[Saadia] Artificial Everything

Frustration at the current over-the-top obsession with AI. Also, Loofah's.

An async podcast by Saadia Carbis and Joshua Wold.

Chapters

• 0:00 Coming in from Germany!
• 0:50 Magic (and no magic)
• 2:29 Product feedback
• 4:54 2 kinds of product market opportunities
• 7:51 The AI hype is too much
• 10:59 Back scratchers, loofah's, and AppleCare

Transcript

0:00

Hi, welcome to Async. My name is Saadia. Async is a back and forth conversation between Joshua and myself about technology, app development and remote work. If you're new to the show, you won't hear Joshua's voice because we send these podcast episodes to each other back and forth between America and Australia because it's Async. And actually, I'm not coming to you from Australia like I normally do, sitting in my caravan with a microphone. I am at Cloudfest in Germany and we're here at Rust Europa Park and having a great time so far. I just arrived today after 25 hours of travel. I'm glad, Joshua, that you enjoyed my previous podcast episode. It was fun to record. I fell into, I could feel myself feeling, falling into a bit of a flow, falling in, I don't know, just the words were coming out. It just felt right as I was doing it. I could tell that there was some sort of magic being made there. So, I'm glad it came across. I recently published a blog post where I just knew as I was in a rush and I was going to publish it and I just knew as I was doing it that I was missing something. And I wasn't happy with it and the magic wasn't there. And I realized several hours after publishing it that I had neglected to make an important point. That was kind of, I don't know, kind of the point of the entire article. So, it's a good reminder not to rush things. Sometimes you have to rush. But it's also a good reminder that sometimes magic happens and sometimes it doesn't. And I think that it's better to have a podcast episode in the podcast feed that doesn't, that maybe is a bit awkward and has a few stumbling blocks than it is to have no podcast episode at all. Which is why these episodes are unedited. They're live to air. Because if we had to edit them, it would take too long and we probably wouldn't do it. So, but that's good because it makes it more real. All right. So, feedback time or, yeah, getting back to follow up. That's the word. Following up from last episode. Product feedback. How much product feedback do you take from customers? That was the question. And you made a really good point around having feedback loops. I completely agree with you. Feedback loops are essential. Feedback loops are these built in points of the product lifecycle. Or, you know, it doesn't have to be so fancy. Just moments while you're developing or designing. Where you can get a bit of feedback. And adjust as you go. You know, either self-reflection or more commonly somebody to just comment and say, oh, that looks a bit off. And there are moments when feedback is not required or unhelpful. And there are moments when feedback is good. But let's specifically talk about customer feedback. Because I think when it comes to customer feedback, people generally in our industry overestimate how important customer feedback is. That's my hot take. Like, I am sick of hearing all the time. Oh, well, we need to find a testable experience so that we can see how people like it and see how people use it. And then we can iterate from there. And it's just like a recipe for never actually publishing, never actually going into production, never actually doing anything live. And it annoys me. So, I am trying to say that feedback loops should not include customers at all until you've launched. Once you've launched, it's a different kettle of fish. Once you've launched, you've got a product out there and you want to see how people are using it and get feedback from your customers at that point. But before you launch, you need to have in your mind a clear problem that you're trying to solve. And you need to have confidence in your ability to solve it. You need to back yourself. And, yeah, you want to get feedback. But get feedback on your user experience from someone who knows user experience. Get feedback on your code from an expert developer. Don't go asking customers whether they think something looks pretty or is usable. Anyway, maybe there are two kinds of product market opportunities. You know, like you were talking on the previous episode how you've been working on something in the medical field and you can't just – you don't have that expertise necessarily. And even if you did, you can't just launch something without significant customer feedback. And because it's medical. If you're not doing it right, then people's lives could be a risk. And that is obviously completely valid. But maybe then we could separate these market opportunities into two types. Maybe there's market opportunities that do require complex solutions. Maybe they're in a critical field like medical. Or maybe it's just like the nature of the problem that you're solving is it just has to be complicated. There's no two ways around it. E-commerce is often a little bit like this. You can't create a simple e-commerce solution, although I actually have a couple of ideas around that. But for the most part, you can't – you know, it's just a complex thing. And so you've just kind of got to embrace the complex thing. But that type of product, it doesn't appeal to me. And this is just a personality thing, right? Maybe it appeals to some people. They want to solve really hard, meaty problems. And they've got a lot of different valences and a lot of different dependencies and can only be solved with taking into account a thousand different points of view. But I don't like those. I want simple opportunities, you know. Not saying these can't be just as lucrative. They can be. But maybe, like, maybe there's a competitor product. There's a product you're using that's really hard to use. It's really complex. And you think you can simplify the user experience. Or maybe there's a product that has just got too fast, too much scope. It's trying to do too many things. Maybe you think, oh, I can just really narrow in on my problem space and fix that one problem. Anything that's got a really, really simple solution to it, that is the type of product that I'm most interested in. And so, because of that, I think my hot take about not including customers is a little bit easier pill to swallow. Yeah. Not saying complex solutions aren't worthwhile, aren't worth your time. They certainly can be. It's just not something I, it doesn't fit my personality type, right? Like, I really, I think a lot of people like this, too. Especially developers want something that they can really just wrap your brain around easily. Something that's completely understandable. Something that's explainable. Maybe. Graspable. Right, I'm at this event, right? This CloudFest event. And it's a huge thing, right? There's AMD and Intel, our main sponsors, as thousands of attendees. And I've been at this first day of talks, and everyone cannot stop talking about AI. Artificial intelligence is the main thing. And it's starting to annoy me a little bit. I know that, like, people are saying, AI is like the next internet. It's, you know, completely going to transform everything and everyone. And that's all probably true. But I'm really starting to feel like it's being overhyped at this point. At what point does AI become overhyped? Can we be at that point? We might be at that point, Joshua. I think we might be there. People are constantly talking about Cursor and these sorts of tools. It really makes me feel uneasy. Because not only as a developer, you know, like, of course, there's some tension there. Is this going to take my job? The developers, are they really necessary? Overall consensus, by the way, is don't let Cursor write your code for you. It's a great tool if you know what you're doing. And if you don't know what you're doing, well, maybe you can prototype something. But do not put Cursor code into production without having a developer look at it. And that's good. But that's not always going to be true. You know, at some point, the AI models will probably be better than 95% of developers. And at that point, it's a different question. Anyway, this is an interesting topic. Will Cursor ever get better than it is? You know, will these AI models actually continue to improve? Maybe not. Because the amount of slop that they're generating now is just diluting the pool of resources that they have to learn from. And so maybe that's it. Maybe that's all she wrote in terms of how smart a model can get. I don't think that's the case. But there is a future where models don't get any smarter than they currently are. I don't know. So I'm just asking the question. Is AI overhyped? Is... I'm not saying it's Bitcoin. I'm not saying it's crypto that comes and goes or NFTs and we never see it again. It's here to stay. It's changing the world. And it will change the world. But the amount of attention that we're giving at it currently just... It's mind-blowing. Every single person here. Every talk. Every sponsor booth. Every advertisement is all about AI. And I just would love to see someone not talking about AI. I'm just sick of it. Maybe that's what it is. I don't know. All right. To answer your two smaller questions. I have never used a back scratcher. I would be interested. If you have a recommended Amazon purchase for a back scratcher. I would give one a go. As I occasionally feel the small itch in unreachable places on my back. But I do use a loofah. And I don't mean a plastic, round, squishy loofah that you buy from the shops. I mean we grow the loofah plant. And we peel it. And we use that in the shower. And it is excellent as a back scratcher. In the shower. I don't think you want to use it as a back scratcher in your regular work day. But in the shower. It's just the perfect combination of it's stiff enough to be able to reach down your back. And it's soft enough to not hurt. And it's, I don't know, just it's perfect. So I highly recommend if you can get yourself a hold of a real loofah. You should give that one a shot. Also, AppleCare on your Mac. Definitely. I broke the screen of my phone one time. And within the hour I was driving to the Apple Store to get it replaced. I think this is the way to go. I think you just need to, as soon as you see any small problem, you take it on the chin. You go to the Apple Store if it's under AppleCare. And you just get it done straight away. And that's the way to do it. Don't dither around. I'll be in Germany for a few days yet. So, you know, I might still have another Async episode to come.