tech:

taffy

Stanford researchers build diver robot

Researchers at Stanford University have built a humanoid robot that can function underwater. Named ‘OceanOne’, the robot was conceived and built from the ground up combining robotics, artificial intelligence and haptic feedback systems, says the university, to study coral reefs deep in the Red Sea, far below the comfortable range of human divers.

OceanOne looks something like a robo-mermaid. Roughly five feet long from end to end, its torso features a head with stereoscopic vision that shows the pilot exactly what the robot sees, and two fully articulated arms. The “tail” section houses batteries, computers and eight multi-directional thrusters.

Each fully articulated wrist of OceanOne is fitted with force sensors that relay haptic feedback to the pilot’s controls, so the human can feel whether the robot is grasping something firm and heavy, or light and delicate. (Eventually, each finger is expected to be covered with tactile sensors.) The ‘bot’s brain also reads the data and makes sure that its hands keep a firm grip on objects, but that they don’t damage things by squeezing too tightly.

Sensors throughout the robot gauge current and turbulence, automatically activating the thrusters to keep the robot in place. As the body moves, motors adjust the arms to keep its hands steady as it works. Navigation relies on perception of the environment, from both sensors and cameras, and these data run through smart algorithms that help OceanOne avoid collisions. If it senses that its thrusters won’t slow it down quickly enough, it can brace for impact with its arms.

In addition to Stanford, the development of the robot was supported by Meka Robotics and the King Abdullah University of Science and Technology (KAUST) in Saudi Arabia.

[Image credit: Frederic Osada & Teddy Seguin/DRASSM]

Just in

OpenAI launches ChatGPT Gov for U.S. government agencies — CNBC

It’s called ChatGPT Gov and was built specifically for U.S. government use; writes Hayden Field. 

DeepSeek’s popular AI app is explicitly sending US data to China — Wired

Users have already reported several examples of DeepSeek censoring content that is critical of China or its policies, writes Matt Burgess and Lily Hay Newman. 

DeepSeek hit with large-scale cyberattack, says it’s limiting registrations — CNBC

DeepSeek on Monday said it would temporarily limit user registrations “due to large-scale malicious attacks” on its services; writes Hayden Field. 

Viral AI company DeepSeek releases new image model family — TC

The models, which are available for download from the AI dev platform Hugging Face, are part of a new model family that DeepSeek is calling Janus-Pro, writes Kyle Wiggers. 

Eleos raises $60M

Eleos, a behavioral health AI company, has secured 120 million in a Series C funding round.