AG
87K posts

AG
@alexserver
Fix and improve. I love to drive. I like to build products. Founder of https://t.co/ZMWkRsxzb4


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.

Los layoffs no son gratis y parten de una mentalidad cortoplacista en muchos casos. Siempre afectan, pero hay casos más justificados que otros. No es lo mismo una empresa que está sufriendo, a casos como Oracle que tienen ingresos record. Ahí hacer layoffs destruye la cultura y la lealtad de los empleados, que ya de por sí no suele ser el punto fuerte de una corpo. Vas a esforzarte igual sabiendo que puedes caer en la siguiente ronda? Si fueran layoffs con stack ranking pues bueno, pero muchos de estos son casi aleatorios, no se le pregunta a los managers. Por eso hay casos de empleados TOP cayendo como moscas. Y eso quita las ganas de trabajar duro. Cuando empecé en Salesforce la cultura era increíble, realmente parecía que se preocupaban por los empleados. Tenía valoraciones de 4.4-4.6 en webs como Blind, Glassdoor etc. Cuando anunciaron layoffs hace 3 años con beneficios record y recortaron los bonus, la cultura se fue a la mierda. Se respiraba negatividad en los all hands, canales de Slack, foros, Blind etc. Dos meses después de los layoffs el CEO salió diciendo que si la gente estaba sufriendo que pillaran un tiempo para desconectar como él, que se había ido un par de semanas a la polinesia francesa xddddd totalmente delusional el cabrón. Cayó a un 3.9 y aún a día de hoy no se ha recuperado, los comentarios de la empresa son horribles por gente que trabajó y aún trabaja allí. La confianza y la cultura se contruye en años, pero se destruye en segundos.



Operación de Estado para exonerar a Pío López Obrador. La presidenta dijo que el hermano del expresidente Andrés Manuel López Obrador no cometió ningún delito y lo respaldó.

🔴 ALERTA MUNDIAL: PIDEN A LA POBLACIÓN QUE VUELVA A TRABAJAR DESDE CASA • La Agencia Internacional de la Energía (IEA) recomienda más teletrabajo, bajar la velocidad al conducir y COCINAR CON ELECTRICIDAD para ahorrar energía frente a “la mayor interrupción del suministro de petróleo y gas en la historia”.


After much reflection, I have decided to resign from my position as Director of the National Counterterrorism Center, effective today. I cannot in good conscience support the ongoing war in Iran. Iran posed no imminent threat to our nation, and it is clear that we started this war due to pressure from Israel and its powerful American lobby. It has been an honor serving under @POTUS and @DNIGabbard and leading the professionals at NCTC. May God bless America.


Tough guy walks into a car meet demanding to take on over 200 people











