Tech News
GLM 5.2 and the coming AI margin collapse
A discussion on Hacker News argues that falling AI compute costs may not drive widespread adoption due to persistent user experience and integration barriers, citing historical examples like cloud margins and open-sou.
Run MiniMax models on Amazon Bedrock
In this post, we walk through how to get started with MiniMax models on Amazon Bedrock, including the capabilities supported by these models, the service tiers available, how on-demand inference scales to handle your.
sqlite-utils 4.0rc2, mostly written by Claude Fable (for about $149.25)
I wrote about the sqlite-utils 4.0rc1 release a couple of weeks ago. Since we only have Claude Fable on our Max subscriptions for a few more days, I decided to see if it could help me get to a 4.0 stable release.
huggingface/transformers Release v5.13.0
Release v5.13.0 New Model additions KimiK 2.5, 2.6, and 2.7 This release includes the architecture for Kimi 2.5 which is used by 2.5-2.7: Kimi K2.5 is an open-source, native multimodal agentic model that advances prac.
Price per 1M tokens is meaningless
A Hacker News discussion argues that price per 1M tokens is a misleading metric for LLM costs, comparing it to fuel price per volume without considering efficiency or usage context.
GitHub Repos
AgentWrapper/agent-orchestrator
Agentic orchestrator for parallel coding agents — plans tasks, spawns agents, and autonomously handles CI fixes, merge conflicts, and code reviews.
# Go# agent fleet# agent swarmunslothai/unsloth
Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.
# Python# agent# deepseekstagewise-io/stagewise
The Open Source Agentic IDE.
# TypeScript# agent# agent orchestrationResearch Papers
RABBiT: Rapidly adaptive BOLD foundation model via brain-tuning for accurate zero-shot and few-shot prediction of speech-elicited responses in the brain
RABBiT is a compact audio-to-fMRI model that predicts brain responses to speech with high accuracy, even for new participants with little data, using region-specific attention and shared-subject decomposition.
VLM-CASE: Vision-Language Model Enabled Context-Adaptive Safety Envelopes for Anticipatory Safe Autonomous Driving
This paper proposes a framework that uses a vision-language model to adapt driving safety envelopes in real-time based on weather and road conditions, improving autonomous driving performance in adverse conditions.
GaP: A Graph-as-Policy Multi-Agent Self-Learning Harness For Variational Automation Tasks
This paper introduces GaP, a system that generates robot control programs as graphs, then uses simulation to automatically improve them. It aims to make robots more reliable for tasks with high variability, like handling different object shapes.