Having to reverse engineer your own product in order to support it gets old really fast, particularly when you are customer facing and have to explain why dates are not being hit...

Has Skype ever pushed an update and had you feeling like "this is great".

Agreed. But I'm talking about fairly simple stuff where there are like 15 lines of code to go look at. And people just don't do it unless I make them.

As an example, we have another platform from a company we acquired that is still running a portion of our business. We rely on webhooks to update state in the main platform. If someone manhandles something in the second platform and bypasses a webhook someone needs to go in and change the state in the main platform. There's an API for those webhooks but instead of looking to see what we do with the missing event it's "hey blackize, how do I do this thing?" As a note this comes up maybe once every couple months and the solution is calling a single method.

It is also always easier to ask someone than to figure it out. Instead of giving them them fish try teaching them to find it themselves: "I'm in the middle of something and don't remember, try looking in such and such a repo." Being the guy with a bunch of system knowledge is a blessing and a curse.

I dunno, I'm still a noob, but in those situations I'm putting in a solid effort to figure out what the code is doing. And will feel bad "wasting" someone else's time due to my laziness.

But I also seem to have a knack for reading other people's code and quickly figuring out what it's doing.

Also we had a career fair type thing for my bootcamp and I did very well.

I was interested in 3 companies and followed up with them the next day, all 3 responded same day and now 48 hours later I've spoke to the CTO at one and VP of Engineering at both the others.

Yeah, putting in a good effort is how I got to be the guy who knows the whole system. No matter how many times I make them read the code, the first instinct is to ask for the answer

Yeah I usually make them go make a good faith effort. But then that leads to the asking for more documentation issue

In your process, do the developers submit a change for review before putting it into the production code? How do they verify and document that they fixed the problem? Do these kinds of issues get discussed in your scrum meetings?

Blackize, I can kind of sympathize with the request for documentation - especially for things that sound like they're changing production data. Code can't tell you everything. Maybe it can't tell you who uses this data and the format it expects. If you need to update the data someplace else to keep things in sync. If there's a second clean up step necessary elsewhere. Etc. Regular operations like you seem to be talking about should be documented. Including with things like 'how do I verify this change worked'.

Our answer to these things is having a common place people look and a shared responsibility to update documentation. So if someone comes asking for documentation on how to do X, then you help them and they're responsible for documenting it for the next person. It's not a perfect system by any means but better than nothing.

So next time walk the person through the steps and tell them to document it.

Here's an article + comment section on Hacker News regarding bootcamps.

https://news.ycombinator.com/item?id=14679061

There is definitely an open market for someone to create a job placement service to test self taught/bootcampers and weed out the bad ones and get the good ones jobs.

No.

Why only bootcampers and autodidacts?

Many bootcamps are interview preparation courses by design (and this is often reflected in the economic model) and so Goodhart's law applies: you don't get people who are good entry-level programmers, you get people who are good at entry-level programming interviews.

That doesn't apply as much to autodidacts.

I just wasn't sure why anyone would specifically target bootcampers and autodidacts. That population is very small, and if there is a thick-line separation between good and bad, then it stands to reason that the employer should never take on a bootcamper or autodidact over anyone with a college education, including someone who went to a technical trade school or community college.

There is no good LCD baseline for programming for college students, bootcampers, autodidacts, or SR devs with 10+ years experience. The fallacy is believing that a bootcamper is worse than a college student by all measures.

Autodidats should take over the world by finding the best resources, collecting them, adding resources on starting your own business, and innovate. Hiring managers and recruiters do not take chances, where an autodictat can.

Currently I am watching a YT channel, Bo Qian, and am learning shared_ptr and the RAII saving tens and thousands of dollars.

Just finished Bo Qian's Advanced C++ tutorial.

YT in spoiler


His presentation is geared towards C++11, which is mostly where the industry is at. I'll call it "Contemporary C++". He has 193 videos total in play lists, all free, and relevant to how C++ is actually used, and why you should or should not use, for example Multiple Inheritance instead of Composition, etc. Then I plan on hitting my book on Concurrency, and find Concurrency YT's, and then Network Programming with Boost::Asio, not necessarily in that order.

You also save loads of time by going at your own pace.

Also, tips, spped 1.5x and listen to music during lecture.

First thank for the info. Second, what kind of work do you want to pursue applying your C++ knowledge? Not sure what that landscape is like now for this but a few years ago if you had the kind of expertise you are obtaining with C++ and excellent knowledge of the Linux kernel, you had a great shot at getting a high paying job with hedge funds/financial institutions doing high frequency trading.

I am going to go all-in on machine learning. In the next 5 years, hopefully 3, I'll have a PhD's worth of knowledge, and earning passive income in the meantime to foster my academic goals. I may go into financial markets w knowledge.

I'll post a best-of thread at some point w YT's. Most of the ML is in Python but TensorFlow is not exclusive to Python. I am going to finish my exhaustive study of C++ first.

I am studying about 40+ hours per week.

I would be interested in that. Good luck and I will be waiting for the thread.

Sent from my SM-G900R4 using Tapatalk

Well anyone who can pay the tuition can go to a boot camp, whereas getting a degree means the person had to at least be able to pass the classes. Also, with regards to autodidacts, Code School sold for $36m so they must have a fair amount of subs - I don't know how many autodidact applications companies are getting, but the more they get, the worse it is for those applicants. I would imagine that there is a larger range of skills/ability for these applications than from those who have a CS degree.

An autodidact isn't can take chances, they must take chances. You are correct, the mind-set and goals between an autodidact applicant and a non-auto hiring manager are very different.

The silly part of running your own business is that coding ends up playing second-fiddle to everything else. No end-user buys a program; they buy a story, a person, or a pretty website / advertisement.

Most of the resources I've found for business development has been at best, philosophical. The only thing that makes a difference is knowing the math or not knowing the math. You are making a huge mistake if you are trying to figure out what works before pushing out ideas and seeing the results. You simply don't know until you try, then you realize no one knows what they are doing when they start out, and no one knows how to make your idea work. The only practical resource I've ever found was Marketing Metrics.

I've been the 40+ hours study person.

I say this to everyone: don't wait. Just apply, get your feet wet, start your business, etc. The worst thing you can ever do is try to become the absolute best before going out and trying to break in. You already said the goals of hiring managers and autos don't align, and this is true. No one cares, no one is looking for diamonds in the rough, and the worst situation you can be in is having zero experience and having the knowledge of someone with 10 years experience. If you can do basic neural nets and gradient descents, you are well above qualified.

It's the reverse. Not anyone can drop $10k for a boot camp and fade a possible year of full-time job searching with no useful income.

The US government has guaranteed school loans, and if you are a resident of certain states, you can get an associates from a community college for free, apply for Pell grants, and at least have some money left over

Bootcamps have limited slots, and few get in. A state or community college basically has to accept everyone.

That's not really an autodidact resource.

I wouldn't know the applicant pool either, but I'd guess that, on average, an auto is well below the CS student in many areas, such as higher-level math, operating systems, and compilers.