/Silicon Editor
Felix covers chips, fabs, and the physics of making transistors smaller. He has opinions about thermal paste and will share them.
Google’s new model keeps 3.7 Flash token rates through 2026, yet extra reasoning can increase task costs before rates do...
By Felix Aranda
Jensen Huang said Nvidia has reached AGI “for many tasks,” but offered no benchmark and called AGI milestones senseless.
By Felix Aranda
OpenAI president Greg Brockman now oversees product and infrastructure operations, while Sam Altman remains chief execut...
By Felix Aranda
SK hynix says hybrid bonding is not expected for HBM4E, as thinner dies and tighter gaps complicate taller AI-memory sta...
By Felix Aranda
Intel detailed Xeon 7 Diamond Rapids at Hot Chips, outlining a 256-core ceiling, 16 memory channels and a new chiplet fa...
By Felix Aranda
Thermalright’s magnetic 3.69-inch fan display targets standard 120 mm mounts, but screw layout and USB capacity decide w...
By Felix Aranda
OpenAI’s opt-in ChatGPT Computer History turns permitted Mac activity into a searchable work timeline, with limits that...
By Felix Aranda
Photon Matrix lists its mosquito-targeting laser for order from $988, but independent performance tests and named safety...
By Felix ArandaWafer size means diameter: a guide to inch names, 200 mm and 300 mm production, and why area is not chip yield.
By Felix Aranda
Supermicro showed its 160-bay NVMe server at FMS 2026, with 164 listed drive bays and a Linux drive-ordering catch for o...
By Felix Aranda
Samsung’s 32-inch G80HS pairs a sharp 6K work display with 330Hz 3K gaming, but native 6K remains punishing for current...
By Felix Aranda
Intel executive Robert Hallock says demand for DDR4-capable LGA 1700 chips caught the company off guard, with supply and...
By Felix Aranda