Hey, Luca here, welcome to a weekly edition of the💡 Monday Ideas 💡 from Refactoring! To access all our articles, library, and community, subscribe to the full version:
Resources: 🏛️ Library • 💬 Community • 🎙️ Podcast • 📣 Advertise
1) 🔍 Codacy launches hybrid AI reviewer
This is brought to you by today’s sponsor, Codacy!
Codacy’s new code review engine actually understands your Pull Requests.
The new AI Reviewer combines the reliability of deterministic code analysis with context-aware reasoning to summarize scan results, ensure intent matches reality, and catch logic gaps conventional scanners miss.
On GitHub PRs, it flags the issues that actually matter, including AI-driven security risks and slop, so dev teams ship secure, high-quality code on every merge.
Use the code REFACTORING26 to get your first month off 👇
2) 🧠 Behavioral signal areas
With AI covering more bases on technical skills, behavioral interviews are quickly growing in importance, both for startups and big tech.
When organizations design these, they do it to assess traits that make employees successful. These traits may align with explicit company values (like Amazon’s Leadership Principles) or simply reflect the characteristics the org needs at its current stage.
These traits are often called Signal Areas and they typically are one or more of the following broad categories:
🎯 Scope / Impact — interviewers assess your scope or seniority and the complexity of problems you’ve solved. They’re looking for evidence that you can handle work at the appropriate level for the role.
🏃♂️ Initiative — are you a “self-starter” who proactively identifies areas for improvement? Do you drive adoption of your ideas?
☁️ Ambiguity — can you break down large, unclear problems into manageable pieces and get started? Do you prioritize work in a structured and appropriate way?
🏋️♀️ Perseverance — when faced with obstacles, how do you respond? Can you motivate yourself to overcome challenges?
⚔️ Conflict Resolution — how do you handle disagreements? They want to see that you can navigate these situations constructively without becoming a difficult colleague.
🪴 Growth — do you demonstrate a growth mindset? Are you intentional about learning from mistakes? Do you respond well to feedback and effectively provide it to others?
💬 Communication — beyond simply communicating clearly in the interview itself, hiring managers want to understand how you communicate in the workplace. What tools and approaches do you use to share information with various audiences?
Understanding these areas allows you to prepare more strategically for behavioral interviews, focusing your prep on stories that demonstrate strength in these dimensions.
We wrote a full guide on mastering behavioral interviews earlier this year 👇
3) 🏦 Design your API for the token economy
A weird constraint that APIs are facing today in the age of AI is that every byte costs money. When an LLM processes your API response, those tokens aren’t free, which creates a fascinating new design pressure.
So should field names be customerAccountIdentifier or just custId? That’s roughly 4 tokens vs 2. Not dramatic for a single field, but multiply that across dozens of fields, thousands of calls per day, and you’re looking at real costs.
On the other hand terse, abbreviated responses lose the self-documenting nature that helps both humans and AI understand your APIs. So you are optimizing for both comprehension and compression — a tradeoff that didn’t exist before.
Here are a few things some teams are experimenting with:
🎚️ Adaptive verbosity — PayPal’s API accepts a
VERBOSITYparameter that controls response detail. ML libraries like cuML use “adaptive” modes that adjust output based on context.🔗 Schema references — Similar to how Anthropic and Google implement “prompt caching” to reuse token sequences, APIs can reference cached schemas instead of repeating structures.
🗜️ Compression-friendly formats — Studies show Markdown is 15% more token-efficient than JSON, while TSV uses half the tokens. Some teams are exploring MessagePack or custom formats for AI consumption.
We wrote a guide on designing AI-first APIs a few months back 👇
4) 🎙️ Is React the last framework?
Earlier this year I interviewed Guillermo Rauch, CEO and founder of Vercel.
One of Guillermo’s ideas that stuck the most with me is that React may be the last major framework of its kind, as future frameworks will be designed with AI collaboration in mind.
“My colleague who’s our VP of AI, Jared Palmer says, it’s very likely that React and Vue and Svelte are the last frameworks... AI decides that the tool is better for its outputs rather than like a first principles new framework.”
New frameworks will also have an uphill battle to emerge because AI will have seen fewer examples and will naturally be less good at those.
Here is the full interview with Guillermo:
You can also find it on 🎧 Spotify and 📬 Substack
And that’s it for today! If you are finding this newsletter valuable, consider doing any of these:
1) 🔒 Subscribe to the full version — if you aren’t already, consider becoming a paid subscriber. 1700+ engineers and managers have joined already! Learn more about the benefits of the paid plan here.
2) 📣 Advertise with us — we are always looking for great products that we can recommend to our readers. If you are interested in reaching an audience of tech executives, decision-makers, and engineers, you may want to advertise with us 👇
If you have any comments or feedback, just respond to this email!
I wish you a great week! ☀️
Luca
The token-cost lens on API design is underrated. Once you realize compression isnt just bandwidth optimization but also an economic constraint, patterns like adaptive verbosity start making more sense. The PayPal VERBOSITY parameter feels primitive now but its actualy ahead of its time as a design idiom. The real question is whethr we'll see protocol-level token accounting built into HTTP/3 or if this stays application-layer forever.