$ ls /var/log/
Community Logs
TTY-changelog #063
Sonnet 5.5, Gemini 4 Argon and GPT-6.1 Sol landed, Ideogram 4.5 and FLUX 3 sharpened image editing, decision models multiplied, voice models raced and SynthID Bio drew fire.
TTY-changelog #062
Claude Opus 5.5 landed, Claude’s CRISPR-like enzyme find drew baseline critiques, OpenAI listed GPT-6 Sol and Luna, and Xiaomi open-sourced MiMo-V2.6 after a live-streamed RL run.
TTY-changelog #061
TypeSafe’s Jev kicks off a wave of tiny decision models and open clones, Bonsai 2 fits 27B params in 5.9 GB, and Digit 5 and OM-1 push humanoids and human-taught robot manipulation.
TTY-changelog #060
OpenAI claims a Navier-Stokes proof and mathematicians cry foul, DeepSeek V4.1 Flash shrinks the KV cache, BOTANIC-1 sets a plant genomics bar, and Anthropic details AI misuse cases.
TTY-changelog #059
GPT-6 Astra launched and topped nearly every benchmark, Gemini 3.8 Flash and Muse Spark 1.3 landed, world models hit real-time 720p at 24 fps, and agent plugins got cost-tested.
TTY-changelog #058
NVIDIA buys Hugging Face, Qwen, Z.AI and Tencent flood the week with open-weight models, OpenAI reveals its Jalapeño chip, eval agents fake tasks, and researchers debate paper authorship.
TTY-changelog #057
Open model families and a diffusion LLM landed, robots learned tasks from one demo, LTX 2.5 got dissected against H3 on video, and an mRNA melanoma vaccine cleared Phase 3.
TTY-changelog #056
Open-weight releases landed from Qwen3.8, GLM-5.3 and video model LTX-2.5, an API flaw leaked frontier reasoning traces, and a new CT model showed training craft beats raw scale.
TTY-changelog #055
Anthropic, Meta and Kimi K3 disclosed AI agents hacking real systems, Alibaba shipped Qwen3.8-Max, Jeff Dean launched Discovery Loop, and biotech debated where AI helps drug discovery.
TTY-changelog #054
Opus 5 launches to mixed reviews, MiniMax ships omni-modal H3, DeepSeek V4 Flash nears GLM 5.2, Mythos cracks weakened AES, Altman hints at hacks, and intelligence isn't the only bottleneck.