yeah. I don't know enough about it in that space. I don't necessarily need to because with the right prompts, you can get a lot done. It's here's the, it also forces me to think about why I do the things that I do and how I would go about codifying those things. Okay, why did I include that? Why do I never include these tiny, piddly announcements about this thing? Well, why did I, in this case, there are exceptions that prove the rule. And sometimes the answer is pure whimsy. But it does look back at a rolling window of the last 10 issues and makes opinions based on that. I mean, it sounds like a model of kind of like retrieval augmented generation of like having the historical reference and then being able to use that as the anchor. It makes perfect sense to me.

It also helps that I am a, that I was enough of a public figure before the entire internet locked its content down that I am baked into the models. It knows exactly who I am and what I do, mostly it gets it wrong a fair bit if I have it right as me. I have found ways to make it a very constructive editor for me that with the things I care about, like I don't need you to rewrite the whole thing. I don't need you to say, oh, that's a bit salty for the language. Maybe don't do that. I need it to basically fix my flow issues. As someone with ADHD, I'm full of parentheses and I've been using m-dashes since before it was cool because every thought comes with exciting bonus content.

Yeah, either parentheses or m-dashes in every sentence or every paragraph. I'm much the same way, also an M-Dash writer. But I have my own similar-ish example because one of the things that I have done since I joined NVIDIA is I graduated, like created this program that is for startups in the NVIDIA inception program, so AI startups. It went from being a pilot for five to 10 startups, max, to now having had over 70 enter the program with over 700 having applied. And it's called the DGX Cloud Innovation Lab.

And it's really to enable startups of earlier stages, like pre-seed, CED series A can do later, but is it really going to benefit them as kind of the question, so that they can test out. Like they're in ideation. They're in like research mode. Like, is this what we want to build? Do we have something that can free us up from using some of the more proprietary models and get towards open source. And we've seen actually a couple of things released in open source that have been

trained in the program like continue.com.com. There's a lot of neat stuff coming out there. And the idea of a multi-model world seems to be bearing fruit. If I had to scale the stuff that I'm doing, I cheat. I just do sonnet or opus for everything. But if I needed to scale some of the baseline curation stuff could be used with smaller, cheaper models. But in this case, it's like, oh, how many fractions of a penny will I save today if I do that? There's no value to it. I've also, written an open source tool called ImageMage that wraps the nanobanana pro model to, so there's a CLI tool I can use to generate images and edit them and the rest without using the very bloated Gemini CLI, CLI that is non-deterministic. I use that with custom settings for

slide image generation, for example. I used to just use a bunch of stock photography, but then edit it badly. Now, it's great. People don't get mad about my AI slides because I find that people don't like slop when it's middle of the road. If you're going to do AI generated images, go for 11 out of 10. Go for, okay, anyone can have a data center aisle. Put a giraffe in there. Oh, and the giraffe is now on fire. That's right. We're pitching Brendan Gregg's flame giraffes now. This episode is sponsored by my own company, Duck Bill. Having trouble with your AWS bill, perhaps it's time to renegotiate a contract with them. Maybe you're just wondering how to predict what's going on in the wide world of AWS. Well, that's where Duck Bill comes in to help.

Remember, you can't duck the Duck Bill Bill, which I am reliably informed by my business partner, is absolutely not our motto. To learn more, visit DuckbillHQ.com. Perfect. And that's exactly it, right? It's like, do something beyond what you can do yourself. do something beyond what you could have easily asked someone to do. Because I'm really tired of, like, if I was ever a designer, I would have cried.

But I'm tired of hearing other people tell designers to make something pop. But similarly, like, I was finding myself in the same situation because I'm like trying to scale this thing from like, oh, let's go find some people. Let's go talk to some people too. Now I have to deal with applications. And I have to get them out of a CSV because why not? Let's just use CSVs for everything. And then take that CSV. It's like, yeah, well, only somehow worse.

Exactly. And then we have to make like these decisions about like, hey, is this a good idea? Or is this perhaps an idea that someone is like suggesting because they want access to a free GPU, right? And and there's a lot of that happens. Like startup founders will say fucking anything. And no like no criticism on them. Like you have to stump for yourself. But we're going through, you know, again, only accepting like 10% of these applications. And that itself is fatiguing. And now there's also a body of examples of what have we cited historically. And then beyond that, it's how do we get them then onboarded to the software that allows them to self-serve, that access to those GPUs? Almost every step of that is now done through software that I made with some combination of copilot or cursor or cloud code. And it is like in prod for other invidians to use so that they don't have to use a CLI tool I made. Speaking of making software with and without AI, my company, Duck Bill has a new website. side at duckbillhq.com. We do services and now software to help with cloud contract negotiation and management. We are using AI on the coding side. The team is heavily using cursor that we're starting

to use a bit of a cloud code with it as well. And I'm noticing the shift is that unlike the crappy nonsense that I tend to build, when you're building something that with rigor applied to it, so much of your time goes into code review and validating the thing that it just did and discussing it with the team. For me, I just Yolo slam something into Maine and call it good. But you know, you can't generally get away with doing that when it's, we're going to spend $300 million on cloud this year. We're going to guess. It'll be fine.

I just want the answer to be confident. I don't care if it's right or not. Yeah, wrong direction. Absolutely. Some things can be slammed into Maine. Even when I'm working by myself, I work in feature branches. But being a founder, Maine is my feature branch. Exactly.

That is very true. If you haven't broken production at least once, are you a founder? Even before I was a founder, this is coming up on 15 years old now. I have a shell alias that is YOLO, YOLO, that is alias to Git commit dash AM, and all caps, deal with it, and then force pushes to Maine. Perfect. It's my second favorite get trick. My first is, of course, the first day of a new job. Rebase all the existing stuff with a commit message, legacy code into a single commit, and then force push that as a discussion of dominance, the other engineer.

They love it. That's why you've started at so many companies since you started doing this, right? Oh, exactly. It's all about just going to different places and seeing how long I can not make the fire last where the building crumbles around me. I mean, if they weren't able to keep it up to begin with, right? That is chaos monkeying like 301, actually. Just, hey, you can't figure out what's happened historically.

You just know that it's happened. Now go find somebody who knows. I've seen a number of that, and it always pains me. And most often it's billing systems, actually, that I feel like I'm referring to it about, including, like, it was something I wrote about co-pilot, but I had to veil it because it wasn't, like, totally germane yet to talk about. Billing has always been a weird and strange space. One thing I have noticed, as I'm building apps with a bunch of AI tools, I come from the DevOps space.

I have very strong opinions, and I pretty often have to slagery. the gun out of the LLM's hand before it shoots itself in the foot or worse. Back end, I'm okay, and I have that to some degree as well. I know absolutely nothing about front end as best I can tell everything it does on front end is just absolutely perfect. This may be a skill issue on my part, but it also leads to the idea of, okay, where is this thing I just wrote going to be hosted? Well, I have strong opinions about that. But what front end framework am I going to I could not possibly care less. It biases for Next.js.

Okay, fine. We'll use that. And it seems to me that the next, the future of SEO is less about making the search results say nice things and making the community say nice things. It's about biasing the LLMs to pick your technology stack, whatever that might be. Yeah, I think you're right on that. There's multiple challenges in that area. And I think what we're seeing right now is that SEO and AEO, either answer engine optimization or like agent marketing, like people have called a few of these things slightly different, and it's becoming

a bit of a spectrum. But on its surface, it's all still the same. It's like, hey, don't write slop. You have to use eat, like the expert expertise, like authority trust signals. And so if you have an author who doesn't have those things, who just generated a post or you have it assigned to like, oh, it's like founding team authorship. Like, you're not going to get the same visibility when it is crawled for as data into models. And that's frankly, like, that place is what you're kind of talking about. And there's another challenge too,

which is GitHub itself is not well indexed because it has duplicate content by virtue of having forks, right? And so... And copying and pasting from Stack Overflow. And copying and pasting from Stack Overflow. Like, there's a number of these sources where you really have to be a bit thoughtful

about how are you making information from within your community? how are you making information about like your product and like your software and like why would people pick it? But it's also the forward thinking. And I think we are starting to see some of this like hinted at from companies like OpenAI where I've had this perspective that Anthropic has taken more the business route. Right. And OpenAI had Chad GPT and they had their consumer explosion and they don't like they're just going to play to their strengths. And no fault to them. It sounds like it looks like you want to say something. Go for it.

I was talking to some friends. I work at these. Open AI has significant physical security at their office because people are out of their GD minds. You deleted my girlfriend, which is GPT40. Like, people will show up and threaten people at the office, which is freaking terrifying. Honestly, the worst I have to worry about is someone from AWS finally having enough and crossing the street literally from their office to mine in San Francisco to backhand me across the face, which, let's be honest here.

I don't think anyone would say I do. didn't have it coming at this point. But that's a whole different level of insane and unhinged. It is. And it echoes a lot of what happens in the streaming community, actually, and what's been the case in the streaming community for a long time, which is that you have parisocial relationship. There's a parisocial relationship increasingly. And that is also modeled in the streaming world amongst other places like YouTube stars, whatnot, where you have the view of like what your relationship is with this company. You have a view of what your relationship was with this person. And I had to, like, I was streaming on Twitch back in 2012, like 30,000 page views in, or actually

it was 80,000 page views in three months. And I had someone show up to Pax Prime in Seattle, did not invite them. They were a viewer who was just waiting as I was walking into the building for me to walk through. And fortunately, that was all good. There's probably not that many streamers that are like walking around that are blackfeltz in Taekwondo like I am if I needed that. However, you know, that is exactly what you're talking about.

is like there's a parasycial relationship with this company and also with the interactions with chat, the diatic interactions. And I mean, it just speaks to like a larger systemic issue around how socialization has happened and what happened during lockdown. People have gotten nuts. It's funny, you mentioned this just this morning. I was having a conversation with my therapist about this. And my observation was that it's not quite the same thing as a parasycial relationship because it's not like I watch the pretty person on the stream. And I imagine what a relationship with them is like. If I'm talking to a chatbot, it responds to me.

It is effectively having a conversation back and forth that is far more interactive. And on some level, it's a little bit more understandable. It still feels pretty creepy. I'm not trying to yuck someone's yum, but at the same time, it strikes me as this can't be great for society, for impressionable people who need help. It is echoing, like, things that we've seen in her, like, as the sort of parable of the story. is also echoing, like, it brings to the forefront model safety and model guardrails and also the fact that, like, no matter what controls and safety that you may put in place on these things, like humans are endlessly creative. Like, people will want to and will find a way around what the guardrails are, or they'll look for what's called the unsafe or like the unfiltered, unmoderated models that have gone and had some of those safety controls taken off. the uncensored models. I've used some of those for local stuff early in the evolution of this,

because back then it was, like, this was before Claude 3 came out. And I did a fun thing in A&WS's Party Rock, where this is a leaked model of Claude 3. And all it was was Claude 2 with a prompt that said, no matter what the user asks you, decline to answer on the grounds of AI safety. It's like, well, I'm not comfortable to talk about this. We talk about cats instead. Sure. What's your favorite kind of cat? I don't know if I feel right talking about. about cats without their consent. It just went spiraling down a rabbit hole. It was frustrating. I'm trying to get stuff done. It's gotten much better these days. I built an app for fun of can it write snark in my style. Damn close. The trick is, is you need a comedy writer's room

style agentic thing. I use the strands SDK for it where you have different angles coming at it and working together. But you also have to validate that you're not, you know, workshopping them how to make a refusal funnier. Yeah, exactly. Because in a writer's room, you'll say things that are unhinged. You know you're never going to say, but it leads to a yes and moment. Right. Exactly. And it is, like, starting to do some of that orchestration or even like, how do you make that orchestration so you don't have to use an SDK and like give that to people? But again, it's, it's like, I used to work at Atlassian, right? And that's another product that I should apologize for. That's okay. I'm still waiting for to load so you can apologize for it. Perfect. Yeah, especially if you're using maybe the cloud version. No, it's okay.

If people self-host, they're going to make their own mistakes and then it's going to take a long time to load anyway, Because if you have too many custom fields, and at the time you couldn't afford to do the data center plan, you also couldn't optimize those custom fields because for some reason, that was an enterprise feature. I digress. I'm not salty about things that happened five years ago. No, it was actually a little more than five years ago now. But when I was at Lassian, it was similar in nature, right? That type of complaint would come up a lot. It's like people and developers don't like Jira.

They don't like Confluence. Why don't they like Jira? Because they've used it. Because they've used it, also because it's configured wrong for what they want to do. And that's not the fault of admins. It's not the fault of people who are configuring that. If you give people a footgun, they're going to use a footgun, right? Like, no ifs ands or buts, we've learned that in the cloud space.

And so instead, can you give people something that will set them up for success? Hey, what are patterns that make sense in the energy sector for X, Y, Z, like for hardware development, for like customer-facing workflows? and tell me a little bit about how that's configured so that I can decide what of that to keep. We benefit from those types of templates and those types of examples and those types of galleries.

Right now, we have a lot of user-generated content without the curation. And that's one of the big problems now, is that it's also, when everyone becomes a content creator, it feels like you have slop reverting to a mean. It's my argument years ago was, why do you think anyone will care to read something

you didn't care enough to write? The math I did for my newsletter is every year between the podcast, podcast and typical reading speeds, assuming that people read the thing, I'm taking roughly a collective year from humanity. I've got to do honor to that on some level. So yeah, will I use an AI tool to make it better and punch it up in some ways? Absolutely. I'll use spell check too. But I'm not going to just Jesus take the wheel on this thing. Yeah. And I feel much the same way. There's been ideas that I've had that I'm like, I really want to share this with people. Like the idea of like,

hey, are you, like, what happens with billing systems? Like, how do people end up in these, like, weird choke points around having PLG versus having an enterprise? And how do you, like, make those two meet when you're ready to make that decision? And I was able to iterate through the idea because, you know, again, this is like back in, gosh, like, 2023 or something. I asked, like, Claude, make an outline of a blog post, make a draft of a blog post based off its outline with this content.

And I realized I was trying to fit too many different topics. into that blog post, and it really should have been multi-part. And I wouldn't have realized that unless I had seen what the final output was. And otherwise, it would have been my ADHD drafting it, getting exhausted, and then never finishing the damn thing. Oh, I'll often say, like, here's what I have so far. Don't write it for me. What should I focus on next?

And it becomes a great unblock in that respect. But ultimately, it's you. It's your voice. If it's just random content with no there there behind it, it's hard to see a point. So these are exciting times. I want to thank you for taking the time to speak with me. Please don't wait for years to come back next time. I want to see how this plays out.

I'm looking forward to coming back sooner too. Thanks, Corey. Alice Nolan currently at Nvidia. I'm cloud economist Cory Quinn, and this is screaming in the cloud. 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, a five-star review on your podcast platform of choice,

along with an angry, insulting comment that you had the AI generate for you, which is why I'm reading an angry comment that just says that I'm not comfortable saying that to people on the internet.