Tangential to the content of the blog, I’m always weary of anything being labeled “anonymous” if it contains a lot of text that you personally wrote. Stylometry is a thing and it apparently works (I do need a citation, but on mobile and don’t have time - a quick Google search will return some results). I can see a future or even present where all it takes is a snippet of text that somebody knows is yours, and then a model to provide the probability that any given text off the internet (comments, blogs, etc) is written by you. People write and speak in very unique ways. As you increase the amount of your text, it becomes even more unique, like a blog.
I had bad insomnia during graduate school, tried just about everything under the sun (medication, sleep studies, lifestyle changes, herbs) and eventually did CBT-I with a therapist, and that was my ultimate solution. She had me start with going to bed at 3am and waking up at 7am for a week. I was absolutely exhausted, felt like my days were dreams. Then, we pulled forward he’s time to 2am. Checked how long it took me to fall asleep, and if I reported it was quick, then pull it to 1am, etc. until I was falling asleep quickly but also felt refreshed. I kept this all down on some worksheets provided.
The key was absolutely no matter what, no matter how tired I felt, I had to wake up at 7am and get out of bed. No naps.
We found the optimal spot to be 11pm lay in bed, and wake up at 7am. I always keep that schedule, deviating only by 30min or so even on weekends. Haven’t had trouble sleeping since!
I also think going from graduate school (little solid schedule) to a corporate 9-5 job helped. Covid has mangled that a bit, but I still keep the same sleep schedule.
CBT-i more or less fixed me too. It should be noted that sleep restriction (the practice you're describing here) is only part of CBT-i but it seems to be the most important.
I also gave up coffee and I stop looking at screens 30-60 mins before my bed time. Before the therapy ~30% of my nights were sleepless.
Really good point - there was a bunch of other restrictions/lifestyle changes that probably helped contribute. I think the main one was going from two cups of coffee (morning, and then afternoon) to one cup (morning) really helped me.
Also the recommendation that if I couldn’t fall asleep within 15-ish minutes, to get out of bed and do something until I’m tired again instead of lay there and “try” to sleep.
Yes the 15-20 min rule helps a lot! I tried a morning coffee again recently (I miss it) but it messed up my sleep that night. But that could be due to my mind playing its tricks on me rather than the caffeine remaining in my system.
Some of these salaries are hilarious. I work in a small city in the Southeast at about 250k cash comp and feel blessed because of it. Leave it to HN to make me feel that tinge of envy. The weird thing is, I never get jealous/envious of someone my age making 1 million. 2 million. Or hitting the jackpot and selling their company for hundreds of millions. Doesn’t give me one iota of “wish I had that”, but the 400k starting salaries really make me wonder if I should be starting the job search again and uplifting the family out of the Southeast lol. I wonder if it’s because I can envision myself working for those companies and actually making it (versus founding a company and making millions). The mind is an interesting thing. I am very grateful to be making what I do, and even more grateful that people who practice my trade (in some form or fashion) are getting paid well for it!
I store moderately sensitive files in VeraCrypt containers. Are there post-quantum encryption algorithms on the roadmap for this type of storage software?
I get the same weird anxiety reading these threads. I was making 50k base as a post-doc, anxious about ex-academic friends making 80k in the "real world". Then I got a corporate job making ~$120k, and felt anxious that I wasn't making the ~$150k some of my bosses and friends were making. Switched jobs, making ~$200k in a small southern city, and I come into threads like this ("$500k comp at FAANG) feeling anxious that I'm not making enough in my mid-thirties and don't have enough saved up for retirement.
The crazy thing is, going from 50k to 200k, I can't honestly pinpoint whether I'm signficantly happier or not. So I try to not get envy about even the more ridiculous salaries. How does one get off the hamster wheel and feel content enough with salary to just live?
~200k in a small southern city if you adjust for cost of living is about equivalent to $500k at a FAANG in SF/NYC. When you know people who brag about how cheap their $2000 a month tiny studio in an old poorly maintained building in midtown Manhattan you'd understand you have it pretty good making a salary like that in a place you actually have the freedom to move around without running into someone and easily afford a really nice house & property.
Recognize that getting more salary is a fun goal to strive for, that might even get you some more options, but that you’ll be fine forever at your current salary regardless.
I guess I’ve made it one of my goals to strive for more, but If I’d end my career at my current salary that would still be a solid endgame.
I posted this on the other discussion related to this, but would love to get feedback here.
-----------------
I can understand this position, but I'd be curious what your thoughts are on how to best (I realize there is no perfect) keep your data private from snooping employees, hackers, or law enforcement.
I've thought about this over and over, and it's hard to come to a solid conclusion about keeping personal data safe (in this context I mean emails and files you may store in the cloud, not browsing history, social media posts, etc.). There are so many options with downfalls for each, and I'm not a security expert. So every time I get excited about trying a new service geared towards privacy, or setting up my own instances, inevitably somebody points out the terrible pitfall in it and I get discouraged.
1. Don't use the internet or internet services, period. <- Not tenable for most of us.
2. Use services who market themselves as geared towards privacy. <- Can't actually trust those services, even with E2E encryption because they could be running different code from what you think they're running.
3. Use regular cloud options, but stack stuff on top - VeraCrypt volumes or Cryptomator with Google drive, GPG for email, etc. <- Really difficult to setup and have a nice reliable way of accessing data on mobile/desktop/etc. No security audits on a lot of the open source software.
4. Host your own services - i.e. a Nextcloud 14 instance on EC2 with an S3 backend, then use client-side E2E <- Difficult to make sure you set the service up in a safe way, and not even a fraction of as much resources in auditing code as, say, a giant corporation.
5. Spread what you do out over multiple services - FastMail for email, DropBox for cloud storage, Standard Notes for notes, etc. <- A real pain.
I know there will never be a consensus on this, but I'd love to hear what your thoughts are on the best way to keep my personal files and notes personal to me. Let's assume I'm not a target of any spy agencies or whatnot, but I want to make it very, very difficult for anyone to read my person notes and files but me.
Encryption, to be done right, always has to ultimately be the end user's responsibility. So there is one solid conclusion but it requires work and discipline and is therefore unpopular.
Encrypt before you send/upload, decrypt after you receive/download. If you transmit or receive unencrypted data you are placing trust in someone else and there is no way to really avoid that unless you created and control the chain end-to-end.
So you can use Google Drive, etc., just encrypt your data before uploading if it needs to be secured. A 7-zip file protected with AES256 is decently convenient (from a PC) and secure.
I can understand this position, but I'd be curious what your thoughts are on how to best (I realize there is no perfect) keep your data private from snooping employees, hackers, or law enforcement.
I've thought about this over and over, and it's hard to come to a solid conclusion about keeping personal data safe (in this context I mean emails and files you may store in the cloud, not browsing history, social media posts, etc.). There are so many options with downfalls for each, and I'm not a security expert. So every time I get excited about trying a new service geared towards privacy, or setting up my own instances, inevitably somebody points out the terrible pitfall in it and I get discouraged.
1. Don't use the internet or internet services, period. <- Not tenable for most of us.
2. Use services who market themselves as geared towards privacy. <- Can't actually trust those services, even with E2E encryption because they could be running different code from what you think they're running.
3. Use regular cloud options, but stack stuff on top - VeraCrypt volumes or Cryptomator with Google drive, GPG for email, etc. <- Really difficult to setup and have a nice reliable way of accessing data on mobile/desktop/etc. No security audits on a lot of the open source software.
4. Host your own services - i.e. a Nextcloud 14 instance on EC2 with an S3 backend, then use client-side E2E <- Difficult to make sure you set the service up in a safe way, and not even a fraction of as much resources in auditing code as, say, a giant corporation.
5. Spread what you do out over multiple services - FastMail for email, DropBox for cloud storage, Standard Notes for notes, etc. <- A real pain.
I know there will never be a consensus on this, but I'd love to hear what your thoughts are on the best way to keep my personal files and notes personal to me. Let's assume I'm not a target of any spy agencies or whatnot, but I want to make it very, very difficult for anyone to read my person notes and files but me.
I'll be completely honest, as your standard "data scientist" who hopped on the bandwagon and came from having a PhD in academia in an unrelated field, I cringe at these articles. I'm not entirely sure why. I think it may be two-fold:
1. A little bit of the selfish "oh no, the secret's out, at what point is my salary going to drop when the demand is met by the dedicated Master's degrees and bootcamps?"
and
2. These articles seem so incredibly corny, it's almost embarrassing. The "hottest job"? Ahhhh, stop it. But these things go in an out of phase, similar to back in the day when "anesthesiologist assistants" (CRNAs, AAs) were the hottest thing for Bloomberg to talk about. It will not last forever.
The irony is that I probably only knew "data science" (always in quotes) existed because I read one of these cheesy articles. I mean, we all know that statistics have been around forever, but that there were dedicated positions where you could run stats, build models, and then deploy them all in a single role was foreign to me.
So it's a combination of a potentially irrational fear of self-preservation, and laughing at the state of affairs where some basic stats work will pull in that kind of money.
I tend to have fears about the future, always wanting to hedge myself so I don't become outdated. In the data science sense, I see the field becoming super super broad and eventually saturated with new supply, so I debate on whether I should pivot into management of analytics in general or not. I.E. getting my hands off the keyboard. Ultimate goal would be to help define, strategically, how statistics/data mining/machine learning/yada/yada/yada are used at a company.
My goodness, are you me? I've been having exactly the same thoughts. Provided one finds a data science role that roughly aligns with your training / interests, the actual work is comically easy compared to what you go through in academia.
A boot camp can easily teach someone to, say, estimate a linear model or run k-means. I dread the future when the industry decides the right way to put up barriers is by creating ever less-realistic interview loops that are even more coin-flippier, dice-rollier, card-shufflier.
A boot camp probably won't tell the process to make sure a linear model is the right choice. It probably also won't tell you when you should and shouldn't use k-means. And in most cases you probably won't have very good answers as to the certainty of your models.
Imagine you're hiring someone to build a house for you. Would you feel comfortable with someone who's just been drilled on how to use individual tools? I would want someone who had been taught a step by step process for how to put together a house.
> A boot camp probably won't tell the process to make sure a linear model is the right choice.
Very true. On the other hand, it's a pleasant rarity when I see positions that appear to index more heavily on, "how well is this person able to conceptualize the problem and choose an appropriate method?" than "can this person do X?"
Lots of folks can do X; fewer can conceptualize a research question and choose the appropriate X; even fewer can carry out the X and communicate robustly what it means.
The latter two start to get into squishy territory, but also are where the value is. They also seem to get the least focus in advertising / recruiting / interviewing data scientists.
It reminds me of studying evaluation methods in planning. One that people are really familiar with (at least anecdotally) is cost-benefit analysis. Conceptually, it's very simple. The problem is that the costs and benefits that are hardest to measure are very often NOT measured. And they're very often the sorts of things that people find the most important. So, you end up with an answer that encodes a ratio of easily measured things rather than important things.
So too with data science. Easier to check whether someone can remember basic probability rules and carry out a linear regression than it is to diagnose whether someone can reason carefully about an amorphous business problem.
If the example of coding bootcamps is indicative, then there will be a period of hiring of Data Scientists with minimal background, and an institutional learning period of "oh they are cheaper but don't deliver results we can use", and thus the job will still remain one with more openings than qualified candidates for some time to come.
Stop worrying & go vertical, aka use these (and any other momentarily fashionable) tools in the niche domain you are a PhD expert, so that you will never become outdated.
I don't think it's irrational. As somebody with just a masters in a tangentially related field (economics), I find it surprising how low the barrier to entry is for this job (I got in, after all). Anybody with a decent undergraduate level understanding of stats could learn to do most of the production stuff that companies generally do in a few weeks, at best.
