Sign in

Martin Klepsch

@martinklepsch.org
168 followers 182 following 183 posts

Web, AI and wherever curiosity will take us :) Oh and Clojure/Script of course

PostsRepliesMedia
Martin Klepsch @martinklepsch.org · 12/04/2025
I've tried a few Obsidian AI plugins but none really stuck, so I did what every sane person would do and wrote « repomix for Obsidian » Introducing `obsidian-context` github.com/martinklepsc...
github.com
GitHub - martinklepsch/obsidian-context: A CLI script to query and export content from your Obsidian vault, works fully local & read-only
A CLI script to query and export content from your Obsidian vault, works fully local & read-only - martinklepsch/obsidian-context
010
Martin Klepsch @martinklepsch.org · 11/04/2025
From what I hear enterprise suppliers are super under pressure to deliver the magical AI efficiency gains everyone has been talking about 😅
000
Martin Klepsch @martinklepsch.org · 10/04/2025
you gotta admire the raw-ness of @antirez.bsky.social youtube channel, lots of interesting stuff and mostly "just putting it out here" vibes 👌
1121
Martin Klepsch @martinklepsch.org · 27/03/2025
YOU CAN USE SVG! Its nuts, like it supports color scheme preferences, scales nicely etc
100
Martin Klepsch @martinklepsch.org · 27/03/2025
I’ve fully come around to file system based routing. yes it’s weird to have some of these [slug] filenames but wow, it’s just so much less boring glueing stuff together
000
Martin Klepsch @martinklepsch.org · 26/03/2025
What for? :)
110
Martin Klepsch @martinklepsch.org · 26/03/2025
TOS as in terms of service? Can you expand on that?
120
Martin Klepsch @martinklepsch.org · 25/03/2025
its the hero of ctxs.ai — maybe it works better in the context of the page
ctxs.ai
ctxs.ai context registry
An open-source, community-curated registry of contexts for use with LLMs
100
Martin Klepsch @martinklepsch.org · 24/03/2025
i think the thing about JS that bothers me the most is the laizzes-faire control flow, return anywhere, crazy syntax constructs for if expressions, … if it only was s-expressions 🥲
010
Martin Klepsch @martinklepsch.org · 24/03/2025
This! I had an important meeting an hour ago and unfortunately was not aware - transcript lost. 🥲
010
Martin Klepsch @martinklepsch.org · 24/03/2025
its becoming annoyingly commonplace that model choosers switch to 'auto' and then don't even clearly label what model was used and why 🥲 like how am I supposed to build ~intuition~
020
Martin Klepsch @martinklepsch.org · 24/03/2025
probably the easiest $12 I spent in a while. simple, plain analytics for all my sites by @drizzle.team at $12/year!
010
Martin Klepsch @martinklepsch.org · 23/03/2025
The thing with RLS is, you only really need it if you want to connect clients to the DB(?), but with Astro on the server that need kind of disappeared for me. A few actions with server-side validation is all it takes.
000
Martin Klepsch @martinklepsch.org · 23/03/2025
I've been using this stack for a project and it feels so much lighter than Supabase. Love Supabase but RLS and auth just try to solve for problems I don't have. - DB: Neon - Auth: better-auth - Frontend: @astro.build - Deploy: @fly.io - Jobs: Trigger
220
Martin Klepsch @martinklepsch.org · 22/03/2025
today I was at a restaurant and ordered MCP instead of MCC... that's what this has done to me
010
Martin Klepsch @martinklepsch.org · 21/03/2025
⋆ * ⋆ ✦ ✧ models <> intelligence ⋆ * ✧ context <> knowledge ✦ * ⋆ ✧ * ⋆ ⋆ *
010
Martin Klepsch @martinklepsch.org · 20/03/2025
Ah yes! I think I must’ve read it somewhere else but that was probably the source then!
110
Martin Klepsch @martinklepsch.org · 20/03/2025
Pricing these models high will allow OpenAI to deprecate those models without much backlash in the case that no use cases justifying the compute requirements are found. All this makes it even more fascinating that Sonnet 3.5 is still everyone's darling when it comes to coding
000
Martin Klepsch @martinklepsch.org · 20/03/2025
Can't remember where I read this but thought it was an interesting take: o1-pro, gpt-4.5 and other expensive models are effectively a "drop" to see if the market can find emergent, useful behaviors. Presumably because they haven't found them internally yet.
200
Martin Klepsch @martinklepsch.org · 19/03/2025
in a world of abundant creation, focus is everything must not forget
041
Martin Klepsch @martinklepsch.org · 19/03/2025
Could not quite make out a difference, but I'm on a particularly bad connection right now 🥲
000
Martin Klepsch @martinklepsch.org · 18/03/2025
it's criminal you can not export a cursor chat history
020
Martin Klepsch @martinklepsch.org · 18/03/2025
Feel free to use this (for free) dynogee.com/preview?id=1...
dynogee.com
Dynogee — Preview
Export dynamic images from Figma using the Dynogee plugin and get customized versions of your image via a simple HTTP API.
010
Martin Klepsch @martinklepsch.org · 18/03/2025
A bit late to this but tried with a use case I had a while ago and was somehow expecting greatness but instead was surprised to find that it garbled quite a few words, “Projektkategorie” turned into “Projektbiologie”, “Baubeginn” into “Boubejims” martinklepsch.org/posts/pdf-to...
I’ve also run a small experiment with Mistral’s latest OCR model after reading Simon Willison’s blog about it. Results were not very impressive unfortunately:

Several words were garbled, “Projektkategorie” turned into “Projektbiologie”, “Baubeginn” into “Boubejims”
I found these interesting because the PDF does have selectable text so these errors were unexpected
Here are permalinks to input and output.
000
Martin Klepsch @martinklepsch.org · 18/03/2025
Well meaning feedback: I find the AI generated OG image doesn’t do the effort that went into this justice - you have great branding, use it :)
100
Martin Klepsch @martinklepsch.org · 12/03/2025
really enjoying your writing! it hits the sweet spot of having great substance and being entertaining at the same time
010
Martin Klepsch @martinklepsch.org · 12/03/2025
explaining.software/archive/complex…
explaining.software
complexity as entropy
In a random dataset, there are no internal relationships; with each element, our explanation must begin anew.
020
Martin Klepsch @martinklepsch.org · 12/03/2025
really enjoying @ztellman.ideolalia.com's recent essays on explaining software, first the sudoku saga, now another great read
120
Martin Klepsch @martinklepsch.org · 12/03/2025
view transitions make this a one liner 🥳
030
Martin Klepsch @martinklepsch.org · 07/03/2025
believe it or not, i need this!
000
Martin Klepsch @martinklepsch.org · 07/03/2025
sometimes i think about just how good of a name "namespace" is its the absolute least you could put into that word
020
Martin Klepsch @martinklepsch.org · 06/03/2025
martinklepsch.org/posts/claud...
martinklepsch.org
Claude, the librarian
Personal Website and Blog of Martin Klepsch
020
Martin Klepsch @martinklepsch.org · 06/03/2025
I stopped myself from vibe coding my way to some organized markdown files and instead tried Claude code. It's cool! (and slow & expensive) But mostly cool. More things should be agents.
110
Martin Klepsch @martinklepsch.org · 05/03/2025
cant wait for when claude code gets fast, really an interesting way of working with local files. it scans and loops and mostly gets it all right
000
Reposted by Martin Klepsch
Steve Klabnik @steveklabnik.com · 03/03/2025
at one point i questioned my desire to help people get into open source image unrelated
a tweet by @primawesome

My neighbor told me coyotes keep eating his outdoor cats so I asked how many cats he has and he said he just goes to the shelter and gets a new cat afterwards so I said it sounds like he's just feeding shelter cats to coyotes and then his daughter started crying
13115884
Martin Klepsch @martinklepsch.org · 05/03/2025
The custom project rules editor in Cursor is a mess, you can put this in settings.json to just use a plain markdown editor. Shoutout to @rdotpy in the Cursor Forum!
000
Martin Klepsch @martinklepsch.org · 04/03/2025
If they’re as smart as me, never 😆
010
Martin Klepsch @martinklepsch.org · 04/03/2025
why do none of the cursor rules directories have rules files in the new format?
000
Martin Klepsch @martinklepsch.org · 04/03/2025
I’ve been thinking that blogs should have that again! Curious to see your implementation, might steal :)
000
Martin Klepsch @martinklepsch.org · 03/03/2025
the same person has about 20 apps in the app store so I think time just moves differently for them :P sindresorhus.com/apps
sindresorhus.com
Apps — Sindre Sorhus
Quality crafted apps by Sindre Sorhus
010
Martin Klepsch @martinklepsch.org · 28/02/2025
Are you making an example eval of a deep research thing for evalite? (I'd love to see that!)
000
Martin Klepsch @martinklepsch.org · 28/02/2025
FYI, I've opened a PR to include it in the docs: github.com/simonw/llm/p... Doing it as a plugin would be cool also but the plugin dev docs didn't quite give me an idea of how that'd work :-)
github.com
Add mention of `llm-web-ui` to logging docs by martinklepsch · Pull Request #773 · simonw/llm
Hi there, I thought this might be helpful to others as well so I thought I&#39;d open a PR and you can decide :) There&#39;s a 20s demo here. Also thought of adding a screenshot but wanted to keep ...
000
Martin Klepsch @martinklepsch.org · 26/02/2025
should I write a web UI for aider? 😆 kind of intrigued but also stop me pls
000
Martin Klepsch @martinklepsch.org · 25/02/2025
8000 to 128,000 is indeed a massive jump!
000
Martin Klepsch @martinklepsch.org · 25/02/2025
love that on Mac audio playback is interrupted when you switch between dark and light mode, dramatic
000
Martin Klepsch @martinklepsch.org · 25/02/2025
What is a comfy chair even
000
Martin Klepsch @martinklepsch.org · 24/02/2025
Thats much longer than what Perplexity generates. Relatedly, does chatgpt not have a "share" button? www.perplexity.ai/search/give-...
perplexity.ai
Give a high-level overview and also a deep dive for...
VideoMAEForVideoClassification represents a paradigm shift in video understanding, combining masked autoencoding's efficiency with transformer-scale...
100
Martin Klepsch @martinklepsch.org · 24/02/2025
I haven’t had an OpenAI subscription in a while so can’t really compare. Would be curious though. Maybe share a query here and I’ll share the same with perplexity deep research?
110
Martin Klepsch @martinklepsch.org · 24/02/2025
Fully back on the Perplexity train. Having access to R1, deep research, great search grounding all in one package is such a killer. Never been so fun to follow your curiosity ✨
120
Martin Klepsch @martinklepsch.org · 23/02/2025
Would be happy to contribute this to an eventual LLM plugin. The HTTP backend is very basic and the frontend could be packaged into a single <script> tag. Either way, would love to get your feedback if you give it a go!
000