2,341 karma · joined July 15, 2012
Worked on machine learning for medicine at Cardiogram. Studied CS at UC Berkeley. Former undergrad researcher and teaching assistant for Berkeley's ML class. Former intern at Twitter on Ads Analytics Infrastructure.
Email: bc@brianchu.com
github.com/bchu
twitter.com/brrrianchu
It's remarkable that at a place like UC Berkeley (where I study CS), the pinnacle of innovation is videotaped lectures and programming assignment autograders. That, and iClickers during lectures (devices that let you answer multiple choice questions). Many universities have worse technology than many Fortune 500 companies - in areas that directly affect students. Apparently Berkeley's online course enrollment system limits the number of times you can login because it's based on a legacy telephone system originally designed for dialing in.
I've been thinking that technical/engineering education is the most ripe to be disrupted - someone with a CS education needs to rely much less on their credential/alumni network than someone with, say, a liberal arts degree. We're already seeing this with the coding schools that are all emerging.
I also know that Apple uses Angular for a few internal (non-public) developer tools/consoles.
Angular is a framework, heavier than the other frameworks too. Twitter and Facebook are particularly good examples because they use both server and client-side rendering, so they get the best of both worlds (fast initial page render and fast subsequent client renders). I love Angular, but Angular can't do this currently.
Others: Google Maps.
I don't find Gmail to be a particularly good example, unfortunately.
I think anyone can make it without a degree given sufficient motivation. I'm doing a degree because I recognize I don't have the motivation to teach myself certain areas of "hard" math, statistics, and theoretical CS. Yet I recognize that these topics will be useful. All of these topics can be learned through MOOCs/MIT OCW today (given sufficient motivation).
An alternative you could consider is to try transferring to a top CS school like MIT/Berkeley/CMU/Stanford/UIUC, or to another all-around top school. You may find that you'll be seriously challenged and stretched at those schools (or maybe not). This depends on your financial situation, though.
If each fellow had to pay $150-250k for college, I bet not a single Fellow would entertain the idea of going to college.
Personally, I took a gap year before college but eventually decided to return to college afterwards, knowing that I would be able to get a $100+k dev salary in the Bay Area if I chose to forgo college. Part of that decision was driven by the fact that my parents would pay for college (and then some).
On a broader note, taking just one year off gave me a ton of perspective on what I wanted to do in life, my priorities, and on the value of college. Everyone should take time off.
I don't put much faith in the arguments that "you'll never go back" if you take a year off - I did, and when I made the decision to go back, it was precisely that - a conscious decision, rather than how most students go to college (which is by following the default path). At any time, the decision is yours to return to college - if you decide not to return to school it will only be because you judge your current opportunity (the startup or wherever your career takes you) to be better than going to school.
Obviously the details of this opportunity are pretty important - how much equity are you getting? How much control? Who do you get to work with (the people you meet and/or work with is probably the single greatest factor in your early career)? How much funding do you get? How flexible is the parent company in letting you make decisions (remembering that if you don't have >50% of the equity, they have final say in all decisions)?
If you want to take this offline, my email is in my profile. I wouldn't be so quick to dismiss an opportunity like this. If you think your skills are up to it, it's worth considering. At the very least, I would take the trip to Japan, and come prepared with tough questions for the executives (questions about control, autonomy, flexibility, the conditions of funding, how they make decisions, their philosophy around product management, etc.)
For structure/content: raw HTML, or less commonly a language that compiles to HTML (Haml, Jade, etc.).
If the HTML is being rendered and served from the backend, you might use a templating language like Django's templating or Jinja (Haml and Jade also have these features).
If you are creating an application that does a lot of AJAX, you will also often use a JavaScript templating system like Handlebars.
For styling: raw CSS, or fairly commonly a CSS pre-processor like SASS, Less, or Stylus. Very often you will use a CSS framework like Bootstrap or Foundation to give you pre-made forms, buttons, grids, and other essentials.
For logic/interaction: JavaScript. Many people will also use compile-to-JS languages like CoffeeScript, TypeScript, or ClojureScript (in decreasing order of frequency). There are a lot of libraries/frameworks that can be used here. Most commonly you'll use jQuery and jQuery UI widgets. For more complex applications, single page app frameworks like Backbone, Angular, or React come into play.
Build process: For more complex projects, build tools like Grunt or Gulp to watch for changes to the filesystem and re-compile files (typically when you're using CoffeeScript or a CSS preprocessor), to package everything for deployment (minifying JavaScript, etc), and to run tests.
Packaging: Often you'll use a tool to keep track of modules/packages. These tools are quite different. Bower, RequireJS, Browserify, Component, and other tools occupy this space.
Often you'll use Node's npm to keep track of build process and testing packages!
Testing: For testing you'll often see frameworks like Jasmine or Mocha/Chai for unit tests/integration tests. For end-to-end tests you might see things like Selenium (run tests in a browser window). There are also test-runner frameworks like Karma.
Debugging: Chrome's developer tools (or Firebug/Firefox/Safari).
For someone starting out this is probably overwhelming. The important things to realize are that:
1) No one uses all these tools
2) Most of these tools are only useful for large complex codebases with large teams.
2) The most common stack is raw HTML, raw CSS, raw JS, jQuery, jQuery UI, Bootstrap, no build process, no packaging, no testing.
In Asia (I think China specifically, not sure), QR codes are ubiquitous and widely used.
Edit (reply to below): I took a year off to work before I went to college, so I think I have a good perspective when I say this. You assume that: 1) college is going to educate someone about the world, get them a good job (not a measly startup job), give them discipline, and introduce them to cofounders; and 2) college will do this so much better than you could by working in industry that it justifies the cost (anywhere between $150-200k tuition for a good school, and possibly $200k in lost savings). Suffice to say, not all of those things are necessarily true.
I'd be happy to talk about this perspective to anyone over email.
The exceptions:
1) when used with new, it's the new instance of the object (pretend that when you use the new keyword, an Object.create() call is inserted in the beginning of the constructor function),
2) when there's nothing to the left of the dot, it's the window object.
3) .call and .apply force this to be something else.
Give it a year or so.
Companies, people, and governments deny things that turn out to be true all the time. I'm not saying Nike is lying. Just that their denial is not clear evidence of misinformation.
All these companies are essentially doing is claiming that you are giving up your rights. It remains to be seen whether these provisions will actually hold up with legal force in court (I doubt most of them will).
I haven't made up my mind whether Mozilla should have fired Brendan.
[1] One of those is hate crimes against certain classes of people.
var funcName = function(){}
function funcName(){}
The problem is that using named functions involves nuances that can be confusing. Generally it's considered idiomatic to just avoid named functions altogether. You can read about some of the nuances here: http://kangax.github.io/nfe/The second time he applied with Dropbox, he was accepted. After being accepted, he got Arash to join him.
It seems to me that about every batch there is a single founder. But they tend to find co-founders after acceptance.
Furthermore, it's not like quantity supplied is perfectly responsive to price. It takes time for new people to enter the industry.
Talking about the long-term is rather pointless because technically, from a theoretical standpoint, everything is supposed to equilibriate over the long term. All shortages self-correct over the long term. Again, this is theoretical.
Google uses Angular for a few projects.
Amazon AWS uses Angular for their dashboard.
Pincus still retains voting control over the company (much like Zuckerberg retains control over Facebook). Any changing of the guard that took place had to take place on his initiative.
Now, there's no doubt that he felt pressure to leave from the stock market. But that's not an ouster.