Man going back to apartment living though

We'll see what happens during my first review here in a couple months and go from there I guess. Some possibility I get RSUs which incentivizes me a bit.

Golden handcuffs incoming

Second onsite today. Only 4 panels in this one, I think I crushed the first 3 and struggled a little on the last one.

One thing that helps you do well on interviews: one of my interviewers gave me almost exactly the same problem as a DBAM guy gave me. It had to do with a matrix of numbers where numbers are sorted from left to right and from top to bottom, like so:

and how you can exploit the properties of that to find stuff quickly. I guess this is either a popular interview question or I got absurdly lucky on sample size. DBAM gave me a slightly more complicated version, which I struggled a little on a final perfectly-optimized solution for - count the number of occurrences of a number in the matrix - but today it was just, return whether or not a number exists in the matrix, ultimately getting to a O(m+n) solution if the matrix's dimensions are m by n (i.e. linear on the number of rows and columns, NOT on the total number of items in the matrix), and after my previous work on this problem I was super ready for it.

What's DBAM?

Desirable Bay Area Megaco, first place I interviewed

So Cloudflare sounds pretty terrible.

http://cryto.net/~joepie91/blog/2016...ave-a-problem/

Suzzer: useful reading re: CloudFlare's attitude towards content moderation: https://blog.cloudflare.com/why-we-t...daily-stormer/

I don't really have an opinion on the technical merits of the service, I've never used them.

There is always going to be problems when theory meets practice. In theory free-speech is great, but in practice there are way more *******s out there than anyone thought, and they don't get more timid when they find like minded people.

In the pre-internet days anonymity was useful for speaking truth to power. On the internet it is used to radicalize people and emboldens the worst views in society. The idea that exposure to the light is the best disinfectant for poisonous views only works when you know who to expose. I don't know what the answer is but it's going to involve some way to reduce anonymity.

It might also help to get some people who grew up with this technology making laws to regulate it instead of a bunch of grandpas trying to understand what their nephews are telling them about it...

Do you guys have some length of time without something better than a cost of living adjustment that you deem unacceptable or just feel like it's an ev loss? Or is that not something conforming to a standard

Depends on how much I am being paid and how much I feel I can make elsewhere. For example if I start a job and I feel I am being well for my skill set and experience and feel I can't make much more elsewhere without a ton of effort, I would be ok with only col increase. But after say 3 years, I may feel I have enough experience that I can make more elsewhere. If I start out underpaid, I may be ready to bail after a year. If overpaid, I could stay for 5 years. Just spitballing numbers here.

Unless there are extenuating circumstances I'd start looking if I got a raise that was CoL or less.

Makes sense. I am an a somewhat fortunate/unfortunate spot of being paid much higher than the norm for my experience in my area, and I consider the experience worthwhile. So overall, it's hard to make a firm decision on it. I don't want my ego to ruin a cushy gig.

Really? You expect more every single year? FTR I got 3.8% after 8 months. I would be highly disappointed to get that again this go round though

Your situation being somewhat overpaid for your experience is definitely something that would make me more likely not to start looking elsewhere. But in reality I think other companies aren't likely to balk since you have that to anchor negotiations.

But basically if you only get a CoL raise, they're saying your value is unchanged. In reality the added experience has made you more valuable and the experience in their org has increased your value for your employer specifically.

Never hurts to go ask the market what they think you're worth in case it's out of whack with what you're making now. (I'm one to talk, doing this for the first time ~ever after 10 years of working as a programmer)

I could use some financial advice. I just signed up for the employee stock purchasing program at my new job.

When I sign up to the program the current stock price is locked in for 2 years. I can choose 1%-10% of my pre tax salary to be deducted from my paycheck and set aside (so no tax savings on that amount). After 6 months they buy stock for the set-aside amount. The price you buy at is either the locked in price if the stock went up or the latest price if it went down. Then you get a 15% discount on that buying price.

I'll have to pay income tax (45%ish) off the 15% discount and capital gains (22%) on any investment profits beyond the discount. I have the option of decreasing the percentage once during the 6 month period. At the end of the period I'm free to sell the stock or keep it, and re-enroll in the next 6 month period.

How does this sound? fairly standard? If I sell immediately the only way I can lose money is if the company goes busto (right?), but it is a fortune 500 so not much chance of that. I have some loans at 5-12% interest rate (mortgage, car loan, appliances). This should offer better returns than paying those down faster, right?

ESPp is a slam dunk. I was purchasing at less than half price for a couple of years recently.

Yeah, if I had no loans it's def a slam-dunk. The stock is projected to rise by about 10% next quarter. I guess I just need to sit down and excel this out.

Max it out and take the free money

From SO bad answer vault:



Top rated answer obv. Can’t even wrap my mind around how bad that is.

I know the fread buffer is probably large enough for it not to matter but you can’t always depend on that

...is it? cppreference:

Sounds like either way, it's gonna make count * size calls to fgetc?

I don’t think you can absolutely guarantee a performance equivalence. And the internal buffer size of fread varies - there is potentially overhead there.

Furthermore the first call will return 1 and the second call will return number of bytes read. The first call can fail if you do not read every byte you expect. There’s a massive difference semantically even ignoring performance.

Anyway I have seen instances on some systems with versions of these two calls where the performance difference was massive. you can’t even guarantee how fread is implemented either, the iso standard doesnt specify anything about getc or the number of calls fread can/should make. It’s a really dumb assumption to make.

I don't fget any of that, so feel free to explain which way is better and why.

Sure, you can't guarantee a performance equivalence, but you also can't guarantee one way or the other because like you said, it's implementation-dependent. I'd probably pick 8192, 1 because it seems more likely to be optimizable based on how I'd theorize an underlying implementation might work - but I'm just guessing, the standard doesn't guarantee or imply anything. In that sense, they're equivalent in what the standard implies about how your code will run.

Like, "benchmark it and figure it out for yourself" is also a good answer to that question because of it being implementation-dependent - but if you had to pick an answer to "which is more efficient" of either "they're the same" or "they're not the same" (the latter implying you can say one is better than the other, which, no, you can't without knowing the implementation) - the first is, semantically, more accurate, at least with regards to the part the question was asking about (efficiency, not the return value).