ssp.sh
Dad. Technical Author, Data Engineer.
Data practitioner (20y) • Writing at ssp.sh since 2015.
Focused on the craft of data engineering & storytelling.
📚 vault.ssp.sh • 📖 dedp.online
❯ #dataengineering, #opensource, #writing, #obsidian, #neovim
3,514 posts
4,962 followers
1,089 following
Regular Contributor
Active Commenter
comment in response to
post
I also tried it, very cool. But unfortunately, it does not work on YT on AppleTV, where I'd use it the most :(
comment in response to
post
It works great for my coffee and book list, but to switch for my business CRM, I'd need the Kanban board (and color coding would be nice).
I'm looking forward to the next improvements and releases of Bases.
comment in response to
post
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
comment in response to
post
Ahh nice, this was a bug - with the latest version v1.9.2 this works:
```
filters:
or:
- file.hasTag("tag")
```
Before that, it gave me the attached error. More info on help.obsidian.md/bases/syntax
comment in response to
post
Thank you, Kepano. Is there a way to set a filter on "normal" tags, not a tag as part of properties?
e.g. `tag:#mytag` as explained help.obsidian.md/tags#Find+no....
I couldn't figure it out. Or is that not yet possible?
comment in response to
post
Correct. Sharp eyes 😉
comment in response to
post
I went with Presenterm, great, lean tool, including export as HTML and PDF 😍
comment in response to
post
Fun fact: The slides are done all in the terminal 🙃.
I really enjoyed Presenterm (github.com/mfontanini/p...), including converters (PDF & HTML).
🧾 PDF: www.ssp.sh/slides/25-06...
🌐 HTML www.ssp.sh/slides/25-06...
comment in response to
post
Amazing, thanks for coming Alexander. It was great connecting. Until next time.
comment in response to
post
Related, I love this by @stephango.com.
comment in response to
post
Yeah, and review for hallucinations 😅🙈
comment in response to
post
Rill has metrics, data models configured as YAML, so the LLM has all the context, plus the direct access to query the OLAP cube. A big step indeed
comment in response to
post
What would be the advantage of this mcp connect?
comment in response to
post
Do you see this if you go to settings? This is where I added it in the Claude Desktop on MacOS.
comment in response to
post
What do you mean, what is hidden? I have the regular subscription, not the max. Not sure if I would need it.
comment in response to
post
As always, I'll collect updates and the latest on vault.ssp.sh. See www.ssp.sh/brain/using-....
comment in response to
post
All it needs to do is add this to your local `claude_desktop_config.json`. Find more on docs.rilldata.com/explore/mcp.
comment in response to
post
I'm just gonna leave this here :)
comment in response to
post
Great read, thanks. To me, it's about having "soul". Nobody wants to read soulless text, or code. Creating the thing, the craft, the art, is the fun to me; not reviewing a blob of generated "stuff".
Attached are more of my thoughts, I also added thoughts from Graham, Hashimoto, AndrewNg, and more.
comment in response to
post
dbt says:
> As long as you aren't offering Fusion as a hosted or managed service to a third party, you should think about using Fusion in almost exactly the same way you think about using dbt Core today.
Questions to me still, is "executing it" considered a hosted service 🤔
comment in response to
post
Interesting thought. But would you consider running dbt fusion as part of an orchestrator, a "hosted service"?
If so, it wouldn't apply if you host yourself. But then it's a problem for all cloud providers, indeed.
comment in response to
post
I added an example as well. Looks very promising.
www.ssp.sh/brain/duckla...
comment in response to
post
Here are some more comments and links to recent announcements on www.ssp.sh/brain/duckla....
comment in response to
post
True, like a good old Static Site Generator. The best. I added Quartro, it seems really simple indeed, amazing.
comment in response to
post
That's the best, and what I do with my website.
Oh wow, you make your whole website with the "presentation" tool? But yeah why not, it's HTML :) Or do you use Quatro?
I will check it out more. The generated PDFs with Presenterm don't look as good. Presenting locally seems fine, though.
comment in response to
post
I hope you enjoy it.
Curious, what's your take on today's open data platform architecture built on open standards and formats?
comment in response to
post
If this interests you, I have provided various demos, such as:
- Live accessing Iceberg Tables with DuckDB and MotherDuck on S3
- Using DuckDB as a lightweight Data Lake access layer
- And a fun Bonus exploration with MCP with autonomous SQL queries to the MotherDuck database with Claude Desktop.
comment in response to
post
So, is that the future of databases or data warehouses, rebuilding database features on cheap storage?
Last week, I explored why open table formats are suddenly popular; this article focuses on where open table formats fit into the broader picture of #dataarchitecture.
comment in response to
post
OTF provides database-like features on top of distributed files. These features include SQL interfaces, versioning, ACID Transactions, and many more.
But does this data architecture represent the next evolution of or extend the Lakehouse core principle? What's the difference?
comment in response to
post
That's where databases shine, right? They offer numerous helpful features and a SQL interface for interaction.
This article dives into combination of these two options with data lakes and an open table format (Iceberg, Delta, Hudi, Lance) into a #opendataplatform.
comment in response to
post
Oh i didnt know that this revealjs is done by Quarto. I started with presenterm as revealjs seems very heavy.
But i like the html generation part. Can you deploy that easily to a webserver (just html and simple JS) or you need all kind of JS libraries on the server, do you know?
comment in response to
post
Will check it out, thanks for sharing. Usually for this kind of task I use Claude Code, it runs all sorts of Linux command to figure out what to do.
comment in response to
post
Shout out to @betterstack.com for the inspiration.
comment in response to
post
Thanks for sharing, and please let me know if there's anything relevant to this discussion. I'd be very curious to know
comment in response to
post
This is great. Also, time is an essential factor for the error/hallucinating rate, as in the first image.
And fun/competence, as well as the illusion of AI productivity. I wrote a little more on www.ssp.sh/brain/will-a....
comment in response to
post
If you want to know more about this, I wrote an article about it a while back.
comment in response to
post
Why speed? Because we can't wait 1 minute to get a simple query back. That's where the OLAP system or local DuckDB databases will shine.
Another key is code-first. As with that, you have a common ground with which AI and humans can interact.
Another example is attached that shows this with Cursor.