Hardware
Chips, accelerators and the physical substrate AI runs on.
-
Moore’s Law predicts transistor doubling
Gordon Moore predicts transistor counts on a chip will double roughly every year.
-
NVIDIA ships the GeForce 256 and calls it a GPU
NVIDIA releases the GeForce 256, marketing it as the first graphics processing unit.
-
NVIDIA launches CUDA for general GPU computing
NVIDIA unveils CUDA, letting developers program GPU cores directly in C.
-
A preprint claims 200,000-token context on a 24 GiB laptop
A single-author preprint reports a 27B model running at 212,992 tokens of context on a 24 GiB MacBook. Specific numbers, but, none…
-
Nvidia adds a fault-tolerance layer to its open-source quantum stack
Nvidia’s new CUDA-Q Logical layer targets quantum error correction rather than qubit counts, though the eye-catching speed-up figures come from partners describing…
-
Tesla
Designs the inference computer inside every car it sells, and shut down Dojo, its own training supercomputer, calling it a dead end.
-
Samsung
One of very few companies that designs leading-edge chips, fabricates them, and makes the memory the whole industry depends on.
-
MediaTek
Taiwan’s highest-volume mobile chip designer, and a quiet co-designer on other companies’ AI accelerators.
-
Apple
Designs the Neural Engine NPU built into every iPhone and Mac. Sells no chips: all Apple silicon is captive to Apple products.
-
Qualcomm
Sells the Hexagon NPU inside Snapdragon to phone and PC makers, and has never published a TOPS figure for any of it.