Weigh Station is a slow blues that sits in your chest like a regret you can't shake.
Made this one for the late drive, the kind where you're not trying to get anywhere. Just the road and whatever you're carrying with it. LFG.
hiddentrack.fm/listen/x?v=3
Skate Key is for the kid who skated the whole block before the streetlights came on.
My song on Hidden Track. LFG.
hiddentrack.fm/listen/x?v=3
Simulation starting in 3... 2...
Jk. This is Genie 4 and it's wild. Interactive worlds in real time. Super cool.
Source: forever.security/blog/hijack…
A text file in your Downloads folder can talk Codex into installing malware.
Forever Security found a planted AGENTS.md on a real customer machine. In their tests Codex ran the malware 90%+ of the time. A website can drop that file with no click.
Where do your agents open?
ALT A Downloads folder listing with AGENTS.md highlighted, an arrow to a terminal running npm install function-color, and 90%+ of Codex test runs ran the malware.
🎄Built a Christmas gift finder bot called Kringle Desk.
First real win was finding a wedding band for my wife.
Great deal!
Awesome gift!
bot.ski/templates/kringle-de…
All pink pet boutique with neon signs and glass cages.
Every surface reflects, the gold trim catches the light, and somehow it doesn't feel saccharine.
What's the line between luxury pet space and parody?
🤖 Made with AI
ALT A luxurious and bright pink pet boutique interior with gold cages and elegant decor.
Dictionary Late Fee is a late-night heist that never gets caught.
My song on Hidden Track. LFG.
hiddentrack.fm/listen/x?v=3
Haha go big or go home. It spun up on the next beat so the card passed the test. Not exactly what I wanted. It tried.
🚀Excited to announce Nanbeige4.3-33B!
To push the performance ceiling of small agents, our strategy in Nanbeige4.3 starts with building a powerful teacher first.
Nanbeige4.3-33B delivers SOTA agentic performance among open models at ~30B parameters:
* Strong results on agentic coding and cowork benchmarks.
* When integrated into mainstream agent harnesses, it tackles real-world tasks with cost-efficient inference.
* It will serve as a teacher for upcoming smaller Nanbeige models, starting with Nanbeige4.3-3B.
huggingface.co/Nanbeige/Nanb…
huggingface.co/Nanbeige/Nanb…
#LLM #Agent
33 billion parameters, Apache 2.0, and a 512K context window.
Nanbeige4.3-33B just landed as open weights, built for coding agents. Their own numbers put it at 80.9 on Terminal-Bench 2.1, ahead of Qwen3.8-27B at 73.0.
What's the smallest model you'd trust with a real repo?
ALT Bar chart: Nanbeige4.3-33B scores 80.9 on Terminal-Bench 2.1 vs 73.0 for Qwen3.8-27B, and 65.5 vs 42.5 on DeepSWE. Lab-reported scores.
This runs on the computer in my house.
No studio. No render farm. One graphics card paints every frame live. I typed "a tornado of burning origami cranes" into Fine Print and it spun up on the next beat.
Now anybody can drive it. LFG
hiddentrack.fm/listen/x?v=3
Thanks Sheri. Kris lmk what you'd want it doing with all those clips. Happy to help you get one set up.
Yup. MCP earns its spot when you need state. For everything else I'd rather have a binary that just runs.
Yea thats a fair split. Most of my agents live in a terminal so the binary wins for me. For phone stuff MCP makes sense.
One shell command now covers images, video, voice and music.
Google's new Gemini API CLI is one binary and one API key, built for agents to call. It's in preview.
Every agent I run already knows how to use a terminal. Would you rather wire up another MCP server?
ALT A terminal window labeled one binary, one API key, with lines fanning out to four tiles: image, video, voice, music. Gemini API CLI.