Nephila Radar [1/26 | Aug 31 – Sep 13]
Table of contents
How does the Nephila team keep up with this crazy, ever-changing world of tech? We don’t just follow what’s happening. We discuss the ideas and perspectives that we find most relevant and work out together how they can bring value to what we do every day.
Starting today, we’re sharing some of the most interesting reads we come across. Every two weeks, you’ll find a new post on this blog featuring the resources that got us talking and thinking.
Want to join the conversation? Come and find us on our LinkedIn page. Sharing ideas and comparing perspectives is the part we enjoy most.
On our radar
Christoph Nakazawa, Engineering Director at Cloudflare and CEO of Nakazawa Tech, shares concrete examples from projects he has shipped to explain how he integrates AI into his workflow without giving up control. His approach puts human oversight front and centre: in one of his React libraries, for example, 80% of the code is AI-generated, but control remains firmly in human hands. He remains sceptical of fully autonomous agents.
Worth reading for a closer look at hands-on approaches to AI-assisted development: You Are Absolutely Right!? - Christoph Nakazawa
Pavel Samsonov, author of The Product Picnic newsletter, criticises “vibe prototyping”: AI prototyping tools allow anyone to build prototypes while bypassing the contributions of designers and developers, all in the name of “speeding up” the process.
Worth reading for a clear-eyed critique of how AI, when used without intention, can erode processes rather than accelerate them: Vibe Prototyping Isn’t Solving Any Problems (But It’s Creating Many New Ones) - Pavel Samsonov
Laura Summers, Lead Design Engineer at Pydantic, explores the psychological impact of AI-assisted development on developers. She describes the “fatigue of supervision”: spending days specifying requirements for an AI, reviewing dozens of AI-generated pull requests, and constantly correcting course can lead developers to question their own role.
Worth reading for an honest perspective—neither hype-driven nor alarmist—on the human side of AI-assisted development: The Human-in-the-Loop Is Tired - Laura Summers
From the Nephila blog
AI is reshaping where value actually gets created in a software project. At Nephila, we're rethinking our workflow from analysis to delivery and we packed what we're learning into this piece.
Worth reading to understand why architecture and documentation can't be fully handed off to an LLM: Vibe coding or software engineering? How AI is redefining value in software projects - Iacopo Spalletti
Try it: can you debug AI Code?
Writing code with AI is just part of the job now. But can you actually spot the bugs it generates? Find out with the AI Debug Challenge we built. It's a fun but genuinely tricky quiz that will show whether you need to sharpen your debugging skills, or whether you're already on the right track. You can share your result too!
Worth trying to put your debugging instincts to the test: AI Debug Challenge → Take the challenge
***
We wrote this article with the support of AI. If you’re interested in how we use these tools as part of our writing process, you can read more here.
