NEAR AI Confidential Inference Goes Live on Bittensor’s SayGm Gateway
NEAR AI Cloud’s “confidential inference” is now accessible through SayGm, a Bittensor Subnet 28 gateway designed to keep routing and model execution inside hardware-isolated environments.
Related
SayGM (SN 28): Your Private Gateway to Frontier Models on Bittensor
Bittensor's gm (SN28) routes your Claude, GPT, and Gemini prompts through an Intel TDX enclave. No API middleman reads your data. Here's how it works.
Latest
Reflection Launches Beam, a 501B Open-Weight Model for Coding and Agents
Reflection says Beam pairs 501B total parameters with 23B active parameters, with Apache 2.0 weights planned after final red-teaming.
RobCo Becomes a Unicorn as Employee Share Sale Values Industrial Robotics Startup Above $1B
The Munich-founded company doubled its valuation in nine months and set a March 2027 commercial launch for Alfie, its autonomous industrial robot.
Meta Opens Muse Gadgets With Open-Source Hardware SDKs and Muse Home Link
Meta is opening Muse to DIY hardware with open-source ESP32 and Linux SDKs, while shipping a limited USB-C Home Link dongle for U.S. subscribers.
Tavus Introduces Griffin Human Interaction Model After 48% of Testers Mistook It for Human
The research preview combines real-time video, speech, and perception while Tavus works on safety controls before a wider release.