Monthly Archives: October 2024

Get in, get out, get back to work: Wes Bos on the McMaster-Carr website

I was introduced to McMaster-Carr corporation early in my postdoctoral career, now well over thirty years ago.  The firm is renowned for stocking all sorts of parts and equipment and getting them to customers fast.  I went up one floor to the rapid orders room, checked the massive catalog, filled out a simple slip, and had electrical, mechanical, and/or vacuum parts the next day.  Sometimes they came the same day if I got the slip in early.  The paper catalog is not easy to get but I managed when I started my research position at UCLA after that postdoc.  The warehouse in Santa Fe Springs enabled California McMaster-Carr to meet the responsiveness of the New Jersey operation.

Fast forward to the era of e-commerce and the firm inspires affection even from the cynical, world-weary souls of HackerNews.  mcmaster.com is simple, well-organized, cruftless, and faster than any of its competitors in the Maintenance, Repair, and Overhaul market.  Consumer websites aren’t even in the discussion.  I’ve even ordered from them for my personal needs when the slightly higher prices are more than offset by the speed from order to delivery.  Youtuber Wes Bos dives into why the website is so fast, marvels at the developer team that enables it, and the management team that maintains focus on the customer’s desire to search, order, and get back to business.

Youtube Channel: Wes Bos

But to what end?: Branch Education on GPUs

Branch Education produces stunning videos comprehensively explaining aspects of modern technology for the interested layperson.  Whether it is a simple 3 way switch or a wafer fab, each video contains deep research, exquisite 3D animations, and honest narration.  It is one of the few sites I’ve whitelisted on SponsorBlock.  The recent release on GPUs is up to their usual Gold medal standard.  I knew the gizmos were powerful but I did not know that they were running at 36×1012 operations per second or that they had so many specialized subcores for raytracing and matrix processing.

Youtube Channel: Branch Education

These chips are based on years of R&D in physics, chemistry, materials science, chip design, fabrication engineering, algorithms, and 137 other developments to do it at a large scale to create reliable products at a price that at least some consumers can afford.  The performance gains are what basic researchers dream about when they invent some small piece of that puzzle with some vague hope of making the world a little better.

Where does it actually go?  Videogames, cryptocurrency, and now AI – whatever that is.

Youtube Channel: 3dfxhistory