idunno246

My account team tried to sell anthos, but they knew nothing beyond “it’s for the dc”. They had no idea what it actually was, or what problems it would solve. They’d get us on the phone with engineers who would just ask if we use VMware and when we said no, it’s like… oh well doesn’t work for you. They failed to educate internally or have any cohesive plans for it. They couldn’t even define what it was

I’m not, it’s the other problem after the promo issue. Google is an engineering only company, if you aren’t an engineer you’re second class. So engineers see a cool problem like live migration and solve that. But they do an awful job at building a product customers want(product managers), and supporting(customer support, marketing) that.

If you look at stadia, they built the cool streaming platform, but failed to actually sell it.

At a smaller company, definitely not. The focus changed at least yearly so planning that far out was fruitless. The one year plan was see what is making money and do more of that, keep the lights on, and try a couple quarter long experiments.

At the last big company, yes I was part of a multi year initiative. Part of that is probably mythical man month, we just couldn’t execute quickly. There were many teams and layers and dependencies that needed coordination, scheduling, headcount. Part was there’s enough legacy stuff that is core and can never break that every step had to be derisked a ton.

i know thats the common refrain, im sure its true most of the time, but aside from this last year or so ive gotten more increases from promos/equity than from job hopping(which have mostly been flat)

Right, like the company raised bands for new hires but didn’t raise current employees commensurately, that is in line with needing to job hop, and tracks with the other comment of a couple thousand. But the difference in pay now is so drastic now basically everyone with 1+ years is looking or left

There’s a difference between variable and a new hire at my current level would get literally double my current comp as a person at the company for five years

unfortunately demand isnt driven by the crisis, its driven by the states' willingness to pay

> Again that would be request tracing, not application logging.

from what ive seen people who rely on logs do not have proper request tracing and/or metrics. its much simpler to dump a bunch of unformatted junk to stdout and throw elasticsearch at it, but not a good experience.

i agree overall, i find logs to be a last resort with really bad signal to noise ratio.

these tend to be for one-off analytical queries. you want ever user with flag X >10 joined against five other tables each with similar filters. you don't know ahead of time what that query is, your analyst thought of it this morning, so you cant make indices ahead of time. and itll never run again so you don't need to take the performance hit keeping an index. and someone has to decide which indices to keep, but app engineers arent best utilized figuring out indices for analysts.

the indices is nice, but the bigger selling feature for me is if you have many services, and each services data are in the warehouse, you can join against them all together.

nah the article was really overblown. 'we've approved budget for five years. and our goal is to beat azure.' more or less, but that doesnt mean in five years there wouldn't be more budget approved, and it didnt seem like an ultimatum connecting the two. businesses dont plan infinite years out.

while google cloud as a whole i dont see going away, individual products within it(like iot core) are gonna always be up for the chopping block

> I wish we'd used containers on Lambda from the start.

Does this actually solve a problem? You still need to implement the lambda interface. so as someone who hasn’t used them, containers just seem to replace the zip file format with no actual portability benefit

the incentive part of ISO is you don't pay regular tax - other stock options you do though. unless you're talking about an 83b election(early exercise), or double trigger rsus, but those are not really types of options. also, amt limits are rather low, they didn't adjust for inflation for a long time so its surprisingly easy to hit, anecdata but I know a ton of people hit by it

the irs defines the rules for an ISO, so to be incentivized it must exercised within 90 days of leaving. with an NSO(or ISO), the limit is ten years from grant, again set by the irs - you can still be employed and have to deal with this. until recently, companies typically gave ISO - this means they had to do ninety days. so first, blame goes towards the IRS.

ISO are generally more favorable because you dont owe regular income tax at exercise - with NSO you do owe tax on the difference between grant and FMV prices, even if you can't sell it. However, AMT does not recognize ISOs so you may owe AMT against that same difference. AMT is weird though, you can get that money back over future years by claiming tax credits

luckily, the trend seems to be to convert ISO to NSO and extend the expiration, but thats still a relatively new change and not everywhere.

Even more fun, my primary care office billed a visit to insurance under the PAs name, so insurance treated it as a specialist. So they tried to charge the higher copay… had to get them to resubmit with the doctors name, despite not seeing them, to get the cheaper rate

That said, my experience with pa vs md hasn’t really been that different, they all just follow the same flowchart

I think I’ve just had too many discussions where someone’s complaining they aren’t given time to fix tech debt, and they think by calling it that it automatically qualifies for needing to be done. And the person that they’re asking time from just hears someone being a perfectionist. At the heart of it, it’s miscommunication.

So I agree that the term is here to stay, but there’s better ways to communicate it to product/management. Like your example, making it clear to product what shortcuts were taken and how it will need time to fully implement. Code improvement vs tech debt doesn’t really make a difference, it’s just getting that conversation convincing

i think riot put together a good article on technical debt: https://technology.riotgames.com/news/taxonomy-tech-debt

i stopped using the term technical debt. 'technical debt', 'refactoring', etc are not helpful to product owners. if something is contained in an area youre never touching, its not really worth even tracking because making it better won't improve any business metrics. tickets get lost, better to throw a comment in the code if anyone starts touching it later. If new functionality requires code changes, then that's just part of the requirements for that functionality. If some bit of code is slowing you down, then its a first class ticket to speed up developers, with explicit numbers on hours saved.

at the end of the day, the business doesn't care about code quality, it cares about revenue/profit, so reframe any tech debt in those terms

it depends how you want to grow. if you want to be really deep on one subject, a larger company will get you that more easily since teams tend to be more siloed and focused. if you want to learn a breadth of things, jump from data to backend to frontend to devops, a smaller company will be better. if you want to grow in training/mentorship/leadership abilities, high growth small company will give more opportunities but nobody to help you, whereas a high growth team in a large company may be much better supported to train you how to mentor/lead(there should people that can mentor you).

to me the title sounds like its cancel all flights, like its closing. but that doesnt really make sense, it's just phrased badly. its borderline clickbait, it sounds really bad(worse than it is), and the only way to understand what the title means is read the article. 'any more seats' - i think a 'more' in there would make the title ok

a quick search on google seems to say that theres no asset check for section 8, just income checks(including interest/etc from assets)

this is really tough for people with disabilities, since a lot more stuff then depends on medicaid. An easy example, if you qualify for medicaid then you qualified for the emergency broadband benefit, which was good for 50/month off internet. Or more complex, you get a budget to spend on home health care workers, transportation, respite care, therapeutic activities, etc., go over and you lose thousands of dollars of services necessary to live. thats even ignoring the actual medical benefits of medicaid

able accounts that are listed have pretty big caveats. and are themselves confusing [like 529, you can buy them in any state, all with various fee structures]. the best course of action is to set up a special needs trust and able account and move money between them cause they each have different restrictions. but setting up a trust itself costs a couple thousand for the lawyers, which you arent allowed to save.

my sisters disabled, but luckily NJ has a program called workability that raises these limits if you have a job. otherwise its impossible to save.

outsourcing doesnt necessarily mean india/eastern europe. we had great luck with south america and they were able to work the same time. this was a few years ago, the contracting company thought we were a little abnormal at the time, but everyone put on our project went through a regular interview so they were talented, and were treated as a member of the team instead of some remote contractor.

its also free under ~50k AGI, though it did keep trying to upsell me to the paid version, the free file program is pretty readily available from many providers. but they all do the type of thing that they just slapped intuit's wrists for and try to trick you into needing to pay when using free file, or trick you into your "free" taxes only being federal

sre means different things to different people, it’s an overloaded term like devops. but I feel like he spent a lot of this article defining what it meant for this team - all that about embedding a dedicated person is their definition of sre

I found that only happens if you try to hit the most common letters in your first guesses. If you use middle of the road letters you tend to get only a couple while eliminating a lot - s and e would be bad early letters because they don’t reduce enough words

yea that friction is good. i wrote some code. someone liked it and added a dependency of their app on my app. i needed to update my code - all the sudden i was responsible for updating some other random app and ensuring it kept working - behavior we considered a bug and they didnt, so both code versions needed to work. the monorepo let them make an api where one was never intended to exist

my company turned off all our conference room video conference hardware to force people to dial in individually. the argument being that the remote people have less representation if theyre a small box on a screen vs everyone in the room

in the US the IRS is pretty known for working with you - if you call theyll set up a payment plan or whatever.

ive known a number of people that got mildly burned by taxes, nothing crazy just unexpected, but with bigger numbers it could have been bad. When you have ISO(incentivized stock options), one of the incentives is you dont pay tax until you sell, which sounds great. There's this other way of calculating taxes, AMT(alternative minimum tax) which does not respect ISO, and you pay whichever is bigger. normally people dont get near AMT amounts, so most people dont know it exists. So when you exercise you think theres no tax, but then you do your taxes and realize actually you owe tax on money you never received.

It’s not never, but it’s rare. They renamed to cloud map and took a free product and started charging for it

https://mobile.twitter.com/0xdabbad00/status/106819770559422...

my experience is the opposite. there's a lot of really insecure defaults in gcp, some of which are really hard to override - gcp makes it easier when the default service account that every is ridiculously overpermissioned, and for the longest time you couldnt change the service account in cloudfunctions or app engine, i forget which.

The reason you hear about s3 buckets and not gcs buckets is because usage of s3 is way higher and way older.

time limits put an upper bound on how long youre spending on it which is good for a lot of people. ive given take home tests and some people clearly put more time on it than we expected.

as an applicant, ive refused takehomes that were auto-sent on receiving the application, but if its after a phone call with the hiring manager or engineer its not so bad. but then, i did end up having to defend why i didnt do a bunch of parts ancillary to the assignment - 'why didnt you build a config loading system?' 'uh cause i timeboxed it and that gives you fairly little signal compared to the business logic'