AI practitioner specializing in large-model evaluation and on-premise deployment, agents, AI programming, Vibe Coding, general AI, and broader tech trends, with daily original technical articles.
OpenAI's GPT-5.6 arrives with three tiered models (Sol, Terra, Luna), new max and ultra reasoning modes, benchmark breakthroughs in programming, biology, and security, extensive multi‑layer safety guards, a steep pricing structure, and a tightly controlled preview rollout.
Qwythos‑9B, a Qwen3.5‑9B model fine‑tuned with over 500 M Claude‑style tokens, offers a 1 M‑token YaRN context, native function calling and tool‑augmented self‑correction, outperforms its base on MMLU and gsm8k benchmarks, and provides GGUF quantizations for consumer‑grade GPU deployment.
The article analyzes Tencent’s dual‑track AI Agent strategy, detailing how the consumer‑focused WorkBuddy leverages the company’s ecosystem while the newly acquired AiPy from security firm Zhidao Chuangyu targets enterprise and government markets with on‑premise, code‑as‑agent technology, and evaluates the competitive landscape and future prospects.
OpenAI has partnered with Broadcom to launch Jalapeño, a purpose‑built AI inference ASIC designed in nine months, promising superior performance‑per‑watt, integrated networking, and a full‑stack AI hardware‑software optimization cycle that could lower inference costs and reshape future data‑center deployments.
The author open‑sources a universal video‑download Skill and then introduces a companion Skill that automatically extracts audio, frames, and visual insights from a local MP4, runs Whisper and qwen3.7‑plus to generate a structured summary webpage with player, key points, timeline and actionable items.
The author presents an open‑source Skill that consolidates video downloading into a single command, leveraging yt‑dlp to handle various platforms, offering safe default strategies, detailed failure reports, multi‑format support, and a downstream AI pipeline for transcript and webpage generation.
The author demonstrates how OpenAI Codex can serve as a full‑featured computer manager on macOS, automating cache cleaning, software uninstall, startup service control, large‑file detection, browser data cleanup, material organization, and daily inspections through tailored prompts and screenshots.
On Father’s Day the author built an AI Agent skill called z‑father‑concept that, given any term, traces its lineage through concrete ancestors, functional roles, societal issues and finally a philosophical theme, illustrating the process with examples from fans to loneliness.
The author walks through a complete Qwen3.7-Max workflow—knowledge collection, cross‑document understanding, article creation, and video production—highlighting tips, challenges, and the model's ability to handle 1 M‑token contexts and long‑chain tasks.
The author reproduces Anthropic’s Fable 5 using Gemma‑4‑12B‑v2, showing a 3.5× improvement on the telecom tau2‑bench versus the base model, details the agentic, coding, and general training data, compares quantization sizes, provides llama.cpp launch commands, and notes speed gains from speculative MTP decoding and current limitations.
