Simon Späti 🏔️

I loved reading all the Harry Potters when I was young. Maybe I need to find another genre. But they'd need to be not so thick

The best I got is biographies, which I much more enjoy as someone has lived through it, instead of just a made-up story. Which fiction do you like?

I know, I tried many times 😬. I still have "The Graveyard Book", recommended as the best fiction by Tim Ferriss, and the all-time favorite The Hitchhiker's Guide to the Galaxy downloaded on Audible.

They just bore me after a while, I don't know why.

I think slowing down is needed, the more I use them. (just thinking out loud here)

I think there's also the tricky part. If you can explore a code base much faster, same as summarizing a book, but did the statistical model highlight the parts you'd have evaluated as well?

Or is it showing stuff you'd never have looked at when exploring manually, and slower?

Vibe Coding: Build Fast, Think Slow.

As Daniel Kahneman said in his book Thinking, Fast and Slow, it might happen the same to working with AI, but with «Building Fast, Thinking Slow».

What do you think?

Vibe Coding Vibe coding is the next evolution of [[Prompt Engineering]], when you instruct your [[AI Agents]] to run autonomously. The best example is [[Claude Code]], but there are more.

More thoughts at the usual place. www.ssp.sh/brain/vibe-...

This was also the message of PewDiePie's video about training his own model, where he struggled so much that he learned a lot. With vibe coding on autopilot, this is going away fast (?).

Is craftsmanship gone with AI? You can vibe code something on demand, but you have no attachment to what it's built, so how are you gonna sell or rave about something you don't even have any connection to?

No hardship, no challenges, no learnings with AI.

My book recommendations for history, starting from 2018 to 2025.

I love books, and even more so, I love recommendations from other fellow readers.

That's why I share my book recommendations, starting from 2018 to 2025.
www.ssp.sh/books/

One more reason why you should still write your blog, and why I still blog.

I tested a bit more. It's great. But still lacking features compared to Obsidian-project plugin:

- Kanban-style (e.g., for my Sales Funnel)
- Color tagging

Rill | Building an Agent-Friendly, Local-First Analytics Stack with MotherDuck and Rill The convergence of embedded analytics engines (DuckDB/MotherDuck), declarative BI-as-code (Rill), and AI agent protocols (MCP) is creating a new architecture for business intelligence, one where dashb...

Again, if you like this, read the full article at: www.rilldata.com/blog/buildin...

What do you think, and which agent-friendly, local-first analytics stack do you prefer these days?

We explore three examples:

1. Stack Overflow Developer Survey with zero-pipeline analytics

2. Updated Multi-Cloud Cost Analyzer to add MotherDuck with `DLT_DESTINATION=motherduck`

3. Including a real-world setup from Drio Tech, how he uses Airbyte/dlt + BigQuery/MotherDuck/dbt + Rill.

And we check #limitations the future holds due to the imprecise way of natural language, e.g., "give me the analytics for this week?" Did you mean "from today until last week"?

Or how data workloads need to be #deterministic and #reproducible to backfill faulty data, but AI agents aren't that.

We explore more:

→We discuss agentic coding—no more dashboards needed?
→Natural language interfaces: convenient, but inaccurate?
→Does self-serve BI-as-Code provide the #context we need?
→How language choice matters—SQL, YAML, chosen for Agent use (not human familiarity).

Getting started with an example project - zero pipeline setup.

In this new essay we look into general data engineering workflows and working with new agent tooling, why you'd use @motherduck.com, and how @rilldata.com can help you in this stack.

We look at #ConversationalBI, turning dashboards into code (and code into agent-readable context).

Iterate with prompt, generate and ship

These stacks are a joy to work in, and the iteration cycles are super fast from:

`ambiguous prompts -> verifiable configs -> shipping visualizations`

Increases quality by allowing domain-experts knowing the business best to own more of the stack, while everything is still versioned, automatable.

Rill | Building an Agent-Friendly, Local-First Analytics Stack with MotherDuck and Rill The convergence of embedded analytics engines (DuckDB/MotherDuck), declarative BI-as-code (Rill), and AI agent protocols (MCP) is creating a new architecture for business intelligence, one where dashb...

CLI-first is eating development. Email, calendar—they all have CLIs now. Why not your business metrics?

For data/analytics engineers building with agents: DuckDB + MotherDuck + Rill give you an #agentfriendly, #localfirst frontend—exact context via SQL and YAML.

I just parse the notes as I have a second brain, and my blog is all locally. So it's very easy to see and check, plus I already have a graph view from my second brain, provided by Quartz (a framework for Obsidian notes publicly).

Great point - I just added a small weight to show notes that are multiple times linked or link forth-and-back are bold and are centered more.

All on the notion that the future of blogging is more connected.

And updated link management on my website and second brain, to better see internal (keep you on the page) and external links.

I hope you enjoy.

This is part of recent updates on my overall websites — e.g., added a graph to my patterns book too
bsky.app/profile/ssp....

backlinks and links on the interactive graph on https://www.ssp.sh/blog/linux-omarchy-the-good-bad-and-fixable/

I added an interactive graph for my blog posts too (collapsed by default) — so you see outgoing links to other posts, second brain, or backlinks brain visually at the end of every blog.

E.g., here is a recent Omarchy Linux article (see ssp.sh/blog/linux-o...).

What do you think?

Daylight savings, one of the hardest data engineering problems :)

Open-Specific Files with Obsidian and Neovim I wanted to open my todo or daily note with a single vim-shortcut or also hotkey from Obsidian. There was not possible, so I create a very simple plugin. Check it out at: Obsidian-open-specific-notes (A small and simple plugin for Obsidian to quickly open a specific note (e.g. todo, daily journal).

Configs and settings for Neovim and Obsidian here: www.ssp.sh/brain/open-...

Todo's: Need Constraints One thing I learned about todos, it’s never about the [[Todo Apps|todo app]], it’s about having constraints to not add anything immediately. Important things will resurface again. Urgent matters in the moment, but irrelevant tomorrow, will save us a lot of time.

Check more at www.ssp.sh/brain/todos. Or accessing the weekly-todo note quickly with `space+space` on Obsidian and Neovim - check out the configs and settings below.

It's the digital version of a weekly todo-list in addition to the physical one.

I do that too. And for long-term tasks, I add them to my Todo-apps (no default reminder, no review - just adding them to be ready when I have time).

It's also about constraint. Most todo apps make it too easy to add todos, making it overwhelming. With AI auto-adding tasks, it's no better.

Cal Newport's suggestion: delete social media apps. Use the browser instead. You'll get fewer notifications, less tracking, and natural friction.

Better yet: try a 30-day sabbatical. See what you actually miss versus what you thought you'd miss.

The most advanced technology companies employ neuroscientists to make their apps addictive.

You are already addicted. The question is what to do about it.

ssp.sh/brain/digit...