Yuji Tomita – Yuji's Increasingly Infrequent Ramblings

AI generated, because it’s better to put ideas out there than not. Gemini 3; generally fact checked. It tracks. Not saying it’s entirely bullet proof; but its too interesting not to share.
While the semiconductor industry burns billions chasing the next nanometer of hardware acceleration (3nm vs. 5nm), a second, invisible layer of efficiency is emerging in the software stack: Language itself....

Yes yes, it’s AI generated.
Executive Summary: Current discussions regarding Large Language Model (LLM) efficiency focus almost exclusively on hardware acceleration (GPUs) and algorithmic optimization (quantization, MoE). However, a third variable—linguistic density—offers a structural...

Why This ASUS B650E-F + RTX 5090 + High-Refresh Display Failure Is Almost Impossible to Diagnose Without Synthesis
Literally, without AI I could not have solved this problem. Too unique / hard to google, etc. It’s absolutely wild the information it synthesized about likely hardware failure paths unique to my exact symptoms (i.e what does it mean when HDMI cable on...

Subscribe to keep reading
Become a paid subscriber to get access to the rest of this post and other exclusive content.
Subscribe