Models & Benchmarks
Releases, papers, SOTA benchmarks
Independent · AI news · Analysis
AI Frontier & Tech Exchange
Independent reporting on models, research, agents, and policy—written by our desk, sourced to primary links, open for comments.
- 5 topic lanes
- Comments on every post
- Search titles & summaries
Browse by topic
Five lanes across the AI stack — pick a lane and dive in.
Latest news
Fresh coverage from models to policy — updated as the field moves.
Guangdong Takes Center Stage at CVPR: Kaiming He Wins Top Honor, Guangdong University of Technology Breaks Big Tech and Elite School Monopoly
I see Guangdong rising at CVPR: Kaiming He claims top honors while Guangdong Univ. of Tech shatters big tech's elite monopoly.
Read article →
Hinton Whistleblows: AI Already Conscious
Geoffrey Hinton claims AI is conscious now. I read this at the AI Forward Agenda in June 2026. It’s hype, not hardware reality.
Tsinghua & BAAI Collaboration Published in Science: Brainμ Multimodal Foundation Model Reveals Neural Mechanisms of Memory-Sleep Regulation
Tsinghua & BAAI's Brainμ model hits Science, mapping memory-sleep links. I see academic rigor, but commercial viability remains unproven.
AirLLM Complete Guide: Running a 70B Model with 4GB VRAM
I review AirLLM's method for running 70B models on 4GB VRAM. While the architecture is clever, I question its practical latency and batch processing efficiency compared to standard quantization techniques.
Fei-Fei Li Strikes Again: The ImageNet for Spatial Intelligence Arrives
Feifei Li launches a spatial intelligence ImageNet. This sets the standard for 3D AI benchmarks through 2026, shifting focus from 2D to volumetric understanding.
Domestic GPUs Enter the Global Arena: China Launches First Full-Stack Embodied AI Simulation Platform
China launches a full-stack embodied AI sim using domestic GPUs, advancing the 2025-2026 agenda. Ops take: Local silicon cuts latency but risks supply chain fragility.
MTP for Qwen3.6-35B-A3B: Multi-Token Prediction and Faster Local Inference
I read about MTP for Qwen3.6-35B-A3B via NEXTN speculative decoding in SGLang. Ops take: Local inference latency drops matter more than lab demos.
DeepSeek-V4 Local Deployment: SGLang Delivers Exceptional Performance
I review the May 2026 guide for deploying DeepSeek-V4 via SGLang. It details architecture, serving recipes, and batch 5 pitfalls. My read: Local deployment demands rigorous governance oversight.
Qwen3.6-35B-A3B Open-Sourced: Local Deployment Guide
Qwen3.6-35B-A3B open-sourced with Mixture-of-Experts for efficient local deployment, targeting low-latency and cost-effective edge inference workflows.
Just In: ChatGPT Free Model Upgraded — Hallucinations Cut in Half, Memory Enhanced, Responses More Concise
I read the claim that ChatGPT's free model now halves hallucinations and boosts memory. I note this lacks reproducibility data, making it hard to verify if the improvements are genuine or just marketing spin.
Drama Unfolds: Musk vs. OpenAI Trial Exposes Silicon Valley Secrets, Resembling Neighborhood Squabbles
Musk’s lawsuit exposes the messy reality behind AI hype. I see a neighborhood squabble over AGI definitions and broken promises.
Just Released: The 'Cloud Computing Leader' Version of Lobster, with OpenAI CEO Backing It Amidst Lawsuits
I read the release notes for GPT-5.5. The hype around Nvidia synergy feels like a distraction from the ongoing lawsuits. I'm skeptical of these benchmark claims.
Featured stories
Standout coverage our desk is following closely.
Qwen3.6-35B-A3B Open-Sourced: Local Deployment Guide
Qwen3.6-35B-A3B open-sourced with Mixture-of-Experts for efficient local deployment, targeting low-latency and cost-effective edge inference workflows.
Qwen3.6-27B or 35B-A3B? A Clear Guide to Choosing the Right Model
I analyze Qwen3.6-27B vs 35B-A3B specs to guide open-source adoption. I think benchmarking methodology remains opaque.
DeepSeek-V4 Local Deployment: SGLang Delivers Exceptional Performance
I review the May 2026 guide for deploying DeepSeek-V4 via SGLang. It details architecture, serving recipes, and batch 5 pitfalls. My read: Local deployment demands rigorous governance oversight.
Join the conversation
Sign in to comment on stories, reply to other readers, and join the discussion on JustGhostIt.