AI 時代的 100 倍開發者:生產力該如何衡量?
100x Developers
**作者質疑 AI 時代「100 倍開發者」的說法,認為這種倍數可能只適用於特定任務或個人專案,難以代表企業團隊的整體生產力。
- - 倍數取決於衡量方式**:生產力可用程式碼行數、MR、程式碼審查、部署、功能交付或業務 KPI 衡量。
- 若把完成任務的影響當成生產力提升,結果也可能失真。
AI 伴隨問答
針對這篇報導有疑問?可直接點選快捷問題或向 AI 追問背景脈絡與技術細節:
Think about it. 10x of something is a lot.
- 10x a salary of $70.000/year (~avg salary in USA) is 700k€/year. Imagine the lifestyle change.
- Lifting 10x my maximum PR on dead-lift goes from 160kg to 1.6T.
- A means of transportation 10x faster is arriving in 9 minute what it took 1.5h
- Avg. vertical jump for an male adult is 45cm. 4.5m if you 10x it.
We referenced the mythical 10x developer casually in the dev world. Now that we have AI, we’re mentioning the existance of 100x developers!
It bothers me. Not because it might be true or false in particular instances, but because the number just scapes me. 100x. It’s a lot if talking about productivity.
Lets put this number in perspective. In 3 days, at 100x and you achieve the work of 300 working days. There are 260 working days in a year. That sounds unrealistic.
Steelmanning the “100x case”
Under some circumstances, I think it’s possible to have such a big multiplier:
Newbies. You know absolutely nothing about programming. Doing a full website, including frontend, backend, DB access, prod setup… that’s a lot of work for someone who’s knows nothing. You’ll be blazing fast using AI compared to the time it would take without.
Super small or specific task. Tasks so small that AI can do them in 10 minutes, were you would take almost 3 hours. Examples:
- AI troubleshooting a production problem involving many services, with access to source code, logs, traces, etc…
- Classifying and prioritising tickets/issues on a given criteria
- Finding the root-case of a flaky CI tests
Measuring task impact, instead of developer productivity. “In 10 minutes of vibe coding I saved 10 minutes a day to every engineer in the organisation”. You’re measuring the impact of the completed task. But if the task would have been done manually taking 1 hour, the impact would have been big as well, right?
10x/100x projects
So, 10x of what? 100x measured how?
What’s the metric being measured, here? It can be: Productivity. Lines of code. MRs. Reviews. Deployments. Impact (how you measure it?), features shipped, business KPIs’ growth, …
I bet that most developers claiming to be 10x because of AI are working on a personal projects or uni-personal companies.
I bet they are measuring lines of code, MRs or deployments.
I bet are projects without other people involved. Without a PM, or a wide-company vision. Without needing to align with older codebase. With a single (small) codebase. No code-debt. Without required MR reviews. Most likely very few (if any) clients.
Remove all human interaction and work alone and I fully believe that you can be a 10x developer.
A 100x developer it’s a stretch. I think it’s still too much of an exaggeration.
But those are great news! I feel 5x to 10x when working on personal projects! I’m having loads of fun. But this doesn’t translate to my experience when working at a company.
What about corporate?
If you’re in a Product team, and AI is improving your feature lifecycle by 2x? It’s already great.
I think 10x is crazy, altho in some cases it might be attainable. 100x is just a figure of speech. It’s just not possible.
Are you feeling down because AI is everywhere around you and you’re supposed to be working at 10x, 50x or 100x? Is this causing you anxiety? Please relax and take it easy. If you’re doing 2x as a team, it’s great already!