Greetings Forum Community,
After several years participating in the forums, and seeing many of my favorite members posting in this thread, it seems the time to post a bio. Mine will be a bit longer, since I’ve lived many more years than most forum members.
I am an involuntarily retired curmudgeon, having worked in a dozen+ computer languages, and seen most of them supplanted by the “next great thing”. It always happens. I started out in high school with Fortran on punched cards, but quickly moved on to APL, timeshared on the IBM 360. In those days, computer pirates stole CPU time, not identities, because CPU time was strictly rationed and we geeks wanted to learn more and have fun.
I graduated from Harvard in Applied Math in 1977. At that time, research in AI was a career killer, but I always secretly kept an interest in it. The problem then and now is the lack of a massively parallel computational substrate. After graduation, it was a privilege to work with Bob Rodieck in his University of Washington lab, mapping retinal neurons and the LGN. I was a lowly lab tech, but eventually co-authored a paper with him. I would like to think that his discoveries about the vertebrate retina have informed the design of ML vision models.
After a period as a database consultant, I started a business that developed and sold FileMaker plug-ins. It had its heyday and put money in the bank. I added a master’s from Naropa University in Transpersonal Counseling Psychology in 2001, but never became a therapist. It’s hard enough keeping one’s own life together, let alone those of clients.
There are many life episodes that will never make it onto a resume. Fortunately, this introduction is not a job application. Meditation teacher, street musician, dance caller, cult member, daycare worker, ultramarathoner, stock trader, ex-Burner, SAT/AP tutor, psychonaut, cancer survivor.
Machine learning hit the public consciousness about six years ago. I was fortunate to find fast.ai and Jeremy’s clear lessons right away. ML may be a “next great thing”, but it is still the most interesting game in town, and it fulfills a life-long dream. My plan was eventually to learn the field and land a job. However, it seems the corporatocracy only wants to hire recent PhD graduates who will work 80 hours for 40 hours pay. C’est la vie and not for me.
I strive to be a helpful and welcoming presence on these forums, and refrain from giving advice. Posting has been a long journey of building confidence. Posting for the first time, in March 2017, almost induced a panic attack from the public exposure, but now I can hardly shut up. Mathy questions are a specialty, as are hard questions that do not get any response, even if I do not have a ready answer. It’s discouraging to post a sincere question and hear back only silence. I also enjoy hard “how to do it with PyTorch” problems. This may be a holdover from APL, where there were competitions to see who could perform a task with the least number of characters.
Looking forward especially to Jeremy’s lesson on Transformers. I understand the “what” (it’s just math), but not the “why”. Understanding “why” is essential for designing a model that fits the problem.
See you on the forums!
Malcolm
