Devansh @ashvend - Twitter Profile

@KP24 Sir Cook is not a T20 Specialist, whereas a talent like Bethell, has the ability to adapt quickly. Agreeing with KP in here, because Cook hasn’t been in an intense environment like IPL he doesn’t know the value.

0

383

Devansh @ashvend

3 months ago

@ItzMeTheProf This too shall pass, STD(Save the Date)

0

2

ashvend retweeted

Andrej Karpathy

@karpathy

3 months ago

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.

3K

60K

7K

108K

22M

Devansh @ashvend

4 months ago

hi @claudeai is your SSO login broken atm?

0

15

Devansh @ashvend

4 months ago

@Riki_tweets @social_man_007 Jacob Duffy is there too.

0

84

ashvend retweeted

TDAWGG

@tdawggVAL

4 months ago

The current Rank 1 Radiant is 810 RR... There are currently 221 Radiants in NA 3 Years ago, Zekken was 1v9ing pro lobbies for 1250 RR to secure Rank 1 while competing Pros are DAILY playing with and against: 0RR Immortal 1s Ascendant 3s Cheating + Wintrading + Sniping

38

4K

43

130

201K

ashvend retweeted

SR meL

@melanji

4 months ago

lukewarm take about the meta incoming: it’s fun to watch double duelist go brr.. but i imagine it sucks to actually play against. doesn’t feel like there’s a lot of counter-play vs a good neon/waylay jumping around (all while yoru is still busted) VALORANT rn is dictated by how well you can dissect micro situations and gain an advantage. taking favorable engagements + strong executes = king unfortunately, there just aren’t good ways to counter how busted these duelist kits are besides fast flooding / hitting insane shots teams have also gotten better at baiting out what little util there IS left in the game, making anchoring sites even harder (doesn’t help that sentinels don’t get a lot of value rn) i do think people are over-simplifying “oh the meta rn its just a bunch of 1v1s” when there’s actually a lot of thought put into how to make these fights advantageous as possible. ppl forget micro is just as important as macro ALL THIS TO SAY: i understand ppls frustration (both viewers and pros alike). it’s a drastic fundamental meta shift from what we are all used to. but the meta is just okay. if it had more more comp diversity and counters to these OP duelist kits, it’d be perfect debate is welcome 🙏

46

2K

27

86

112K

ashvend retweeted