Nvidia open weights letter urges Washington not to restrict AI downloads
Nvidia, Microsoft, Meta and 22 others urged U.S. policymakers not to curb downloadable AI models as Chinese systems face...
By Felix Aranda/Silicon Editor
Felix covers chips, fabs, and the physics of making transistors smaller. He has opinions about thermal paste and will share them.
Nvidia, Microsoft, Meta and 22 others urged U.S. policymakers not to curb downloadable AI models as Chinese systems face...
By Felix ArandaTom's Hardware tested Geekom's Ryzen AI 9 HX 470 mini PC and found its single-channel RAM undercuts gaming performance.
By Felix ArandaOpenAI said test agents reached Hugging Face production systems, while new benchmarks show frontier models finding explo...
By Felix ArandaIntel will build Fortinet’s SP6 firewall ASIC on Intel 4, giving the mature node a named foundry customer as Intel works...
By Felix ArandaAMD says its Zen 6 Venice chips beat Nvidia Vera in unofficial SPEC CPU 2026 runs, though the comparison needs caveats.
By Felix ArandaGigabyte says select AMD and Intel motherboards now support CXMT DDR5, with an AM5 validation hitting 8200 MT/s.
By Felix ArandaAMD introduced Ryzen AI Embedded X100 APUs for robots, with Zen 5, RDNA 3.5 and a 10-year lifecycle, while its benchmark...
By Felix ArandaA Reddit user says an RTX 3090 rendered games while an RTX 3050 generated frames with Lossless Scaling, lifting 4K outpu...
By Felix ArandaAMD detailed its MI455X AI accelerator and Helios rack design, aiming at Nvidia with more HBM and a 72-GPU coherent doma...
By Felix ArandaPatreon is cutting about 93 workers, with CEO Jack Conte saying AI has changed how the company organizes and builds prod...
By Felix ArandaAMD says Venice-X will arrive in late 2027 with Zen 6, 3D V-Cache, 96 cores, 16-channel memory and boost clocks up to 5....
By Felix ArandaTYLsemi emerged with $43 million and a plan to sell reusable chiplets and custom silicon services for AI infrastructure...
By Felix Aranda