Richard Russell ⤴'s Linkedin Analytics

X : What's the future of software engineering? Are they all getting replaced by AI? Is Eric Schmidt right? Me : I think he might be right ... X : Phew. We don't need software engineers. Me : Execs like you are getting rid of them. You'll eventually hire more useful but expensive computational product realisers, machine collaboration architects, digital workflow synthesisers, technology decision auditors, AI enablement strategists, Human-AI systems integrators, codebase curators, technical debt reconciliation officers, platform realisation facilitators - like Alice. X : Alice? She used to work in my IT Group as a software engineer. Me : Alice is now a highly skilled Human-AI system integrator capable of reading and writing code, designing systems and working with many AI agents. Highly sought after and far more expensive. X : But reading and writing code, designing systems and working with many AI agents is what Alice did before we fired her. Me : Do you want to explain to your board why you fired most of your software engineers that you're now trying to hire back? You did the same with your sysadmins in the early cloud days and then had to hire them back as DevOp engineers or have you forgotten? Remember NoCode? Or trying to build your own private cloud ... OMG, I almost forget, VCE. You fell for them all. P.T. Barnum must have thought it was Christmas everyday with you. It's your choice. X : AI was supposed to create 10x software engineers, reduce our costs. Me : It did. But you thought this meant you could get rid of 90% of your engineers. Your competitors realised this meant they could do 10x more work. There's a long tail of unmet IT needs. This is why your CEO is constantly on your back, your competitors look more innovative because ... they are, they have the capability. You can't keep up, you fired the skills you need. X : But I can't hire back software engineers, I told the board we didn't need them. Me : I know. Hence, Alice the Human-AI system integrator. X : How much? Me : I'm thinking, 5x? How much is your career worth? X : 5x? That's robbery. Me : There's an entire market of CIOs who have made the same mistake you did combined with a massive growth of startups building new things on AI and a whole bunch of canny players who have taken advantage and were hiring from you as you fired. The market is blazing hot right now. You should have retrained people. X : Wasn't very bright was it? Me : Listening to some Aston Martin driving ChatPPT consultant who charged you a bathtub of cash for some hogwash on vibing the future was about as smart as getting AI advice from some VC invested in AI companies. -- Addendum For those who missed the arguments first time around that cloud would get rid of sys admins, I made the same warnings back in 2010 - https://lnkd.in/en3bzjpT ... and yes, a typical large company today has several hundred thousand virtual servers.

We just published a short, but I believe important article, talking about the rise of the product creator, and some of the implications for product managers, product teams, and more generally, innovation:

When you understand this? You understand everything. Just 10 words. But a mountain of meaning. The quote in this doc cuts so hard against how we think business works that it’s hard to make it sink in. Instead, it just pings off the surface of our stubborn brains. But if you unpack the mechanics of it, it becomes easier to grasp. And then you start to get it. Give this a read, and hope you enjoy. ___ P.S. For more deep strategic wisdom, subscribe to the number 1 source on the internet: My newsletter, The Hidden Path. You’ll also get 2x essential gifts: - The Ultimate Strategy Doc Template - The No BS Strategy Toolkit Here's the link: https://lnkd.in/eUHhESi

Waterfall might be having a quiet comeback, thanks to AI. As LLMs reshape how we build software, I’ve noticed something surprising: MVPs succeed more when I plan like it’s 2005. Why? • LLMs need clarity. The more ambiguous my prompt, the more off-target the output, especially when GenAI agents are generating full-stack flows. • Speed flipped the bottleneck. With tools like Copilot, Cursor, and GPT-4o, the limiting factor isn’t coding speed. it’s clarity of thought. • Execution is near-instant. I recently built AI-powered apps in days. But the magic only happened when the architecture, flows, and logic were crystal clear before the first line of code. What I’ve seen: • 48–72 hour MVPs are now realistic — but only if you front-load your thinking. • Spec writing is faster than ever. With GenAI, outlining flows, API contracts, and feature logic takes hours, not days. • “Prompt-first” = Waterfall reboot. We’re just not calling it that. • Iteration still matters. But it pays off most after you’ve shipped a coherent V1. But: • Waterfall still struggles in the wild — especially when feedback loops drive product evolution. Remember, MVP doesn’t imply happy users! • Agile still shines in continuous discovery and fast learning cycles. • LLMs can write code, but they can’t (yet) reason about the user, edge cases, or tradeoffs. My takeaway: Maybe the real shift isn’t Agile vs. Waterfall anymore. It’s Clarity vs. Chaos — and clarity wins when AI is on your team. Next questions to answer: • What’s the right size for an AI-powered MVP? • How do we reach clarity faster, without bottlenecking creativity? • What does a full, AI-native SDLC look like?

Seen the graph on the left posted lots of times with the narration "It's the end of software developer jobs" before? So have I. Those posting it speculate that AI is causing a depressing decline in developer postings. So let's look at the companies that are using AI the most: Big Tech (like Microsoft, Amazon, Google, Meta, Apple) "top tech" companies (eg Pinterest, Coinbase, GitLab etc) and VC-funded scaleups and startups (OpenAI, Anthropic, Grafana etc). If the narrarive of AI tools spreading reduce jobs were true: these are the companies adopting these tools the most! So job postings would be down in this group: as they have integrated AI into their developer workflow and products more than any other company has! BUT: they are not! This second group is hiring more than they have in 2 years's time - since AI has gone mainstream (ChatGPT was released in Nov 2022, and it crossed 100M weekly users in Nov 2023. Hiring sped up at these companies right around this time!) So something doesn't add up. Worth asking: what is this? And why are the companies using AI the most the ones increasing tech hiring the most (including software engineer and AI engineer hiring - these two are not that different btw!) Data source for first chart: FRED https://lnkd.in/epht9MZS Data source for second chart: TrueUp at https://lnkd.in/e9j8Xcra. Browse the 236,000+ jobs at these companies listed at https://trueup.io/jobs And see my analysis on what really caused this decline from 2022: the end of zero interest rates in The Pragmatic Engineer at https://lnkd.in/eqTg2Rus (free article) As closing: it's clear that the current tech jobs market is much tougher than the years before. For new grads, it might even be the toughest job market ins 20 years (since the Dotcom Bust). And the massive decline in hiring since 2022, coupled with tech sector layoffs, and (now) stricter performance management all make getting a new job a lot more cutthroat. But the job market started to get tougher even before AI went mainstream: and while it probably contributes some, it's unclear to me if AI is the one and main reason for all of this. (I plan to look more at this strange market. If you have tips for data sources or angles, feel free to comment or send a direct message!)