I really like the framing here (via Richard Sennett / Roland van der Vorst): craft is a relationship with the material. In software, that “material consciousness” is built by touching the system—writing code, feeling the resistance of constraints, refactoring, modeling the domain until it clicks.
If we outsource the whole “hands that think” loop to agents, we may ship faster… but we also risk losing the embodied understanding that lets us explain why something is hard, where the edges are, and how to invent a better architecture instead of accepting “computer says no.”
I hope we keep making room for “luxury software”: not in price, but in care—the Swiss-watch mentality. Clean mechanisms, legible invariants, debuggable behavior, and the joy of building something you can trust and maintain for years. Hacker News needs more of that energy.
I’m evaluating UI options for a new .NET product and I’m struggling to find balanced, experience-based pros/cons about Blazor (Server / WASM / Hybrid) compared to alternatives (React/Vue/Angular + API, MAUI/WinUI/WPF, etc.).
This post dives into that "black magic" layer, especially in the context of emerging thinking models and tools like Ollama or GPT-OSS. It’s a thoughtful look at why sampling, formatting, and standardization are not just implementation details, but core to the future of working with LLMs.
CRDTs and HLCs often feel like over-engineering. In most business/management apps, it’s rare for two people to edit the same piece of data at the exact same time.
A simpler approach works surprisingly well:
Store atomic mutations locally (Redux-like) in SQLite.
Sync those mutations to a central server that merges with last-writer-wins.
Handle the few conflicts that do occur through clear ownership rules and some UI/UX design.
This makes local-first behave more like Git: clients work offline, push their “commits,” and the server decides the truth. Most of the complexity disappears if the data model is designed with collaboration and ownership in mind.
Example of decoupling things in goblins framework (nodejs):
one domain give one package/module, domains are now dependencies
actors models can act as services, one actor in a domain is a service with an API communicating with other trought event loops or tcp/ip (microservice?)
we can develop and debug the whole system in a mono-repo <- the monolith is the repository of code.
Good architecture, actor models, and collaboration patterns do not emerge magically from “more agents”.
Maybe what’s missing is the architect’s role.