The future of Hybrid Hacker! 👓

Following up to last week's experiment

Hey, Luca here!

Last week (actually two weeks ago! I took one off 🙏) I published a different email than usual, which instead of being an essay, was a short list of the best articles and tools I had found that week.

I did that because so many people ask me what to read these days, with AI and everything else going on.

So I made what I read public via this Mailbrew digest that you can clone / subscribe to by yourself, and condensed the very best items in that very email.

Then I asked readers what they thought about it, and the result was… astounding! 👇

The poll from last week’s newsletter

So I got back to the drawing board, and thought deeper about how to turn this into a permanent, high-quality thing.

As most of you know, I am a full-time content creator, and other than Hybrid Hacker I also run another newsletter, Refactoring.

The way I think about my work is that of a curator of ideas. I read a lot of stuff, distill the best, come up with my own ideas, and combine them into long-form essays.

This work, which is largely individual and private, has also a public version attached, which is what makes me a creator, as opposed to just someone who likes to study things 👇

What’s interesting is that not all the things I do have such a public version. Namely, up until last week, both my reading list (sources) and the short list of things I have read and liked have always been 100% private.

So I realized that what I unintentionally did last week was attach a public version to these two remaining bits, in the form of the Mailbrew digest and the Hybrid Hacker newsletter.

I really like this idea, because it fixes two problems at once:

• It creates more incentives for me to do things well — as now it’s literally all public.

• It creates a clear separation between Hybrid Hacker and Refactoring, which have always had a lot of overlap otherwise.

So I am going to keep this going for a while and see how it turns out!

You may have questions, so feel free to reply to this email and I will get back to you. The most common one might be: what if you are a paid subscriber of Hybrid Hacker? I want these weekly digests to be free for everyone, but since you subscribed (and paid for) long-form essays, I am happy to convert your HH subscription into a Refactoring one at no additional cost (Refactoring costs $15 instead of $10, so consider this a Black Friday deal!). Refactoring, if you are not familiar with it, covers the same topic, with the same style, and also has a podcast, community, and large library attached to it!

If you prefer to unsubscribe instead, just reach out and I will issue a prorated discount. But I really encourage you to stay and see what we are up to!

So, if you missed last week’s email, you may wonder what these digests look like, so here is the one for this week! 👇

---

👓 Weekly Readings

---

1) Thoughtworks’ Technology Radar

30 min • by Thoughtworks

This bi-annual tech radar is simply a must-read every time it comes out. Get yourself a cup of your favorite hot beverage and go through at least the Adopt and Trial sections.

---

2) Understanding Spec-Driven Development

13 min • by Birgitta Bockeler (↗ our interview with her)

SDD is one of the latest buzzwords in working with LLMs, and Birgitta did an awesome job at demystifying it from first principles, plus trying a bunch of tools for it. More resources:

↗ SDD: The Waterfall Strikes Back — a counter-take

↗ Kiro • Spec-kit • Tessl — the tools tried by Birgitta

---

3) How to Write a Great AGENTS.md

7 min • by Matt Nigh (Github)

This article was super useful to me to get practical examples of which agents you should create, how to set their boundaries, and how to structure the agent files.

---

4) Will Larson is back to coding at work after a decade away

5 min • by Will Larson (Imprint)

Will is a legend and this is the best article I have read about how managers can get back to coding thanks to AI, and whether they should. It is a balanced take that includes opportunities, risks, and how to do it well.

---

5) Maybe You Are Not Actually Trying

6 mins • by Cate Hall (Astera)

Amazing article about agency. Sometimes we stop trying in one area of life because we got stuck once and never revisited. That frozen response can look like effort, but it isn’t so. It gave me a lot to think about.

---

🔧 Weekly Tools

---

1) Nano Banana Pro

Mind-blowing new model by Google for image generation. It is especially great at diagrams and infographics, most often one-shotting them with perfect text — and in 4K!

2) Google Antigravity

Google was on fire this week and also released its very first IDE. It’s a fork of VS Code with a more agentic spin. I tried on a few projects and it feels different than Cursor, will probably write a longer take at some point. I appreciate it allows for multiple models, but I guess it’s still not 100% optimized because I keep hitting quota limits, apparently for concurrent requests.

3) Kiro / Spec-kit / Tessl

After reading the SDD article above I went into a rabbit hole and tried a bunch of tools for it!

4) LLM Council

Nice open source project by Karpathy where your questions are sent to multiple LLMs and each model reviews and ranks the others’ answers anonymously. Similar to Pewdiepie’s experiment, which I also enjoyed a lot.

---

And that’s it for today!

Sincerely 👋
Luca