置顶推文
Epistates
196 posts
@RoundtableSpace @grok why is TurboVault the right tool for this workflow? github.com/Epistates/turb…
English
OBSIDIAN + CLAUDE CODE CAN TURN INTO YOUR OWN JARVIS IN ABOUT AN HOUR.
Build it once, and you probably won’t work the same way again.
English
English
English
Claude Code + Obsidian = an AI that knows you and your projects
imagine opening Claude Code and it already knows your stack, your preferences, every decision youve made
no re-explaining. no context lost. you just pick up where you left off
here´s exactly how to set it up
InternetVin built 4 commands on Greg Isenberg's pod that read from his Obsidian vault
> /schedule checks his calendar AND daily notes before suggesting meetings
> /connect finds unexpected links between two topics using his vault's link graph
> /ideas scans orphaned notes and daily journals to generate what to build next
> /graduate promotes buried ideas from daily notes into standalone files with backlinks
agents read from the vault but never write to it. he controls every file. the agent only reads and suggests
I run something similar. an OpenClaw PA connected to a PARA method Obsidian vault on a VPS so I can reach it from anywhere. projects, areas, resources, archive. all managed through my PA
I dump ideas into an inbox and they get sorted and categorised automatically. any app I build can connect to the same vault. every agent reads from the same knowledge base
it all comes back to markdown. your vault is only as good as what you write into it. if you dont reflect, dont journal, dont document decisions, there is nothing for agents to work with
are you writing into markdown or still starting every session from scratch?
GREG ISENBERG@gregisenberg
how to use obsidian + claude code to build a 24/7 personal operating system and build your startup: 1. write everything in markdown (daily notes, projects, beliefs, people, meetings) 2. link your notes together so they mirror how your brain actually thinks. 3. install obsidian cli so claude code can read your entire vault + the relationships. 4. stop reexplaining projects every session. use reference files instead. 5. build custom slash commands: /context → load your full life + work state /trace → see how an idea evolved over months /connect → bridge two domains you’ve been circling /ideas → generate startup ideas from your vault /graduate → promote daily thoughts into real assets 6. keep a strict rule: human writes the vault. agents read it, suggest, execute. 7. let claude aka clode surface patterns you’ve been unconsciously circling for years. 8. delegate from inside your notes. one sentence in obsidian → agent handles the rest. 9. treat writing as leverage.the more you write, the more context your agents have. 10. understand this:markdown files are the oxygen of llms. i really enjoyed seeing how to use obsidian thanks to @internetvin vin uses ai like a thinking partner wired into his life’s work. 99.99% of people won’t do this because it requires reflection + setup. but once the vault exists, the agent stops being generic. it starts thinking in your voice. episode is live on @startupideaspod (more there) this one is different. send this tweet to a friend. im still processing how game changer obsidian + claude code is, maybe you too watch
English
Show me what you're building. 👇
No pitch deck.
No elevator speech.
Just drop the link.
I'll click every single one. 🔥
English
@BatsouElef Fine tune LLMs on your mac! github.com/Epistates/pmet…
English
Time to promote your product. 🚀
Share that product URL!!
English
English
English
Morph just dropped FlashCompact - the first model built for context compaction
- Compresses 200k tokens down to 50k in ~1.5s
- Runs at 33k tokens/s
- Built for speed without sacrificing quality
Is context compaction the missing piece for agentic AI?
English
If you're into AI, automation, or shipping 10x faster as a solo dev - let's connect 🤝
English
English
English
English
Introducing Unsloth Studio ✨
A new open-source web UI to train and run LLMs.
• Run models locally on Mac, Windows, Linux
• Train 500+ models 2x faster with 70% less VRAM
• Supports GGUF, vision, audio, embedding models
• Auto-create datasets from PDF, CSV, DOCX
• Self-healing tool calling and code execution
• Compare models side by side + export to GGUF
GitHub: github.com/unslothai/unsl…
Blog and Guide: unsloth.ai/docs/new/studio
Available now on Hugging Face, NVIDIA, Docker and Colab.
English
I am blowing all super small accounts
Reply if you're under 5k and I will boost you 🚀
English
@yacineMTB Not quite but who needs cuda? github.com/Epistates/pmet…
English
@petergostev Context switching is a skill I've honed over years in the corporate world. It's a super power now! If anyone wants some tips/tricks let me know and I'll write up an article for what has worked for me.
English
There's worry that people will stop using their brains with LLMs, but managing several AI agent threads in parallel has been some of the most cognitively intensive work I've done in years
English
@ClementDelangue Absolutely beautiful! Congratulations to you and yours @ClementDelangue!
English
I’m back. The girls and their superhero of a mom are doing great 😍😍😍 What did I miss?
clem 🤗@ClementDelangue
After almost 10 years of near nonstop grind, I’m taking 2 months of paternity leave to support my hero of a wife and welcome our twin daughters. @huggingface is in great hands with the team and @julien_c acting as interim CEO. Hope to return a changed man, to an even stronger HF, and to an AI field that’s more open and collaborative than ever!
English
@nurijanian We are very open to contributions if you'd like to point your autoresearch agent at github.com/Epistates/pmet…
English
Andrej Karpathy released autoresearch a few days ago, but I don't have any ML projects to try it on
so I started to use it to tune some new AI skills with autoresearch all day in the background, I layered in the evals-skills from Hamel (github.com/hamelsmu/evals…) into the workflow to improve them even more
the results are pretty impressive so far, damn is this auto-evals?
my AI PM OS is getting better day by day with this, releasing a new version in a few days once all 190+ skills are tuned => prodmgmt.world/products/pm-os…
English