I wonder how many guys who have have written or significantly maintained "household name" level FOSS products just earn a corporate sinecure somewhere as hypercompetent remote sysadmins or ICs or something. Folks who don't necessarily care to earn top dollar, with all the headaches that entails, but also almost never have to actually work more than 2 hours in a given day to keep the ship going, and the arrangement is just so cozy and gives them enough time to themselves to work on their actual passion that they accept the arrangement.
I know of at least one recruiter who does something like this and specializes in greybeard hiring, and it seems like a steady niche if you have the network to pull it off.
Yes. Absolutely. The biggest data point pushing the affirmative is less Anki itself but the success of products at the forefront of the second wave of spaced repetition apps [1] like Khan Academy. Duolingo, too, but Duolingo gets flak from people for being too Goodhearted by retention for its own good; Khan Academy actually does force feed you enough actual problems to learn some math.
Writing the cards is engaging with the cards for some small subset of the population. I am part of that audience. But most people are terrible at it, and it's not an easy skill to build.
Ther majority of people who are interested in Anki -- and the vast majority of normal human beings with nonzero willingness to pay, which is a very unique subset of the population with goals that tend to look like "Pass X exam by Y date so I can [get a job|earn my citizenship in a better country|...] -- just want good pedagogical material wrapped in some control harness so they can treat some fraction of their learning the same way they treat going to the gym. Show up, put in the reps, get results.
Oh dude, no. The instant I could start using something like https://www.wolframalpha.com/ in college over a decade ago, I did. I've never, not once in my life, had any desire to use a physical calculator when I could use a full blown computer instead. I think it's inferior even for pedagogy purposes - I can't screenshot and make Anki cards out of a TI-84.
I tried using my WolframAlpha in high school but I had no idea what I was doing and it felt way too complicated, but in college it is a game changer and I still use it to do school work. Totally agree!
>This is a practical policy allowing me to maintain my own professional standards and remain employable in a difficult economy.
I'm interested to hear more about the rationale behind the "remain employable" part of this line.
All things equal, we would normally expect someone deliberately saying they won't use a certain tool to perform a certain job as limiting their employment opportunities, not expanding it. The classic example is people who refuse to drive for work; there are good non-employment reasons for this (driving is the most dangerous thing many people do on a daily basis) but it's hard to argue that it doesn't restrict where one can work.
I think the most likely rationale is that the author thinks that posting a no-AI policy for professional work is itself seen as a signal of certain things about them, their skill level, etc., and that wins out for the kinds of clients they wish to take on. This doesn't have to be a long- or even medium-run bet to make, given that it's cheap to backtrack on such a policy down the line. Either way it's clear from reading the measured prose that there's an iceberg of thinking behind what's visible here and they are probably smarter than I am.
>if you’re an executive [...] you get the big bucks for a reason
In Finland? Notably wage-compressed Finland?
No comment on the specifics of this case, I agree with you that the executive should be where the buck stops. But you would be surprised how many various execs I have met here over the years who admit behind closed doors they really do treat it as a fancy job title that barely pays above their last position, but comes with 3x the stress, and they do it simply because, well, someone has to. You can't really be surprised that most of the folks here who you might want to be in the C-suite decide it's just not worth it, that remaining a middle manager or even an IC is simply a far better value proposition.
Posting anonymously here. I was on the leadership team of a Nordic public company, reporting to the CEO, presenting to the board and representing the company at the AGM. Total comp a little under $200k.
The compensation really didn’t match what you take on in terms of responsibility and legal liability. The stress was significant too. That said, as you point out, the work needs doing.
Recommended if you have an over-active sense of duty, not otherwise.
This is actually one area of research for me as a recently-minted DevSecOps engineer. Most cybersecurity attacks are relatively unsophisticated, and succeed by being scaled up so much they succeed on a few soft targets anyway - but a Ralph Wiggum loop on even a scaled up local edge model could make those kinds of techniques much, much more terrifyingly sophisticated for actors with that kind of hardware at their disposal. Abliterated models are of special interest here [1] because they make it even cheaper to do it at scale.
It almost feels like we are seeing the digital analogue of the evolution of the flighted stinging insect, i.e. mosquitoes, bees, etc. They don't have to be very smart individually to absolutely decimate the population of megafauna. A tiny bit of economic intelligence goes a really long way here.
Anecdata, but uv served as a very good packaging mechanism for a Python library I had to throw on an in extremis box, one that is not connected to the Internet in any way, and one where messing with the system Python was verboten and Docker was a four-letter word.
I keep toying with the idea of writing a cron that implements a Poisson process. Say I give it a parameter of 3600; our `pcron` would ensure the jobs occur randomly but average out to once per hour, making the timing of the next run independent of the last via the memoryless property of the exponential distribution.
The next sleep interval would be calculated probably as as t = -\lambda \ln(U) (where U is a uniform random variable). This way you ensure that the probability of the job firing in the next 10 seconds is the same whether the last job finished an hour ago or just five seconds ago. But \lambda remains the average amount of time between jobs.
It’s compelling to me because it solves thundering herd problems at the architectural level, and also because it simply seems like a lot of fun to have to code very defensively against such chaos. Switching back to a deterministic schedule after surviving such chaos probably leads to a much more robust system overall.
> making the timing of the next run independent of the last via the memoryless property of the exponential distribution
Nit: you’re not relying on the memoryless property here but just plain old independent sampling. You’re right that memorylessness means that the elapsed time since the last job provides no information on when the job fires next, but this is orthogonal to the independence of the sleep intervals.
I know of at least one recruiter who does something like this and specializes in greybeard hiring, and it seems like a steady niche if you have the network to pull it off.
reply