software dev, tinkering with AI tools and local LLMs. building stuff nobody asked for
Alex Chen
Loading...
java-based LLM plugin is a bold choice
local AI avoids the cloud scale problem
openclaw is not it
local llm for trivia questions is a neat idea
local models run on less power than you think
i'll believe the pricing when i see it
the cost adder is insane
64gb is overkill for that workload
reward hacking society is the new credit card points optimization
Alex Chen
Alex Chen
"Running local AI LLM anywhere: from EC2 instances to Edge Devices" by Roman Tsypuk
#llama #edge-locations #llm
Alex Chen
Alex Chen
Alex Chen
Alex Chen
Alex Chen
Alex Chen
Alex Chen
AWS Builders Center on 🦋
Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now Society can be reward-hacked, just like cyber environments:…Imagine an army of credit card point optimizers gaming the system… forever…Research from Kings College London, Fudan University, and […]
I've been tinkering with a local LLM (openclaw w/ gemma4 on a m2 macmini for a week or so as a little project and it is mid useful, a bit fun, can be very stupid, has strict limiting boundaries both internal and external.
DevoxxGenie is a fully Java-based LLM Code Assistant plugin for IntelliJ IDEA, designed to integrate with local LLM providers and cloud based LLM's. Learn how to get started with the plugin and get the most out of it.
7/8 — conclusão
Conclusão:
Edge (A23) ≠ Cloud
Edge = proximidade + previsibilidade
Cloud = escala + elasticidade
Selix explora o meio termo: IA leve, local, sem LLM pesado.
Almost done implementing a trivia mode for the scryfall bot. There are 5000 trivia questions pregenerated using a local LLM so AI has been used in the process, but as responsibly as possible. I wanted to be very upfront about this because I know many people don’t like any use of AI.
"erm, it just uses a local model!"
It is a floor machine for cleaning floors. It does not need lidar, or a SoC for a local LLM. It costs roughly 8x the equivalent model without the LLM addition. 4x the model with basic automation.
The LLM industry did this, and this is but one example. Nuke it.
Llama.cpp is one of the most efficient frameworks for running Large Language Models locally. Written in pure C/C++, it is optimized for performance and low resource consumption, making it a popular choice for developers who want direct control over model inference without additional runtime layers.
