Eduardo Yamauchi

1.7K posts

Eduardo Yamauchi banner
Eduardo Yamauchi

Eduardo Yamauchi

@dudu314

Katılım Nisan 2009
182 Takip Edilen56 Takipçiler
Eduardo Yamauchi
Eduardo Yamauchi@dudu314·
@Pshichenko I ran the experiment. In my city none were available, but I found a set of plates about 150km away (just the plates, not the bar) for $1000 😩
English
1
0
0
22
Roman Pshichenko
Roman Pshichenko@Pshichenko·
I can now deadlift at home, thank you fb marketplace
English
3
0
3
114
Andrej Karpathy
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.
English
2.8K
6.7K
56.3K
20M
Eduardo Yamauchi
Eduardo Yamauchi@dudu314·
@DuquesaDetax Já foi pro Japão? Lá é imbatível, o desconto é no ato da compra. Pena que vai deixar de ser assim.
Português
0
0
0
30
Maria Carolina Gontijo
Maria Carolina Gontijo@DuquesaDetax·
Surreal a experiência do tax free na Itália. Tudo em uma máquina, ninguém deles pra conferir se eu estava com os produtos, pá, pum, 5 dias depois e os euros estão na wise. Esquisita demais essa eficiência.
Português
18
5
445
29.3K
Rodrigo Campos
Rodrigo Campos@roderix1966·
Estou achando essa série Pluribus da Apple simplesmente sensacional. Em todos os sentidos. Atuação, enredo, filmografia…muita gente não curtiu mas estou amando! Louco?
Português
34
2
169
14.5K
Felipe Leme
Felipe Leme@LEME12·
Essa é a última sobrevivente. Se vierem com "nova formula", precisaremos pegar em armas
Felipe Leme tweet media
Português
228
78
3K
107.2K
Rafael Zattar
Rafael Zattar@ZattarRafael·
Agora tenho mais uma filha… Compro uma cartela de 30 ovos a cada 2 dias … 14 pra mim e 16 pra frigideira. Aqui o FGC não cobre. Aliás, F G O (fundo garantidor de ovo). LAELE
Rafael Zattar tweet media
Português
276
7
650
120.6K
xandeco telecoteco
xandeco telecoteco@XandeReformed·
fiquem tranquilos crianças o AleXANDECO de Moraes jamais deixaria vocês na mão
xandeco telecoteco tweet mediaxandeco telecoteco tweet mediaxandeco telecoteco tweet media
Português
4
1
27
539
Eduardo Yamauchi
Eduardo Yamauchi@dudu314·
@roderix1966 Uma vez tava num voo internacional e tô escutando meio de longe um pessoal falando português, meio alto mas não dava pra entender... Quando me aproximei, vi que tavam falando é russo!
Português
1
1
51
2.3K
Rodrigo Campos
Rodrigo Campos@roderix1966·
Sabe que tive uma amiga cujo pai foi diplomata na USSR. Ela me disse que se pegasse um texto em cirílico e colocasse no nosso alfabeto (a-z) e lêssemos em voz alta em “português” eles entendiam tudo!
Português
19
8
515
31.6K
Rafael Zattar
Rafael Zattar@ZattarRafael·
Hahahahhahahahahahhahahha
Rafael Zattar tweet media
Filipino
10
1
63
11.5K
Biologia Braba
Biologia Braba@biologia_braba·
Fato curioso: Os tubarões não conseguem fazer a manutenção de um ar-condicionado LG dual inverter.
Biologia Braba tweet mediaBiologia Braba tweet media
Português
412
3.8K
24.8K
1.9M
O Grafista | Investidor 🇧🇷🇬🇧
Hoje o abriu as pernas é trade foi no Dow Jones. Ainda preso no índice e no dólar? O mercado internacional é muito mais barato. Comenta livro que te mostro como
O Grafista | Investidor 🇧🇷🇬🇧 tweet media
Português
20
0
28
3K