How does the company decide whether an engineer truly understands their code?

It asks extremely detailed questions about implementation, such as what useState does or why clicking again hides the div. Engineers must answer correctly before a PR merge is approved.

Why does the company say relying on AI raises costs?

When code misbehaves, engineers repeatedly ask AI to fix it until it works, and rising usage frequency increases costs. Simple tasks that framework commands could finish in seconds also get routed through AI.

What happens when engineers cannot use AI, such as on offline servers?

The company says AI-dependent engineers struggle in offline on-premises environments, cloud infrastructure AI cannot directly operate, or sudden database and server operations. Those with system knowledge can respond immediately.

AI Coding AssistantsAI Business & IndustryZenn AI/ML2d agoOct 19, 2026, 22:00 JST

Firm bans AI for juniors, grills them on code

Send on LINE

3 Key Points

What happened
The company bans AI in client support and training, and approves PR merges only after engineers answer detailed questions about their own code. Most low-skill engineers fail and must re-investigate.
Why it matters
The firm says AI-generated work breeds overconfidence: engineers cannot explain logic, locate code, or review others, and repeated AI fixes raise costs, so client complaints follow.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

The company frames its ban not as a rejection of AI but as a return to how system development is supposed to work. It argues that banning AI pushes engineers back to reading documentation and tutorials, and that fixing review feedback without AI provides success experiences that build real skill. A recurring concern is evaluation: when AI masks an engineer's true ability, managers cannot measure practical skill, and engineers themselves mistakenly believe they can develop systems on their own. The firm's proposed remedy is simple to state but demanding to apply: make engineers explain their work in detail, which it says surfaces gaps that normal PR reviews miss and raises the technical dialogue skills needed for client support.

FAQ

How does the company decide whether an engineer truly understands their code?: It asks extremely detailed questions about implementation, such as what useState does or why clicking again hides the div. Engineers must answer correctly before a PR merge is approved.
Why does the company say relying on AI raises costs?: When code misbehaves, engineers repeatedly ask AI to fix it until it works, and rising usage frequency increases costs. Simple tasks that framework commands could finish in seconds also get routed through AI.
What happens when engineers cannot use AI, such as on offline servers?: The company says AI-dependent engineers struggle in offline on-premises environments, cloud infrastructure AI cannot directly operate, or sudden database and server operations. Those with system knowledge can respond immediately.

Zenn AI/MLRead Original Article

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.