I haven't blogged in a while. Given things in the news like generative AI you'd think I'd dive into thoughts there, especially interesting aspects like the use of probability, with my usual meanderings to related topics like Terrence Fine, Cornell, and formalists versus frequentists, etc. Or posts like https://www.cisa.gov/news-events/news/call-action-bolster-uefi-cybersecurity-now where I could possibly opine for another 900pp+ (like https://link.springer.com/book/10.1007/978-1-4842-6106-4).
Instead I was reminded of late 90's when reading https://www.quantamagazine.org/complexity-theorys-50-year-journey-to-the-limits-of-knowledge-20230817/#, specifically my time of the University of Washington (UW) in the UW PMP (Professional Masters Program) https://www.cs.washington.edu/academics/pmp. The PMP started in 1997 and I was accepted prior to the 2nd quarter having commenced, or the '2nd wave of inductees'). This was fortuitous for many reasons. One, Intel was the largest recruiter from UW at the time, including donating a lot of equipment, and two, the other regional software companies hadn't flooded the application pool with talent form Microsoft, UW, Boeing, etc. If I had waited a few more quarters, the probability of acceptance as an 5-year-in-industy-EE-undergrad may have been tough going.
Also, I had my first daughter during that last quarter. New tech job + around-the-clock-diaper-change + after-work-hours-studying-and-projects nearly killed me.
My 6-month-old little Husky didn't look so happy in her purple onesie on that hot day back in 1999.
Fast forward to 2023
My 6-month-old turned out to eschew the UW path, but her younger sister ended up choosing UW and is in her 3rd year
and my 2023 Husky next to her older sister
and a couple of my favorite things my UW daughter gave me include a box years ago
and then this coin when she entered UW
As part of that program, the computer science department hosted a welcome event at the Pacific Science Center in downtown Seattle next to the Space Needle. There were many UW professors in attendance, included a newly minted graduate from Caltech named Chis Diorio who had worked with Carver Mead on hardware neural nets. Part of his research entailed work on multi-level cell storage, and Intel had a significant history and products around NOR storage at the time, so Chris asked if I could connect him with any other Intel folks. Diorio has done pretty well in the intervening years https://www.impinj.com/library/blog/impinj-named-a-finalist-in-two-categories-2023-geekwire-awards with Impinj, too. My learning from this interaction is as follows: always assess if the value others see in the conversation with you are because of the 'person' or the 'platform'. In this case I suspect it was the 'platform', namely my employer.
Speaking of Carver Mead, it's interesting to see the wisdom of pioneers potentially resurface. Namely, reading https://www.techradar.com/pro/nvidia-beware-ibm-has-a-new-analog-ai-chip-that-could-give-the-h100-a-run-for-its-money made me think of https://www.amazon.com/gp/product/0201059924/. This is definitely a contrast to the digital state-of-the-art sort of read I found in https://www.amazon.com/Efficient-Processing-Deep-Neural-Networks/dp/168173835X, for example.
So after a couple paragraphs, how does this relate to the Quantum magazine article? Well, another professor was Richard Karp, viz., " In addition, the American complexity theorist Richard Karp proved that the universality property identified by Cook (and Levin, though Karp and Cook didn’t know of Levin’s work until years later) was itself all but universal". More details on Karp can be found in https://en.wikipedia.org/wiki/Richard_M._Karp, including "Apart from a 4-year period as a professor at the University of Washington, he has remained at Berkeley. From 1988 to 1995 and 1999 to the present he has also been a research scientist at the International Computer Science Institute in Berkeley, where he currently leads the Algorithms Group." And what was my interaction with this esteemed Turing Award winner? After Karp introduced himself I replied, "ah, I guess you get to work on things like compilers and stuff." My learning from this interaction 20 years later is as follows: always do your homework on the group with which you are interacting ahead of time.
The late 1990's had other great interactions. Having come up to Seattle from Houston, Texas, I couldn't but help to name my proposed Test Executive, or test framework for firmware, 'TEX", when working with Eric Hudnell. Or the spirited discussions with Andrew Fish in 1997 about ASIM (Advanced Server Information Modules), or a way to decompose our server firmware prior to the next of many, many 'modularization' initiatives, such as Intel Boot Initiative (IBI) and Davinci/Kittyhawk BIOS in 1998, EFI in 1999 Framework in 2000, ...... My retrospective learning: infrastructure evolution takes time. There are few 'short wins.'
A lot of those interactions above also entailed wisdom from others. I still recall a director who told me about the plan/clue matrix for engineers. Namely, you aspire to working with someone who has a 'Plan + Clue.' They are effective and directed. The 'Clue but no plan' are often visionaries/strategists in that they know where to go but not how. The 'Plan but no clue' are the scariest since they are akin to a bull in a china shop- they have a lot of kinetic energy to expel but perform the drunkard's random walk in plotting a path. Finally, the easiest to handle are the 'No plan and no clue' since they are largely inert/ignorable. Cook reminded me of this early 2000 wisdom in his posting https://www.johndcook.com/blog/2010/12/27/dumb-and-gets-things-done/.
Speaking of wisdom, I was asked to give some patent training last week. One of the pieces of wisdom I always offer to folks on patents is to have a sense of urgency, especially with the normalization of invention primacy with 'first to file' and the American Inventors Act a few years back. I often put this urgency in the context of the industry, namely invoking Joy's Law "no matter who you are, most of the smartest people work for someone else,” https://en.wikipedia.org/wiki/Joy%27s_law_(management). What that means for inventors is that given similar problem statements and the frontiers of knowledge, others may come up with similar ideas. But what you have within a large company is perhaps earlier access to the problem and/or the frontier of knowledge since Fortune 100's like Intel often 'create' said knowledge.
I also try to temper the patent presentation to put patenting in context. First the positive where I mention the ability to get patents as a software person in a hardware company (data also mentioned a year ago in http://vzimmer.blogspot.com/2022/09/new-milestones.html), viz.,
but then I remind the audience that patents are part of delivering value, not a means in-and-of-themselves (including quoting this blog from 10 years back http://vzimmer.blogspot.com/2013/12/invention-and-innovation.html), viz.,
From the patent milestones you see the title "Fellow." This is one of the highest designations of a technical-track employee, as distinct from management-track, employee at Intel. Many tech companies have similar designations, and I culled a comparison of the Seattle area shops below https://www.levels.fyi/?compare=Amazon,Intel,Google,Microsoft&track=Software%20Engineer
NB
Someone might question why I continually show an image from a website that others can find themselves. I guess I have two answers. One, I learned that content on the web ages out or evolves, such as top inventor Wikipedia started to subset its listing. And two, for better or worse a colleague introduced me to the MS Windows 'snippet' tool and the barrier to image creation was significantly lowered for me :)
Speaking of old blogs in 2013, I should probably reprise http://vzimmer.blogspot.com/2013/03/a-technical-career-path.html since I discovered the T-shaped model https://collegeinfogeek.com/become-t-shaped-person/ and have used it a lot since in mentoring people. Then the other part of me notes that I also usually advise people that 3-7 years is optimal occupancy time in a grade level, say 9, prior to becoming a 10. Too short a time and you don't have time to generate sufficient evidence to prove you delivered at that last level, and too long you become stale and people ask 'why now?' As I roll into year 10 as a Senior Principal Engineer or grade 11, I am in the 'why now' camp where people can dismiss earlier accomplishments in the decade as not fresh or of insufficient merit since if they were, I would have been acknowledged then. As such, I challenge myself as to whether I am the right person to be dispensing 'career growth wisdom' if my career isn't growing.
There are many great engineers in the Sr. PE community, including the two I mentioned in the last blog who retired, Sham and Kirk. And as the company matures, this cohort continues to grow in cardinality. But it does make me wonder if it's a terminal level, or sort of a tech elephant's graveyard
I couldn't help but use that metaphor since found the image arresting from reading Tarzan? or some such as a youth.
More likely it's the other career advice I give folks, namely that getting promoted is a lot like real-estate opportunities, namely 'timing and location.' Since promotions are a meld of business impact, technical acumen, and leadership, not all role or organizations afford opportunities for the same, or that triple may not be as business relevant as the market moves or the exigencies of the employer change.
I like to stand outside of the phenomena and observe it sometimes. Thanks Marcus Aurelius and Seneca, too, for the classical doses of introspection and stoicism, respectively.
Maybe part of it is the reality expressed by Corey Quinn
https://twitter.com › QuinnyPig
From what I can see, getting promoted at a big tech company is roughly four times
harder than getting that job at another company.
(I apologize for the rough copy-paste of a Google-search-of-a-Tweet). The gist of the sentiment above is that the new company wants to believe in you and leans in, whereas your present employer knows all of the defects of an internal candidate and perhaps subconsciously over-indexes on them. Another area where I have seen Quinn's assertion in practice is that if things are not going well in a company, outsiders from successful companies are seen as more attractive since the internal candidates are the ones who 'contributed to the company and its ills' (i.e., why promote the arsonists?); of course this confuses correlation with causation on both internal and external candidates to a first order since rarely is a single individual responsible for the ills or gains of any large company. Finally, one reason that Quinn's logic holds is that as the business evolves, it is necessary to important new skills for the new business domains, as distinct for taking the time or risk of 'up-skilling' the present bench; I still recall in my early days why hiring was prolific that the 2nd line manager told us 'we need the exact skills. There is no slack in the schedule to train someone up over three months.'
Or maybe it's related to the business cycle? Just as Ben Horowitz opined about 'wartime versus peacetime' CEO's https://www.linkedin.com/pulse/wartime-vs-peacetime-ceo-insights-from-hard-thing-things-pandey/, maybe there's a similar dynamic that applies down the stack, including technologists? The market dynamic might argue for loading the bench with more 'wartime'-minded folks? Or does the company need more 'referees' versus 'risk takers?'
So the magic bullet is job flipping employers then? Maybe, or maybe not. I've always found the promotion-through-job-change risky in that the grass isn't always greener. You'll see from zoom in of above
that my present job code overlaps with Microsoft 'Partner,' for example, I recall one Partner Architect at Microsoft, now retired, telling me that at MS the title 'Partner' is also code for "throw the other Partner under the bus" job description. Namely, it's an understandably competitive peer group when total compensations can easily exceed $1million/year. Similar for my slight overlap with Amazon Sr. PE. I recall asking an Amazon executive what happened when someone was hired at too high a grade level. His glib reply was 'Oh, that? We take care of those situations quickly.' Woof.
So what's my parting advice on this blog postings sidetrack on careers? Enjoy the trip. Add business value. Learn. Appreciate contributing to and receiving wisdom from others. And promotions are a Faustian deal, too. As Goethe expressed in this epic, be careful what you wish for. namely the pursuit of power can entail a deal with the devil, etc. The $1million/year compensation isn't because the employer admires your bio and portrait as a 'new partner.' Instead, it represents an expectation of output against both business goals and your new peers commensurate with this compensation. And these days it's delivering value on 'internet time.'
I've never worked in the high-flying world of silicon valley or startups either, but I suspect that there's no 'rest and vest' on a California tech building roof like https://silicon-valley.fandom.com/wiki/Big_Head in almost all cases. But if you enjoy the trip, moving up the tech job ladder can be awesome. You will be at a technical level aligned with Vice-President+'s of the management track and the role should provide more of an opportunity to make a dent in the universe as a technologist as you advocate for and drive insanely cool technical progress.
No comments:
Post a Comment