💻 "We don't need juniors anymore since we have AI" —— What awaits at the end of that road? Ruby creator Yukihiro Matsumoto's warning isn't mere nostalgia. It's a visceral concern about a future where technical knowledge transfer has ceased, leaving behind nothing but "scorched earth."
What Is the "Juniors Are Unnecessary" Argument?
With generative AI evolving at a breathtaking pace, code generation tools have become remarkably accurate. This has fueled an active debate in the IT industry: "We don't need junior engineers anymore."
Videos of AI creating Tetris in seconds go viral on social media, leading some to declare, "Engineers are obsolete." According to Stack Overflow's survey, AI coding tool usage reached 84% among developers by 2025.
This trend is clearly affecting young engineers' employment. A Stanford Digital Economy Study found that employment for software developers aged 22-25 declined nearly 20% from its late 2022 peak by July 2025.
Yukihiro Matsumoto's Deep Concern
Against this backdrop, Yukihiro Matsumoto—known as "Matz," the creator of Ruby who has been at the forefront of programming for over 30 years—has voiced clear concerns.
What awaits after dismissing juniors as "unnecessary" is a "scorched earth" where technical knowledge transfer has ceased, Matz warns.
According to Matz, AI-powered "vibe coding" is "like a gacha game"—you never know what you'll get. While AI can reproduce existing simple programs with nearly 100% success, trying to create something new fails "99% of the time" in his experience.
In other words, AI is merely a "capability amplifier" for people who can already program. Without humans who can guide AI, fix its mistakes, and design new things, work simply cannot get done.
Without Juniors, There Will Be No Future Seniors
The core of Matz's concern is the "discontinuity" in talent development.
Junior engineers grow by handling small tasks and gaining practical experience. Through this process, they understand system architecture and learn why things were designed certain ways.
But if AI replaces junior-level work, the opportunities for young engineers to grow will disappear entirely. And if juniors don't develop, there will be no senior engineers in the future.
AWS CEO Matt Garman has also pointed out that replacing junior developers with AI is "one of the dumbest things" companies can do. Young people not only grow into future seniors but already know how to work effectively with AI coding tools.
Japan's Unique IT Talent Challenges
This problem is particularly acute for Japan.
According to the Ministry of Economy, Trade and Industry's estimates, Japan could face a shortage of up to 790,000 IT professionals by 2030. Additionally, Japanese IT workers tend to have lower skill levels in international comparisons—while engineers in the US and India average skill levels above 4, Japan's average remains in the lower 3s.
Furthermore, Japan's IT industry's unique multi-layered subcontracting structure compounds the problem. Companies handling downstream processes (programming, testing) are doing exactly the work most susceptible to AI replacement.
The "2025 Cliff" problem has also materialized. While companies struggle with legacy system maintenance and fall behind in adopting new technologies, the engineers capable of handling these challenges are retiring.
Four Skills Needed in the AI Era
Matz proposes four skills that young engineers should develop to survive in the AI era:
1. Ability to Guide AI The power to detect and correct AI mistakes. AI might output 4-space indentation even when you specify 2 spaces. Humans who can catch such errors and provide appropriate guidance are essential.
2. Programming Fundamentals AI excels at reproducing existing things but struggles with creating new ones. Without design and programming skills, you cannot effectively utilize AI in the first place.
3. Problem Discovery Ability Deciding what to build is a human job. The ability to determine "what" to ask AI to do becomes crucial.
4. Communication Skills Software development is teamwork. In an era of human-AI collaboration, the value of human-to-human communication increases.
Sustainability of OSS Communities
Matz also addressed the sustainability of Open Source Software (OSS) communities.
OSS that underpins modern software development—Ruby, Linux, MySQL—is maintained through community contributions. However, as projects grow larger, maintenance costs also increase.
"Unless large companies say 'We benefit from Ruby, so let's give back,' long-term maintenance becomes difficult," Matz notes.
Not just funding, but human contributions matter too. If junior engineer development stagnates, fewer people will be able to contribute to OSS, threatening the entire ecosystem's sustainability.
A Global Discussion
This issue is being debated worldwide, not just in Japan.
David Heinemeier Hansson, creator of Ruby on Rails, argues that current AI tools "still can't match most junior programmers." AI sometimes produces correct code, but it's like a "flickering light bulb"—unreliable.
Meanwhile, Anthropic CEO Dario Amodei suggests that within 1-5 years, half of entry-level white-collar jobs could disappear. CEOs of Google and Microsoft have also stated that AI now handles 20-30% of coding work.
Amid this mix of optimism and pessimism, one thing is certain: change is happening.
How Should Japan Respond?
How should Japan address these challenges?
First, don't stop developing young engineers. Reducing junior hiring for short-term cost savings leads to long-term technical hollowing out.
Second, view AI not as a "threat" but as a "tool," exploring new development styles where humans and AI collaborate. Let AI handle what it does well while humans focus on design and judgment.
Third, foster a culture of continuous learning. In an era of rapid technological evolution, skills learned once quickly become obsolete. Both organizations and individuals must maintain a learning mindset.
What About Your Country?
In Japan, AI evolution and IT talent shortages are progressing simultaneously, highlighting the challenge of developing young engineers.
What's the employment situation for junior developers in your country? How widely have AI tools penetrated your development workplace? What discussions are happening about training young engineers?
Please share your country's situation and your thoughts!
References
- https://atmarkit.itmedia.co.jp/ait/articles/2601/26/news002.html
- https://atmarkit.itmedia.co.jp/ait/articles/2511/13/news001.html
- https://gihyo.jp/article/2025/05/rubykaigi-2025-keynote-report-day3
- https://stackoverflow.blog/2025/12/26/ai-vs-gen-z/
- https://www.meti.go.jp/shingikai/economy/daiyoji_sangyo_skill/pdf/001_s03_00.pdf
- https://www.finalroundai.com/blog/ai-can-not-replace-junior-programmers
Reactions in Japan
I think Matz is right. Companies that stopped hiring juniors will end up with 'nobody can read the legacy system' in 5 years. That's exactly what's happening at my company.
I use AI for code generation, but I still do the detailed fixes myself. Saying 99% fail is an exaggeration, but about half need manual corrections I guess.
I'm a new grad wanting to be an engineer, but job openings have dropped so much... Seniors who job-hunted 3 years ago said it was a 'seller's market.' Times have changed too fast.
People saying 'we don't need juniors because we have AI'—who's going to review AI's code? It just increases senior workload.
Won't AI actually accelerate the multi-layered subcontracting in SIers? I have a feeling lower-tier work will disappear and cheaper companies will go under first.
Matz's talk about 'reverse alpha syndrome' at RubyKaigi 2025 was also impressive. We shouldn't become slaves to AI.
Honestly, my productivity has gone up since I started using Copilot. But I think I can 'use it effectively' because I've written code myself in the past.
I thought the Ministry's estimate of 790,000 shortage could be solved with AI, but it's not that simple. Never thought about the discontinuity in talent development.
My company halved new grad hiring saying 'efficiency through AI,' but it just increased burden on trainers. With so few young people, all the grunt work falls on mid-career staff.
I'm attending a programming bootcamp, and hearing 'AI will take your jobs' makes me anxious. But Matz's words reassured me a bit. Fundamentals matter.
The 'scorched earth' expression is strong but accurate. Technical knowledge transfer is often undervalued because it's invisible, but once broken, it's really hard to recover.
Thinking the opposite way, does becoming an engineer now mean you'll have scarcity value? If everyone relies on AI, people who can actually write code might become precious.
Did the '2025 Cliff' actually happen? There was talk about solving legacy system problems with AI, but it's not that easy, right?
The talk about OSS contributions is important too. Too many companies use it for free. If you're making money using Ruby, please give back to the community.
I heard some overseas companies like Salesforce have stopped hiring juniors. Will Japan go the same way...?
AI coding tools are convenient, but they're 'tools for those who understand.' Beginners using them right away can't tell if the output is good or bad.
Matz's 'four skills' basically mean to polish what only humans can do, right? Not competing with AI but having a collaborative perspective is important.
Congratulations on Ruby's 30th anniversary. Matz's words always get to the essence and are educational.
I'm a senior engineer in Silicon Valley, and I'm experiencing exactly this problem. Junior hiring has decreased, and so have mentoring opportunities. Our team's energy is declining.
I work at a UK startup. AI tools are definitely convenient, but code review burden has increased. In the end, human eyes are still necessary.
In Germany, our vocational training system is solid, so the discontinuity in junior development hasn't happened yet. But we should watch this discussion closely.
In Indian IT companies, junior hiring is still active. However, demand for AI-skilled talent is growing, and traditional programming alone is becoming insufficient.
In France, labor laws are strict, so we can't easily cut juniors. That might actually be working in a good direction. Continuity in talent development is maintained.
In Korea, job difficulties for junior engineers are serious too. Large companies only want ready-to-work talent, and SMEs can't afford training. A vicious cycle continues.
In Latin America, remote work has us competing with US companies. But in the end, we might just be used as cheap labor.
I work at an offshore development company in Poland. Work is coming to us because Western European companies stopped hiring juniors. Mixed feelings about this.
Australia's tech industry is small, so discontinuity in talent development could be fatal. Matz's warning should be taken seriously.
In China, the government is investing heavily in AI talent development. No time for 'juniors unnecessary' talk. Talent is everything to win in international competition.
Ireland is an EU tech hub, but talent mobility is too high—people leave for other companies after training. So companies hesitate to invest in development.
In Scandinavia, 'AI for Good' discussions are active. There's awareness that protecting tech workers' employment is part of social responsibility.
I'm a CS major. Feeling anxious about employment after graduation. We use AI tools in class, but I also keenly feel the importance of learning fundamentals thoroughly.
The Middle East tech industry is growing and still actively hiring junior talent. Western discussions are informative, but our situation is different.
In Vietnam, offshore development is thriving and junior training is active. But due to AI impact, simple coding projects are decreasing.
I agree with DHH. AI is like a 'flickering light bulb.' Unreliable. Human programmers are still very much needed.