Thierno

this dude created a Gemma 4 that is 6X faster than Google's version! Gemma 4 is notorious for: • being the most optimized free open source local AI • that works with even old laptops and smartphones • and now it's insanely fast 6X Dflash Gemma 4: github.com/z-lab/dflash

"Mon argent est en sécurité à la banque" "Au moins il n'est pas volatile, "1€ = 1€" 🧊Pendant ce temps l'argent :

People are asking this question

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.

Ce livre co-écrit avec @PAPlaquevent sera, à mon sens, incontournable pour comprendre ce qui se déroule au Proche-Orient, en #Israel, et pour saisir la dimension apocalyptique de notre époque. Préfacé par l’historien @edouardhusson En prévente 👇 cultureetracines.com/essais/142-isr…

Throughout the west, criticizing the Netanyahu government is now a crime punishable by imprisonment. Glenn Greenwald on the end of free speech. (0:00) The Current State of Free Speech (12:52) Are Jewish Students Underrepresented at Ivy League Schools? (16:10) Israel's Influence in America (23:54) Why Are Republicans Attacking Free Speech? (28:51) Where Are All the Protests? (34:34) Why Is War the Only Thing Both Parties Agree On? (36:36) Ron DeSantis's Strange Trip to Israel (40:44) Is Ethnic Conflict Being Intentionally Promoted? (50:04) The Future of American Politics and the Israel Lobby (52:43) Charlie Kirk (57:57) What Will America Look Like After This War? (1:02:10) Will Americans Be Arrested for Their Opinions? (1:09:07) Will We See Violent Attacks in the US?

Map of U.S.-linked locations hit by the Islamic Republic of Iran, and high value Radars confirmed damaged or destroyed.

Seyyed @fr_Khamenei , un homme d'exception Retour sur la personnalité exceptionnelle du Guide Suprême de la République Islamique d'Iran, Seyyed Ali Al Khamenei, et sur sa modestie légendaire.

🇮🇷🇺🇸 THE 7 WAYS A U.S. STRIKE ON IRAN COULD GO OFF THE RAILS, AND NONE OF THEM ARE “SMOOTH” Everyone’s saying the U.S. looks days away from hitting Iran, and the menu of possible outcomes ranges from “hopeful fantasy” to “welcome to the apocalypse.” 1. The Disney ending Precision strikes, IRGC bases gone, leadership collapses, democracy magically appears. Cute idea. Unfortunately, Iraq, Syria, Libya, and literally every Western intervention ever would like a word. 2. The Venezuela special The regime survives but gets bullied into behaving because the U.S. parks enough firepower around them to blot out the sun. They stay in power, but on a very short leash. 3. The “new boss, same as the old boss” coup The regime collapses… and the Revolutionary Guard just takes over. Protesters lose, Trump loses, everyone loses except the guys with the guns. 4. Iran hits back hard Missiles rain down on U.S. bases, Israel, Qatar, UAE, Saudi, take your pick. Air defenses stop most of them. “Most” is doing a lot of work here. 5. Hormuz becomes a booby trap Iran mines the Strait of Hormuz, 20% of global oil traffic chokes instantly. World economy: “sick, awesome, didn’t need stability anyway.” Iran: also hurts itself in the process, but desperation is a hell of a drug. 6. The nightmare scenario Iran swarms a U.S. ship with drones, torpedoes, speedboats, and actually sinks one. Total humiliation. Captured U.S. sailors paraded on Iranian TV = Carter 2.0 but with memes. 7. Full-blown national collapse Regime falls, and the country snaps along ethnic lines, Kurds, Baluchis, Azeris, everyone armed, everyone settling scores. Humanitarian crisis, refugees, chaos, perfect storm. This is the scenario keeping Saudi, Qatar, and UAE awake at night. Bottom line: Trump has marched everything right up to the edge. If he hits the button, no one, not even him, knows which version we get. Source: BBC

🚀🚨 : AFRIQUE EN MODE PUISSANCE. L’ingénieur et entrepreneur éthiopien #Bijay_Naiker va révolutionner l’espace africain : il construira une usine de fabrication de satellites au Burkina Faso, permettant au continent de produire ses propres #Satellites ! L’#Afrique reprend son destin technologique en main… la course spatiale ne fait que commencer.

ClawPhone is alive. I installed OpenClaw on a $25 phone and gave it full access to the hardware. I haven’t explored everything yet, but it’s a cool formfactor for agents to run on.

Des résultats « exceptionnels » : les puces Ryzen génèrent désormais des revenus colossaux pour AMD @Frandroid frandroid.com/produits-andro…

Introducing the Raspberry Pi AI HAT+ 2, priced at $130. This add-on board for Raspberry Pi 5, powered by the Hailo-10H AI accelerator and 8GB of on-board RAM, is built to run LLMs and VLMs locally, without cloud or network dependencies. The Raspberry Pi AI HAT+ 2 delivers 40 TOPS of generative AI performance and integrates directly with Raspberry Pi OS, including native support in rpicam-apps. raspberrypi.com/news/introduci…

We, over 300 academics, diplomats, journalists and concerned Swiss citizens are calling on our country’s top diplomat, @ignaziocassis and the Swiss Federal Department of Foreign Affairs @EDA_DFAE to intervene diplomatically toward the European Union for the sake of our compatriots, Col. Jacques Baud and Nathalie Yamb. Our open letter is now published here. neutralitystudies.com/wp-content/upl…

Everything is planned. Netanyahu’s plan was “7 wars in 5 years” Jeffery Sachs exposes Israel and America. next Venezuela & Iran . Netanyahu is our master.

Capitaine Ibrahim Traoré au sommet de l’AES Ma conviction, mon analyse, est claire : quelque chose se prépare en Afrique de l’Ouest. Je l’ai appelé l’Hiver noir. L’Hiver noir arrive. Ce sera un hiver très froid. Un hiver sanglant. Un hiver meurtrier. Vous me demanderez pourquoi.

⚠️ INTÉRESSANT 👍 GROK CHOISIT ASSELINEAU COMME LE PLUS APTE À REDRESSER LA FRANCE

n8n now offers certification courses! docs.n8n.io/courses/level-…

Les pays africains devraient se rapprocher du gouvernement américain pour demander que @DOGE leur fournisse le détail des montants et de tous les destinataires des financements USAID des 20 dernières années. Certains finiront par comprendre ce que l’on sait/dit depuis longtemps.

Comminuqué : cela fait 50 jours de détention arbitraire du président Aliou Bah.