Today, @IBM and @Cisco announced plans to build a network of large-scale, fault-tolerant quantum computers — a major step toward distributed quantum computing and the foundation for a future quantum internet.
We’re combining IBM’s quantum hardware + software with Cisco’s leadership in networking to tackle the challenge of scaling beyond a single large-scale FTQC. This builds on our commitment to deliver Starling in 2029 and scale to Blue Jay in 2033.
News: https://t.co/9TDcl2lHSd
Officially launched the RISC-V Keccak instruction "internal review" today.. #riscv
This is one of the main ISA Standardization milestones we have. https://t.co/Xt21nuB28s has pointers to qemu patches, PoC HW implementation etc. Direct spec PDF: https://t.co/uSnvVimDv2
If a CPU cycle (0.3 ns) felt like 1 second, a 300 ms request would feel like 32 years.
I’m trying to say that 300 ms is an eternity for a computer. Software should be way better than what we just accept these days.
Architektura Apple Silicon okazała się strzałem w dziesiątkę pod LLMy. Magia połączenia szybkiej pamięci, przepustowości i zunifikowanej architektury. Mac Studio M5 Ultra (512 GB, 1.2 TB/s) mieści pełnego GLM-5.3-Flash (320B) w FP8 i wyciąga ~60 t/s offline, na biurku. Alternatywa na PC? 10x RTX 5090 za ponad $20 000 i własna mini elektrownia.
My dad has been working at TSMC as an engineer for the last 20 years, and he’s been in semiconductors as a whole for the past 30. A few days ago, he gave me one of the textbooks he’s referenced from when he was working at SK Hynix as a design engineer in the late 90’s.
The book (published in 1997) everything to do with CMOS circuit design and simulation, and what I find interesting is that the fundamental math and physics behind it has not changed in the last 30 years. What really has changed, is the simulation technology behind circuit verification that has evolved in the last 50 years, all the way from when SPICE v1 was released in the 70s, to whatever we have now.
It really does feel cool to be working on the next iteration of a technology that has been around for so long, yet is so difficult to use and develop even now.
I just think semiconductors and simulation is so interesting. I’m really happy that I get to work on this full time with my own startup, with others who also believe in the vision.
.@UMassAmherst researchers slashed AI computing needs by 90% using brain-inspired memory hardware to process language right on your device.
https://t.co/Lw9CPioQUz