Tech Industry Disruption: Lessons From Kent Beck and Martin Fowler
Published on 07.04.2026
Tech Industry Disruption: Lessons From Kent Beck and Martin Fowler
TLDR: At the Pragmatic Summit, Agile pioneers Kent Beck and Martin Fowler discussed how AI's arrival differs from past technological revolutions (microprocessors, object-oriented languages, the internet). They warned that company incentives will remain misaligned, "snake oil" vendors are already appearing, and a "middle" of developers focused on money rather than craft will face career uncertainty.
How AI Differs From Past Tech Shifts
Fowler observed that nothing has hit with the magnitude of AI. Object-oriented languages scared people, but there was room for debate. The internet had a huge impact, but some people genuinely didn't think it was important. With AI, there is no argument about importance. Everyone knows it matters.
Beck drew a parallel to the Intel 4004 in 1971. Before microprocessors, computers were immobile, expensive machines. The arrival of the chip suddenly expanded the possibilities of computing. If you could figure out how to write software and design hardware around it, you could do things you hadn't imagined.
Beck predicts AI will expand software engineering similarly. He's writing ridiculously ambitious projects now, ones he wouldn't have attempted before. Libraries in Rust that rival commercial packages. A persistent Smalltalk. The expansion of imagination is part of what's happening.
But there's a critical difference in adoption speed. Past revolutions took time. AI is hitting fast and everywhere simultaneously. This speed means less time for industries to adapt, fewer places for the "middle" of professionals to find refuge.
Agile Adoption Teaches Us About AI Adoption
When Agile emerged, Fowler and Beck were part of it, so they could push the agenda effectively. Today, they're observing from the sidelines as companies adopt AI in ways that repeat historical patterns.
With Agile, company incentives were often misaligned with actually achieving faster, cheaper, better software. Kent puts it bluntly: "Inside some companies, the incentives are misaligned with actually achieving that. And so as geeks trying to achieve these improvements and saying: 'it's 40% better, 12% cheaper and less fattening,' people will punish you if that doesn't align with their incentives inside organizations."
Fowler adds: "I suspect there will still be some similarities with Agile. The core notions behind Agile and extreme programming are solid and good, but a huge snake-oil industry appeared around it – the 'Agile industrial complex', as I refer to it. This is also happening with AI right now, and it's often hard to see the difference between snake oil and the real stuff."
The "middle" of developers faced similar disruption during the Dotcom boom. "People who got into programming as a way to make money" ended up in real estate or left the industry. That middle is much bigger now.
AI As Amplifier
Beck frames AI as an amplifier. If you're young and learning quickly, AI amplifies your learning. This is the golden age of the junior programmer. But the middle who got into programming for salary? Beck doesn't know where they'll go. And they're a much larger population than in 1995.
AI is not a "let's get rid of programmers" trend in intent, but it creates that effect. Automation always does. The real fear is justified, though Beck and Fowler suggest it's partly about programmers themselves. "We should think about why people periodically want to axe us," Beck says. "Some of that's about us as programmers, and some of it not. Still, we should think about why people periodically want to axe us."
The Re-Soloing Problem
One of the most interesting warnings came from Beck on what he calls "re-soloing" – the reduction of collaboration. Extreme Programming was built on creating a safe social environment for people who are naturally antisocial. People talked to each other for hours a day because the process was set up to be positive.
Now Beck observes programmers saying, "I've got six agents, so really I'm managing a team." That's not team management. It's using six tools simultaneously, which is fine, but it's very different from having a conversation with someone who sees things differently, who has different energy on that day.
"We used to have programmers in individual offices with doors, and you'd shut the door and slide the pizza underneath. That was easy to manage, but then along came this messy, social, complicated, chaotic process of software development, which just happened to produce really good results."
Martin counters optimistically: "Are we seeing two-pizza teams becoming one-pizza teams because agents don't eat pizza, or do we see two-pizza teams staying and becoming much more effective and capable? My bet is on more effective two-pizza teams."
Beck's experience of pairing with two humans plus one or more AIs has been very positive. "The fact the AI is slow is really nice. Every time models come out and are faster, I'm like, 'Oh, there's less time to talk.' When the AI goes away for three minutes, we can talk about our philosophy of naming, or how we express conditionals."
Key Takeaways
- AI's arrival is unprecedented in speed and magnitude of impact compared to microprocessors, OOP, the internet, and Agile
- Company incentives remain fundamentally misaligned with the stated benefits of new technologies, just as with Agile
- A significant "middle" of developers oriented toward salary rather than craft faces career uncertainty, similar to the Dotcom aftermath
- AI acts as an amplifier: it amplifies the learning of juniors and the effectiveness of experienced developers, but destabilizes the middle
Why Do I Care
Fowler and Beck have been through multiple technological revolutions and survived because they focused on craft rather than trends. They learned to balance skepticism with curiosity, and to ask hard questions about incentives and what's actually being solved.
The most important insight is that technological change doesn't fix broken incentives. It just creates different winners and losers. If your organization rewards speed-at-all-costs or short-term metrics, AI won't fix that. It'll amplify it. The "middle" finding refuge is real and worth thinking about seriously. If you're in that middle, the time to move is now, toward either mastery or management.
On the social side, their warning about "re-soloing" is worth heeding. The best work in software has always come from collaboration. AI doesn't eliminate that need. It just makes it easier to pretend you don't need it. Don't believe that.
Cycles of disruption in the tech industry
Stop Scrolling Job Boards: Let Claude Find the Right Job For You
TLDR: A tutorial on building an AI Job Hunter system using Claude. The workflow has two phases: analyzing your CV to find target roles and gaps, then having Claude search real job boards to find matching positions with tailored applications.
Know Yourself First
Most people apply to job titles they already know. But the best opportunities are often roles you haven't considered that fit your skills perfectly. Claude can do that analysis at scale.
Start by uploading your CV to Claude and asking for a brutally honest assessment. Have it identify your top three strengths, top three weaknesses, and calculate a first-impression score. Request an ATS (Applicant Tracking Systems) compatibility check, which most job hunters never think about. Ask for a realistic assessment of what seniority level you actually qualify for.
The key is asking Claude to be direct, not encouraging. You need the truth about how recruiters will see your CV in the 6 seconds they scan it.
Next, ask Claude to identify the top 5 job titles you should target, ranked by likelihood of getting hired. For each one, ask what percentage of the job requirements you likely meet, and expected salary ranges for your region. Request 3 "hidden gem" roles you probably haven't considered but would be a great fit. Ask which popular roles you should avoid because you'd get filtered out. Finally, ask which industries would value your specific skill combination the most.
Third, have Claude rewrite your CV sections. Get a new professional summary optimized for your top target role. Have it rewrite your three weakest experience bullets using the STAR format (Situation, Task, Action, Result). Ask for the 10-15 keywords you should add and exactly where to place them. Request a section reorder if it would improve your targeting. Ask what you'd cut if you had to make the CV one page.
Find the Job
Once you know your strengths and target roles, Claude can search real job boards. This is the phase where it gets practical. Have Claude connect to live job boards (the tutorial doesn't show the exact implementation, but the concept is real-time search against sites like LinkedIn or industry-specific boards).
Claude can filter by location and salary, matching results against your skills. It can rank opportunities by how likely you are to get hired. Most importantly, it can help you prepare winning applications. Tailored cover letters, interview prep for your top matches, addressing specific job description requirements.
The efficiency gain is dramatic. The old way: three hours scrolling job boards, applying to 20 random listings, hearing back from zero. The new way: 30 minutes with Claude. Targeted roles, matched listings, tailored applications.
Key Takeaways
- Uploading your CV to Claude for honest assessment reveals strengths, weaknesses, and ATS compatibility that most job hunters never identify
- Target roles should be identified by likelihood of getting hired, not by job title recognition
- Rewriting your CV using STAR format and strategic keyword placement dramatically increases screening pass rates
- Letting Claude search real job boards and filter by your profile is far more efficient than manual scrolling
Why Do I Care
This workflow reveals something important about how AI changes job hunting: it commoditizes the parts that are actually coachable. Resume writing, target role identification, even cover letters are now AI-handleable. The bottleneck isn't drafting a good application anymore. It's having the confidence and clarity to apply strategically.
If you're job hunting right now or will be soon, this is a concrete workflow you can execute today. The biggest value isn't the individual steps. It's the permission to be strategic. Most people spray and pray because they don't have a systematic way to think about their own skills. This gives you one.