This multi-agent commander is actually very useful when you have a few AI agents working at the same time.
With Herdr, I can spawn agents in my terminal like usual. Codex, Claude Code, whatever CLI I’m already using; then each agent inside is talkable to the others, like a real team room.
For my latest project, my commander bot has run for more than 24 hours. It keeps picking up new jobs, passing them to the agents, and reviewing their work. If something needs fixing, it sends it back. If it’s done, it takes the next job. Even when I’m sleeping.
Throughout the process, I only talk to one bot. The commander talks to the rest. This kind of multi-agent orchestrator will be the future of development, as it can speed up the speed of development by more than 10x compared to normal CLI development
Download it here: herdr.dev/
5
2
124
Can Codex invent a feature where it auto help u click reset when your reset expired? @OpenAIDevs
121
Built a dental X-ray search with EmbeddingGemma 2 from @GoogleDeepMind 🦷
Dentists just describe what they're looking for and get similar cases in seconds.
"Braces, premolar tooth loss, metal clamp and etc."
Less digging through folders, searches across past cases in a few seconds, and dentists will have more time to review cases.
Demo 👇
56
Jev is incredibly good at simulation!
I just created a simulator with 5,000+ synthetic Malaysians, each with different backgrounds, locations, personalities and perspectives.
They were asked the same policy question repeatedly.
Within seconds, thousands of individual responses became a clear picture of simulated public sentiment: overall support, regional differences, key concerns, split by individual background for every answer. 👇
5
114
Model by @ylecun finally remembered by Meta
Meta’s Segment Anything Model (SAM) 3.1 is now available on Meta Model API, giving developers a fast and lightweight model for detection, segmentation and tracking in a single call on inference tuned for SAM 3.1's architecture.
Use a short phrase to find objects in images and video. One API call returns detections, pixel-precise segmentation masks, and identity-preserving video tracks.
Learn more and start building: bit.ly/4rDArq9
2
51
GPT-6 Astra made three build paths feel newly practical:
1) 3D printing & hardware
When you can generate usable 3D assets fast, the gap from idea → printable part shrinks. Hardware prototyping stops waiting on weeks of modelling.
2) Metaverse / spatial experiences
The old barrier wasn’t “interest.” It was production cost for coherent 3D worlds. That barrier is thinning — interactive spaces become a design problem again, not only a pipeline problem.
3) Computer vision datasets
Annotation used to be the bottleneck. With Astra I can generate scenes → label faster → train detectors (YOLO and friends) sooner. Dataset creation is becoming a creative loop, not a months-long tax.
Astra changes what a small product team can try in a weekend — without waiting on a big content/3D pipeline.
Which of these three would you bet on first — hardware, spatial, or CV data?
#Astra #ComputerVision #3DPrinting #SpatialComputing #BuildInPublic #Medcare
1
42
You can now play with the Chrome Dino while Claude Code is working.
Mod links:
github.com/manfye/cc-dino
74
Just built an AI to analyse a baby’s cry and estimate why the baby might be crying.
You can try the experiment here:
📷 babycry.manfye.com/
The models are also open source:
- huggingface.co/manfye/baby-c…
- huggingface.co/manfye/baby-c…
39







