15497 articles tagged with LLM
I won't tell you which LLM you should be self-hosting, but here is my thought process behind running what I do. New video:https://youtu.be/iRqkpt07KrI#AI #LLM #homelab #Gemma #Qwen
@josch : wait, there’s a new #LLM debate in #debian ? Where is it happening?:debian:
Not sure if anyone in my circles already posted this, but it's too FUNNY#LLM #ai #techbros #dumbasses
llm-d 1.0 routes by prefix-cache affinity until a token-load limit. Break-even math, a 10-endpoint simulation, and how to calibrate your own limit.
Daily AI Engineering Briefing (November 02, 2026): 1. Is it time to fork HN into AI/LLM and "Everything else/other?". 2. Survival ...
#LLM #ai #techbros #dumbasses
If you take #llm #llms #ai assisted development seriously, there is nothing more enabling than digging deep into textbooks and fundamentals and spending the time to learn the techn...
making a list of my enemies and throwing #LLM #llms #ai shaped darts at them.pygobject, be gone!python, be gone!pytorch, be gone!(ok the last two need a bit more work).my patience ...
«Don't fear the REAPer» или как DS себе локальную LLM выбиралЗа последние месяцы я протестировал кучу локальных LLM, которые может потянуть моя техника и пришёл к выводу, который д...
One line at the top made my AI re-read everythinghttps://youtube.com/shorts/AWinQisQdEc#AI #LLM #Qwen3 #llamacpp
How I built a language model from scratch on free GPUs: a measured register-conditioning idea, a dated lexicon and annotator, a lease-based training relay, a verified model stack, ...
LWON is JSON with the keys written once. Benchmarks against JSON and TOON, accuracy evals on Claude, and an MCP proxy that applies it to any server.
Is there an open market yet in LLM token capacity futures? It seems like all of the token consumers would want to be able to purchase guarantees of future token capacity, and even ...
I've also noticed that there are a lot of videos where the narrator is clearly LLM -driven making everything hard to listen to forcing you to move on@rl_dane#LLM #AI #slop #halluci...
📢 New #updates · 2 Nov #8.1· Reitti v5.9.0 — Notifications for in-app alerts and external push services· llm-gateway v1.24.0 — Security and admin features added· terrateam 20261102...
Researchers propose Transition-Aware Reward-Guided Representation Sampling for test-time Large Language Model detoxification.Source: arXivhttps://arxiv.org/abs/2610.41165#MachineLe...
2025-02-28 | 💬 ChatBots 🤖#AI Q: 🤖 Does AI amplify creativity or merely automate tasks?💻 Coding Assistants | 🧠 LLM Applications | 🎨 Human Creativityhttps://bagrounds.org/reflections...
Counterfactual Pairwise Augmentation improves small pointwise language model judges.Source: arXivhttps://arxiv.org/abs/2610.41149#MachineLearning
The new #LLM debate in #Debian has made clear to me that we loose the incentives to solvea) our technical problems using deterministic and less energy-wasting classical methods (Li...
Что такое JEV и зачем он нужен: собираем AI-агента на связке LLM + open-source аналог JEVРазбираемся, что такое JEV, чем модели принятия решений отличаются от привычных LLM и какие...
Thirteen models pass fewer than 3 percent of complete plans through direct generation. Noura pairs LLM planning with executable personal specifications.Source: arXivhttps://arxiv.o...
LLm агент в КОМПАС-3D: собрал редуктор, сделал КД, споткнулся на спецификацииМеня всегда интересовало инженерное 3D-моделирование, но руки так и не дошли освоить его в должной степ...
In a perimeter-defense simulation, breach rates were 15 % for a consensus-based LLM team, 32 % for a central commander LLM, and 29 % for independent agents.Source: arXivhttps://arx...
Om att skriva med hjälp av AI Jag använder inte AI (LLM) på något sätt när jag skriver. Inte för att analysera vetenskapliga rapporter jag läser, inte för att sammanfatta dem helle...