AI hardware startup Humane has given its users just ten (10!) days notice that their Pins will be disconnected. In a note to its customers, the company said AI Pins will “continue to function normally” until 12PM PT on February 28. On that date, users will lose access to essentially all of their device’s features, including but not limited to calling, messaging, AI queries and cloud access. The FAQ does note that you’ll still be able to check on your battery life, though.

Humane is encouraging its users to download any stored data before February 28, as it plans on permanently deleting “all remaining customer data” at the same time as switching its servers off.

[…]

Today’s discontinuation announcement was brought about by the acquisition of Humane by HP, which is buying the company’s intellectual property for $116 million but clearly has no interest in its current hardware business

[…]

Source: All of Humane’s AI pins will stop working in 10 days

Today, our mobile phones, computers, and GPS systems can give us very accurate time indications and positioning thanks to the over 400 atomic clocks worldwide. All sorts of clocks — be it mechanical, atomic or a smartwatch — are made of two parts: an oscillator and a counter. The oscillator provides a periodic variation of some known frequency over time while the counter counts the number of cycles of the oscillator. Atomic clocks count the oscillations of vibrating atoms that switch between two energy states with very precise frequency.

Most atomic clocks use microwave frequencies to induce these energy oscillations in atoms. In recent years, researchers in the field have explored the possibility of using laser instead to induce oscillations optically. Just like a ruler with a great number of ticks per centimeter, optical atomic clocks make it possible to divide a second into even more time fractions, resulting in thousands of times more accurate time and position indications.

“Today’s atomic clocks enable GPS systems with a positional accuracy of a few meters. With an optical atomic clock, you may achieve a precision of just a few centimeters.

[…]

The core of the new technology, described in a recently published research article in Nature Photonics, are small, chip-based devices called microcombs. Like the teeth of a comb, microcombs can generate a spectrum of evenly distributed light frequencies.

“This allows one of the comb frequencies to be locked to a laser frequency that is in turn locked to the atomic clock oscillation,” says Minghao Qi.

[…]

the minimal size of the microcomb makes it possible to shrink the atomic clock system significantly while maintaining its extraordinary precision,”

[…]

Another major obstacle has been achieving simultaneously the “self-reference” needed for the stability of the overall system and aligning the microcomb’s frequencies exactly with the atomic clock’s signals.

“It turns out that one microcomb is not sufficient, and we managed to solve the problem by pairing two microcombs, whose comb spacings, i.e. frequency interval between adjacent teeth, are close but with a small offset, e.g. 20 GHz. This 20 GHz offset frequency will serve as the clock signal that is electronically detectable. In this way, we could get the system to transfer the exact time signal from an atomic clock to a more accessible radio frequency, ”

[…]

“Photonic integration technology makes it possible to integrate the optical components of optical atomic clocks, such as frequency combs, atomic sources and lasers, on tiny photonic chips in micrometer to millimeter sizes, significantly reducing the size and weight of the system,” says Dr. Kaiyi Wu.

The innovation could pave the way for mass production, making optical atomic clocks more affordable and accessible for a range of applications in society and science. The system that is required to “count” the cycles of an optical frequency requires many components besides the microcombs, such as modulators, detectors and optical amplifiers. This study solves an important problem and shows a new architecture, but the next steps are to bring all the elements necessary to create a full system on a chip.

[…]

Source: Microcomb chips help pave the way for thousand times more accurate GPS systems | ScienceDaily

Earlier this week, Nvidia confirmed in its official forums that “32-bit CUDA applications are deprecated on GeForce RTX 50 series GPUS.” The company’s support page for its “Support plan for 32-bit CUDA” notes that some 32-bit capabilities were removed from CUDA 12.0 but does not mention PhysX. Effectively, the 50 series cards cannot run any game with PhysX as developers originally intended. That’s ironic, considering Nvidia originally pushed this tech back in the early 2010s to sell its GTX range of GPUs.

PhysX is a GPU-accelerated physics system that allows for more realistic physics simulations in games without putting pressure on the CPU. This included small particle effects like fog or smoke and cloth movement.

[…]

a game like Batman: Arkham City […] with an Nvidia RTX 5070 Ti, and when you try to enable hardware-accelerated physics in settings, you’ll receive a note reading, “Your hardware does not support Nvidia Hardware Accelerated PhysX. Performance will be reduced without dedicated hardware.” [ …] The in-game benchmark shows that with the hardware accelerated physics setting enabled on the RTX 5070 Ti, I saw a hit of 65 average FPS compared to the setting off, from 164 to 99. The difference in ambiance without the setting enabled is striking.

[…]

In other games, like Borderlands 2, it simply grays out the PhysX option in settings. As one Reddit user found, you can force it through editing the game files, but that will result in horrible framerate drops even when shooting a gun at a wall. It’s not what the game makers intended. If you want to play these older games in their prime, your best option is to plug a separate, older GeForce GPU into the system and run 32-bit PhysX games exclusively on that card.

[…]

we see Nvidia deprecating its own hardware capabilities, hurting games that are little more than a decade old

[…]

Source: Nvidia Drops Support for PhysX on Its RTX 50-Series Cards