Jerod Santo: The recurring cycle of ‘developer replacement’ hype

Danilo Alonso has seen it all before:

Every few years, a shiny new technology emerges that promises to make software developers obsolete. The headlines follow a predictable pattern: “The End of Coding,” “Anyone Can Build Apps Now,” or my personal favorite, “Why Your Five-Year-Old Will Be Programming Before Learning to Read.”

The executives get excited. The consultants circle like sharks. PowerPoint decks multiply. Budgets shift.

And then reality sets in.

In this post, he counts the rotations of the ‘endless carousel of replacement promises’, from the no-code/low-code revolution, to the cloud revolution, to the off-shore revolution, to the current ai-coding-assistant revolution. Then he finishes up with an explainer on why the “AI will replace developers” crowd have it all wrong:

Code is not an asset—it’s a liability. Every line must be maintained, debugged, secured, and eventually replaced. The real asset is the business capability that code enables.

If AI makes writing code faster and cheaper, it’s really making it easier to create liability. When you can generate liability at unprecedented speed, the ability to manage and minimize that liability strategically becomes exponentially more valuable.

The skill that survives and thrives isn’t writing code… It’s architecting systems. And that’s the one thing AI can’t do.

Jerod Santo: The AI 2027 scenario

The AI Futures Project is a new non-profit organization led by Daniel Kokotajlo, a former researcher in OpenAI’s governance division, with a goal of forecasting the future of AI. In their first publication, “AI 2027”, the group predicts that the impact of superhuman AI over the next decade will be enormous, exceeding that of the Industrial Revolution.

They lay this potential future out in a scenario that:

… represents our best guess about what that might look like. It’s informed by trend extrapolations, wargames, expert feedback, experience at OpenAI, and previous forecasting successes.

What’s cool about this work, unlike so many you can find online, is its incredible detail. The story goes quarter-by-quarter through the next 2.5 years, weaving one potential narrative of how things might proceed. There are two endings, though: a “slowdown” and a “race” ending. The overall goal is not prescriptive, but predictive accuracy. And they’re so serious about it, that they’re offering cash incentives for alternative scenarios that prove out to be more accurate.

Beyang Liu: I mean, I it, too. I think if you go back to 2013, I think there were a lot of – a lot of people read sort of negative connotations into it… Like, “What are you trying to do? Turn every development org into a feature factory?” Because that’s what I think when I think of the Industrial Revolution.

Jerod Santo: Write blogs so LLMs have something to read

Why create content in the age of AI slop? Nikola Ðuza has a dystopian answer to that question:

LLMs are getting better and better, but they all need some kind of input to be trained on.

And that’s where WE (I hate to say it, the “content creators”) come in. The world needs human touch, for now. The AI needs some human touch. The LLMs need to be trained on the good stuff. That’s why we need to keep writing, recording, and creating.

This is giving me strong Matrix vibes. (SPOILER ALERT) You know, where the machines’ only reason to keep us alive is to harvest us for energy, so they create a digital world to keep our minds busy while they feed on our bodies. But maybe that’s just me… I do agree with Nikola on this point:

In the sea of generated content, the “custom” “hand-made” “locally produced” content will always stand out. It’s the human touch that will make the difference. Furthermore, blogs and writing will highly unlikely die. To write is a way of getting your thoughts calm, organized and composed. Certain humans will always need to write, no matter how easy it is to generate content with AI. In that sense, blogs with human touch will always have a place.

Always is a strong word, but I do believe there will be a market for “human-crafted” content for the foreseeable future, just like there’s still a market for “hand-crafted” goods hundreds of years after the industrial revolution.

Salvatore Sanfilippo: Just that – I see that, in some way, even if we say that our field in some way in these 15 years changed. However, imagine something as big as AI in the hands of other industries, for example pharmaceutical industry or finance. They will never – it will never be like that that we have most of these tools for free. We have frontier models under the open source, a lot of other models and all the informations available. So I believe that still for IT there is hope, because we are in the environment that created the most incredible revolution is creating after like the industrial revolution, and still there is this idea of sharing, of making the tools accessible.

Also, if you think at the price of Open AI and Anthropic right now, $20 per month is very low. It’s very democratizing this tool. So in some way, I continue to be optimistic. I am more about the [unintelligible 01:26:40.27] side, which I believe that in the short run the problem that we will face with AI - as a society, we will be able to overcome them, but I will focus on the existential threat… Because the agentic nature of these models, as they become stronger and stronger, even if super-intelligence can be achieved without consciousness, still an agent that breaks all the computers around… This is the worrying part for me, that there is the – we will survive social change, but we may not survive some extreme event due to AI.

Jerod Santo: I think that pendulum swing that you’re describing finds its way into all kinds of stuff that we do. And it’s like “New technology, or new shiny thing - let’s use it everywhere. Let’s throw out the old, and let’s only do the new. It’s going to fix everything.” And then we go that route a little while, and start to see the cracks or the misapplications where it’s like “Actually, it’s not so great in this particular…” And we learn, and we realize that broad brush could be more precise. But then sometimes we pendulum swing back the other way, and throw out the new for the old… But eventually, I think we hone in on what is practical and useful, and not just in push buttons… But in your book, you talk about kind of the history even of the button, and knobs, and switches… Can you give us, you know, not the whole history, but maybe a primer for all of human history for us pushing buttons? I assume it started somewhere around the Industrial Revolution, or I don’t know, when we had things that could be on the other side of a button.

Autumn Nash: Think about it, if you have no home and no cell phone, how do you get mail? How do you get multi-factor authentication? You know what I mean? Like, even to fill out a lot of forms… We’re trying to partner with the Recovery Cafe, because to fill out a lot of forms that do help people who are unhoused, and who need it - cool. Those things exist. But they don’t even have access to them. They don’t know how to use it, and they need like so much more than they have access to actually utilize the things that are for them.

So I think the first time with tech – like, when we talked to people for the retro shows, it was something cool that they kind of stumbled on. They enjoyed it, and then it became this thing. When we talk even with Pete, about like getting into early internet and all these things, and it was a very like niche of things… But now there’s no niche. Like, it is everything. Tech has encompassed everything. It’s like the second revolution of tech. And at first it was like sad, and now it’s kind of like hopeful, that maybe this means we get to go somewhere cool after all this…?

Adam Stacoviak: That’s right. And they were an early sponsor of ours, and we were early days with them back in the droplet days, when it was a brand new noun in our lexicon, essentially… And they were driving down to like $5. These very fast, blazing fast, on SSD… This whole new revolution. SSD was newer than, a lot newer, and to get to that price point on something that fast… So the state of the market almost makes the user want what you’ve just described. My insignificant projects to be completely free, and my significant projects to be mostly free.

Matt Colyer: I think what’s most exciting about this – I mean, when I grew up, I remember when we first got the internet. And it was like the first ISP out there, and it was like a dial up modem, and there was like a BBS… It was just – that was like the first wave of technology to me, and that’s how I got into this field. And then I feel like the mobile revolution caught me by surprise. I think at the moment – I knew when the first iPhone came out, I was like “I need to have one of those.” But what I didn’t expect was how much the world would change after that. And it feels like this time around – I wasn’t a big believer in Web3. I was like “What is this Web3 business?” But I feel like this is, again, that same kind of shift… So I’m just going to ignore Web 3; I think this is the real Web 3, it’s AI.

And so how does this play out, right? I think what’s different this time is that for the last, I don’t know, 70 years that we’ve had computers, we as humans have had to conform to how computers work. At first we wrote literally bits. Then we wrote assembly code. Then we’re like “Well, maybe we should have languages.” And it’s like “Okay, so we’re slowly crawling there.” And then the next revolution is like “Oh, we should have point and click”, and so [unintelligible 00:46:30.01]

So now we have a world where everybody spends eight hours a day clicking on little colored boxes, and then typing into other colored boxes, characters on a keyboard. And I think what’s fascinating to me is that we’ve become – we’ve shaped who we are to conform to how computers work today… But I think this point in time – I’m sorry, I’m giving you 10 years out from now.

Mike Lewis: [00:58:05.17] Well, I will stand by what I said initially, that if we never got another innovation, these tools are good enough as they are to justify going down on the timeline of humanity as a massive technological revolution. And the tools that I wish for, that are already sort of – they’re almost done baking. The tools that I wish for are really low-latency, in and out, audio, video, image agents. I’ve built so many tools that interact with people. One class is called like Mind Mappers, where we’re essentially trying to convert tacit knowledge to documented process. So you’ve got this single point of failure person in your organization… How do you map out their subject matter expertise?

So what would be awesome is if there was a low-latency conversant model. So like what OpenAI kind of has, but it’s walled off right now. An API where I could give a person an option between an email chain, text messages, a phone call… And right now I can’t do the phone call. And that bothers me.

And then I think too, just as people start to see the value of the tools, that maybe they’re a little less weirded out and fearful of interacting with them. And so, of course, it makes sense that people feel nervous that they’re going to lose their job. The reality is they are going to lose their job as it exists today if there’s knowledge work. They should, because that’s like continuing to want to farm land with an ox, and whatever the thing behind – a plow, I guess. That’s wishing for keeping that, when the tractor dealership just showed up and you’re like “No–” We’re not going to want to go back to the days of like trudging through tedious knowledge work that could be automated.

But it is going to be a process that requires empathy on the part of us, the AI engineers, the architects, the developers, the leaders, and to appreciate, like, this is a scary thing. Another thing I wish for is just to get past this adolescent phase, to where the tools feel less scary. And then I’ve got to go back to - I am very excited to see what it means for learners, who maybe their teachers don’t have time to go on every curiosity journey with them, or explain every little lesson… That’s really big. I could give you such a long list… So I’ll just stop there. I am very excited for how this technology develops, but I will say, the good guys have to have a hand on the wheel here, Daniel. And if you leave it up to BigCo, we won’t have the stuff that really makes life better for a wide range of people.

Jerod Santo: There you go. Well, let’s dive into our topic for today. This is a follow-up to the last time we had you on, which - it has been a few years. But we talked modern Unix tools… And that was a lot of fun, going through a lot of the common Unix tools, like ls, and cd, and cat and etc, etc, etc. And talking about modern alternatives to those, which have bells and whistles. So that was that episode; people can go back in the feed, episode 451. We will link it up. “Modern Unix tools, with Nick Janetakis.”

Today we’re going to talk terminal again, or focus in on TUIs. I feel like TUIs are having a moment, a – I won’t call it a revolution… A renaissance. There you go. A renaissance. Probably in light of a lot of the tooling that’s now available. We’ve had shows with people like Charm, the Charm Bracelet folks, who are providing Go-based tooling for doing all kinds of text-based and terminal things… And then there’s also - is it Will McGugan, I believe, and the Textualize folks, over there providing similar tools for the Python community… And a lot of the stuff that we’ll talk about today is built with Textualize. And then there’s also something going on in the Rust world, with a toolkit called Ratatui.

[00:22:44.25]

And the truth is, I have no talent at all. But this rat, he’s the one behind these recipes. He’s the cook.

Autumn Nash: [00:23:45.16] Yeah, that was open source too right there. I got coffee, I sat [unintelligible 00:23:47.08] Well, because it’s gonna set a precedent, right? From now on, every time somebody goes and licenses something that was open source… These things didn’t happen before, and all of a sudden, the more this happens, this is going to become the playbook of either “How do you fight back? How do you keep something open source [unintelligible 00:24:04.24] forked?” So it’s almost like being – we’re watching the new revolution of open source and how it’s going to survive in this new environment.

Kris Brandow: Maybe it’s an unpopular opinion, but I think anybody that can use the computer can learn how to do these. I think we introduced them too late. If we had started with this passkey revolution 20 years ago, like we could have, I think everybody would be used to it… Like, “Oh, why are you still using a password? That’s so weird; you should be using this other thing that we have.” But we just didn’t build up the infrastructure.

I think the thing that’s pushed passkeys forward is the biometrics in our phone, and on our computers, where now you can actually confirm that the person sitting at the computer is who you think it is, within a degree of certainty. But yeah, I think that a lot of this is – it’s hard to understand, because we’re bad at explaining it, not because people fundamentally can’t understand it. I’m sure your mom doesn’t remember every single person’s phone number. I bet she’s figured out how to use the Contacts app on her phone and put all the phone numbers in there, and just be like “Oh, I’m gonna call this person. Let me click on their name in the Contacts app.”

Mark Russinovich: And to put that into context at the height of PC revolution, when hardware was coming out and advancing very quickly, the capabilities, the number of gigahertz or operations per second for PC or CPUs grew by 100x in 10 years. So this is advancing at 10x the rate of what CPUs were advancing.

Chris Benson: Well, this is kind of a thread we’ve talked about off and on on different episodes over time, and… You know, AI models are hosted in software, and they’re going to always be wrapped in that. And software expands from the cloud all the way out into every device that we’re already using, and so it’s only natural that as AI becomes more accessible, and cheaper to deploy, that you’re going to start having models that are kind of ramping up existing software out there. And there’s of course, we’re going through all the hype that’s associated with that, but the way I see it is it’s simply just a natural evolution of where software development would go.

[00:05:45.13] And to your point a moment ago about the hardware side - we don’t talk about that a lot. We’re very software-focused in general… But the hardware side is really going through a revolution. I know that in your business you have a partnership with Intel, and are seeing that in that capacity, and I certainly see that in my day job… And so there are so many more hardware capabilities coming out to support these functions, many of which are targeting low-power, disconnected environments… And so this is a timely topic. And if you look back for a precedent before that, we’ve always seen over the years in software and cloud development kind of a shifting back and forth between local capability, as suddenly you get a new generation of hardware and things will go a little bit more local, and you’ll have your own equipment, and then things will move back into the cloud… And that natural give and take is part of the flow, and I think right now we’ve been so cloud-focused the last few years, because that was really the only available option, that now we’re seeing a lot of new capability rolling out on both hardware and software models that are going to enable edge functionality to really explode over the next few years.

Scott Guthrie: [00:12:25.12] I think one of the things that’s interesting about AI - and I think this is similar to kind of, again, going back to whether it’s the internet, or whether it’s kind of the iPhone revolution, is it has entered the zeitgeist of people. And so even though there’s lots of developers that haven’t built an AI application, I think it’d be hard to find a developer on the planet who hasn’t tried ChatGPT, who hasn’t tried some generative AI app. And increasingly, we’re seeing just millions of developers that are using GitHub Copilot as part of their day to day activities.

So I think generative AI is something that we’re all – most developers now are using in some way, shape, or form, or have certainly at least tried. And to your point, I think where we are is still in the early innings; it’s sort of 18 months since ChatGPT was unveiled… And so people are still trying to figure out “Okay, how do I actually pivot to be an AI-first app? How do I incorporate AI into my app?” And part of what we try to do with Build is not just talk about the science and the art of the possible, but actually from a practitioner perspective “Here’s how you do it.”

And the first demo, or one of the first demos in my keynote was showing how you build a customer-facing chat experience into an existing website, or it could be a mobile app. And things like our Copilot Studio are really designed so that you can actually safely and securely build your first app, integrate it into an existing app, and not have to be a data scientist, not have to understand groundedness.

You know, Copilot Studio isn’t going to be the tool for every scenario, but if you’re looking to kind of build 80% of solutions, or 90% of solutions, it’s probably good enough out of the box. And then what’s nice is it has that no clips extensibility, so that you can call any API. And that was sort of the second demo that we did, that was live in the Azure AI section, where we used Python, we used Prompty, we basically grounded a model, it could use vector databases etc. And I’d probably encourage people to start and build a simple app using Copilot studio, and then as a developer keep going. Hopefully, what we’re showing with each of the events that we’re doing, each conference, it is becoming easier and easier. And ultimately, I do think in the next two years, I think pretty much every organization on the planet is probably going to have a custom AI app that they built.

Don Beyer: Number one safety issue is “How about job elimination?” We know that it’s going to replace many, many jobs. But the exciting thing is what they call ambient clinical documentation… Doctors and nurses say 25%, 30%, 50% of their time is filling out data on the clinical visit they’ve just finished, or in the middle of it. Now, there’s software or hardware that listens to the conversation between Chris and his doctor, and writes it all down. By the time the patient leaves, the doctor can read it and check it, and “Yup, that’s what we talked about.” Saving an immense amount of time.

But then – so job elimination… And what do we do? We know that that’s happened in every revolution - agricultural, industrial, information… But we also know that it will probably happen much more quickly now than it ever has before. So much less time to react and for people to adapt to that change.

A second level is all the misinformation, whether coming out of the large language models, either unintentional or the intentional stuff. How do we protect against that? Then there’s the whole notion - and some people take this seriously, other people say “Nah…” That with desktop ability to generate DNA, and synthesize DNA, and the ability to look up “What’s a smallpox vaccine? What’s the DNA of that?” Or go from COVID-19, let’s make COVID-27. The whole notion of bioweapons, or others things that can be used based on the information that comes from language models and AI. And then all the way to the existential threat. It’s interesting that so many of the computer scientists I talked to just really say “No, no, no, we’re nowhere near Artificial General Intelligence. And even if we were at AGI, that’s not going to be conscious, it’s not going to have will. It’s not going to plan its own things.” I tend to be on the humbler side, which is “We don’t know where consciousness comes from. And we don’t know when and where it’s an emergent property from what.” If we’re building machines that can think some things thousands or millions of times faster than we can, why are we so sure that there won’t be an emergent consciousness coming from this? And in my conversations with Elizabeth Kelly at the NIST Safety Institute, my plea is that there’s at least some subset of that group that always is keeping the existential threat at the top of mind.

Autumn Nash: Sometimes I think that AI will make us more efficient, but it may also complicate things more than it’s worth… You know what I mean? I don’t know, it’s gonna be really interesting. I feel like we were talking the other day about how far infrastructure has come, and how much it’s changed in the last 5 to 10 years… And now we’re about to go through a whole other revolution of open source, and AI… So it’s interesting, like, where will we end up later.

Jerod Santo: What about the job market? Do you think it’s fundamentally affected by tools like Devin as they progress over the next three to five years? Because we’re not talking about humans out of the loop. I think we’re all in agreement here that that’s not feasible, or smart, at least, in today’s technology plateau of LLMs… But less humans in the loop. That seems like it’s very feasible if these tools continue to iterate and even just not have revolution, but evolutionary advancements from here.