nishil

following my curiosities | prev @chapterone @decentralisedco @biconomy

[nishil @_nishil_]

🧵 Stablecoins and the Overhaul of Global Credit

Lets explore how stablecoins are helping the US distribute its debt globally — and why stablecoin issuers may become the world’s most efficient credit allocator.

[nishil @_nishil_]

Big fan of @ajay_shah , his opinions and the way he presents them.

[nishil @_nishil_]

Jai bhim ke lie na kar, par jai hind ke lie karde?

It’s fair to also want talks about reservations. But it’s like asking for talks on chori when there’s been no answers on murders.

Stand for the continued unaccountability if nothing else, it’s so much bigger than our personal agendas now.

[nishil @_nishil_]

welll guysss shut up with your gatekeeping this fucking mid everyone would know cafe and your cringeworthy worshipping of a fucking AI model improvement and the cocksucking of your love for latent, there’s a crisis going on in the country, let’s fucking give fucks about that for a bit

[nishil @_nishil_]

need a ‘tujh sang preeeeet hamne lagayi hai….’ moment to my writings

[nishil @_nishil_]

@peegeekay @zamana_hq not sure if they for indian consumer

[nishil @_nishil_]

@peegeekay @zamana_hq is the only one I’ve heard of

[nishil @_nishil_]

@anieasyy @peegeekay can you share deets

[nishil @_nishil_]

@mmmaahol ❓

[nishil @_nishil_]

fucking stop the internal monologue

[nishil @_nishil_]

@mmmaahol how

[Jaya Gupta @JayaGup10]

Anthropic and OpenAI are like TikTok for enterprise data, except the feed is your work and the engagement signal is your judgment

[nishil @_nishil_]

How is there no consumer facing product yet that makes it easy to store and access your thoughts and research?

Seems like a no brainer for startups to chase this problem. A product with UX done right is an immediate sell to strategists, creators, founders, researchers.

What am I missing?

[Andrej Karpathy @karpathy]

LLM Knowledge Bases

Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So:

Data ingest:
I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them.

IDE:
I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides).

Q&A:
Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale.

Output:
Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base.

Linting:
I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into.

Extra tools:
I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries.

Further explorations:
As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows.

TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

[nishil @_nishil_]

@codein_artout registered!!

[nishil @_nishil_]

just watched common side effects, such an interesting show

really tempted to get a tortoise and name it socrates 🐢

[nishil @_nishil_]

@hypersoren when we say that outcomes are only observable by the client, it might be a generalisation of the problem space - some might not need internal information, as long as the deliverables are defined prior and externally verifiable.

the one that do need are worth thinking about fs

[nishil @_nishil_]

@jinglingcookies thankss!

[nishil @_nishil_]

https://t.co/o6ZIHCBToY

[nishil @_nishil_]

@kakashiii111 Interesting!

I believe this will give rise to outcome based pricing with companies competing on costs, and it wouldn’t matter what model is used in the backend as long as the set standards are met and verified.

[nishil @_nishil_]

must be doing something wrong for me to have discovered @edzitron so late

[nishil @_nishil_]

flirting on a habit tracker app has to be the peak of wholesome