Orcas are “exploding” fish for fun — scientists have recorded unusual behavior among the “killer whales” The animals accelerate to incredible speeds and ram ocean sunfish with such force that the fish literally burst apart and scatter into pieces. Biologists believe there are several possible explanations: it could be a brutal way of preparing prey so that younger orcas can feed more easily, or the whales may simply be playing and experimenting with their strength.

Growing numbers of orcas and whale sharks have started appearing off the coast of Mozambique, in an unusual development that’s coincided with the formation of the El Niño weather cycle.

Implemented batching support for footless-orcasaq-2-27B-nvidia-p100. Benchmarks show 40 tok/s at concurrency = 4 running OrcaSAQ-2-27B (@OrcaRouter) on a single NVIDIA Tesla P100. More updates to follow. • Repo: • Demo (pure prose):

Someone from the community just pushed OrcaSAQ-2 Cyber 27B to 219K context on 2× RTX 3060s. 68 tok/s. 15.7GB model. Correct long-context retrieval. His verdict: “the best local 27B result I have measured so far.” This is why we built SAQ-2: more intelligence per GB for everyone, on any hardware.

There are only ~75 southern resident orcas left in the Salish Sea, and hidden in the sounds beneath the water is information that could help protect them. Now, AI is helping conservationists listen at scale, identifying whales in real time and alerting mariners when they're nearby so they can slow down and steer clear.

Imagine bungee jumping above the ocean and seeing a pod of orcas below. Of course they let them pass, but how would you feel—cool or scary? I’m betting an adrenaline junkie would love it. Happy Wednesday, friends!

Alright, so I just got OrcaSAQ-2-27B (Qwen3.8-27B quantized by orcarouter) up and running on an Nvidia P100. Obviously, performance isn't the greatest, but I honestly didn't expect to run a modern-architecture model using trellis-based quantization on a 2016 GPU with no Tensor Cores that currently goes for around $100–$200 on the used market. Pretty incredible. Status: No MTP, fine-tuning required, PPL matches model documentation. - GPU: Nvidia P100 16GB - Model: OrcaSAQ-2-27B by @OrcaRouter - Engine: footless (In-house development, currently private) Kernels & runtime: Built with assistance from DeepSeek v4.1 Flash by @deepseek_ai .