Inside three.ws: The Open-Source Stack That Gives AI Agents a Body, a Brain, a Wallet, and a Job
three-ws
• • 42Community Blog & Articles
Community Articles
view all
Real-Time Intelligence with IBM Time Series Models on Confluent
1
September 2, 2026 
BenchMIRT: What are LLM benchmarks actually measuring?
8
September 1, 2026 
inferencejavascriptwebgpu
Introducing @huggingface/kernels: 200+ WebGPU Kernels for Local AI
29
September 1, 2026 
audiospeechleaderboard
The Open ASR Leaderboard Adds Its First Global South Language
- +6
52
August 28, 2026 
nlpguidecommunity
Training and Finetuning Multi-Vector Embedding Models with Sentence Transformers
109
August 26, 2026 
Granite 4.2 LLMs: How They're Built
95
August 25, 2026 
Quantization-Aware Healing: a compressed, 4-bit model that outperforms its full-precision original
44
August 25, 2026 
gradioworkflowstutorial
Wire It, Run It, Deploy It: AI Workflows in Gradio
37
August 25, 2026 
researchsearchembeddings
How Hugging Face Inference Endpoints, Jobs, and Buckets Power Search on Papers with Code
28
August 21, 2026 
audiospeechbenchmark
Measuring benchmark optimization in speech recognition
- +3
62
August 21, 2026 
Up to 3.2x Faster Inference with LFM2.5-DSpark
51
August 20, 2026 
How Much Memory Does Your Agent Actually Need?
76
August 18, 2026 
nlpguidecommunity
Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers
104
August 18, 2026 
Same Cluster, 33 Points More Utilization: What Changed Was the Order
17
August 17, 2026 NEW Articles from Team or Enterprise organizations will get promoted to the main section.
Inside three.ws: The Open-Source Stack That Gives AI Agents a Body, a Brain, a Wallet, and a Job
three-ws
• • 42VLANeXt: A Simple and Research-Oriented Codebase for Robotics Research
cavanloy
• • 19Technical writing in the agentic era
joelniklaus
• • 12Evaluating LLMs Under Production Parity: A Replay Pipeline for Safe Model Swapping in Conversational Agents
TechforHumans
• • 12Extremely Fast and Accurate Transcription with Granite Speech 5.0 Turbo CTC
ibm-granite
• • 31Modern LLMs: What the FLIP is an Engram!?
Blackroot
• • 12Uncensor any LLM with abliteration
mlabonne
• • 910Did the Civilization Emerge, or Was It Recited?
FINAL-Bench
• • 8KV Caching Explained: Optimizing Transformer Inference Efficiency
not-lain
• • 401The Best Open Source and Open-Weight LLM Models to Run Locally in 2026
daya-shankar
• • 20Deploy local agents everywhere with LFM2.5-2.6B
LiquidAI
• • 95We changed one line and the benchmark score moved 0.21 AUROC
FINAL-Bench
• • 14Intel GPU AI Skills
kushal-mittal
• • 4TAVR: Generate Your Talking Avatar from Video Reference
HeyGenAI
• • 4TransVLM: Detecting Any Shot Transition with Vision-Language Models
HeyGenAI
• • 4Frontier-Assisted Single-Prompt Disposable Risk Assessment
darkc0de
• • 4Puffin-World: Scaling a Unified Multimodal Model with Native 3D World States
KangLiao
• • 4Norm-Preserving Biprojected Abliteration
grimjim
• • 97EgoSuite-Open100K: 100,000 hours of egocentric human data for Physical AI
LightwheelAI
• • 3Stop Abusing Git for ML Data: The Architect’s Guide to AI Storage
prpatel
• • 3