nate_martin

>The first problem every startup solves is scalability.

I don't think this is true at all. The first problem they solve is typically finding product market fit, which startups will do by sacrificing scalability and quality for speed of execution.

>The compiler itself is to be developed in OCaml.

These seems like a misstep that I've seen in a few other compiler implementation courses. For some reason these programming language professors always insist on completing the project in their personal favorite language (Haskell, OCaml, Standard ML, etc).

As a student this makes the project significantly more difficult. Instead of just learning how to implement a complier, you're also learning a new (and likely difficult) programming language.

It takes energy to remove carbon from the atmosphere. With abundant energy from fusion you could just remove the carbon added by humans.

Compared to Netflix and even YouTube, TikTok has an extremely tight dopamine feedback loop. Any buffering of a video after a scroll event could result in the user's attention being lost and the session ending. My guess is that they've spent a lot of time optimizing video load because of how correlated with their revenue it is. For Netflix doing this optimization might not be worth the increased infrastructure cost.

Crypto soaks up excess liquidity like a sponge.

This looks really cool, but I'm a little bit confused about why this needs to be decentralized. You could theoretically just implement this network as a saas instead of on top of a blockchain. You could even still do all the financials with ether but just maintain the state on servers somewhere.

The big fear I have is immutability of blockchain data in the case that something goes horribly wrong. If there's a problem with the ethereum core protocol, solidity compiler, wallet software, colony contracts you could lose everything. Vitalik already said that there probably wouldn't be anymore hard forks.

It's a little annoying when companies put framework names in their job titles. Asking for a "react engineer" or a "boost programmer" will likely turn away a lot of people who would actually be a good fit for the role. My suggestion is to just be open with the company and tell them what you do know and how quickly you can get up to speed. Something like: "I haven't used React on a paid project before, but I have a lot of experience with other frontend frameworks like x/y."

The online matches only allowed 30 seconds per move.

People have wildly different experiences there. Some hate it and burn out after a year. Others really like the work environment and get promoted quickly.

One good thing I'll say about Amazon is that it's a great way to break into the "big four" tech companies. If you didn't get into Google or Facebook initially but got into Amazon you can pretty much guarantee you'll get recruiter spam from lots of top companies.

Might be smart to remove their logo from your site. I would expect a cease and desist if this gets popular.

For someone familiar with the products, yes. But for someone new to AWS most of these options don't make sense. I'm looking at a menu with hundreds of options like Snowglobe and Beanstalk and I have no idea where to start. Are they trying to brand their trademarks as the defacto name for the service (like Kleenex and ChapStick have done)?

So many new named services. AWS will soon get to a point where their product dropdown wont fit on a laptop screen. Maybe it's time to consolidate some of these services into more general products.

This doesn't address the root comment at all. Does Amazon actually think MXNet is the best? Or did they simply choose the next best thing that isn't already backed by another "big four" company (Google -> TensorFlow, Facebook -> Torch). It's hard to believe this is actually about scalability without any data.

"flatland" corporate structures don't really seem to scale. You eventually get to a point where the cost of coordination everyone needs to do is greater than the cost of having a bureaucracy. Also, I'm guessing that these structures are only effective in firms like Valve: a small number of experienced employees, no hard deadlines, and lots of specialists who know how to best allocate their time.

It isn't the same as doping in other sports. Climbers will often specify if their ascent involved O2 or not. It's still considered a big achievement in mountaineering to summit say K2 or Annapurna with O2.

Maybe someone who works on deep learning could comment on what this provides vs other open source systems like theano, tensorflow, torch, etc.

Technically I think you could make any n-sided die for even n > 2:

1) Take an n-gon cylinder.

2) Smooth off the two ends to be spherical so that it never lands on either.

Now all sides have equal probability of facing up after a roll.

I wonder if Uber will ever offer a programmatic courier service. It seems like the "Uber for X" companies will struggle to make money if they have to maintain their own delivery fleet, and tapping into a large pool of Uber drivers could allow smaller companies to auto-scale their business just like with cloud compute.

Probably something like this:

template< class Key, class Hash = std::hash<Key> > class BloomFilter;

This example works well for raw data but not for complex types. You could make the filter a template, taking the key and a "hasher" function as template args.

Not sure if the "USA" chant after the first stage landed was warranted.

You're betting on a a startup's product to stick around longer than a tech giant's?

Getting a 503.

From web.archive.org: https://web.archive.org/web/20150403043329/http://dolekemp96...

Are there any examples of companies that achieved success by way of language choice other than Viaweb?

Fun fact: there has been a problem with Googlers calculating hash functions that pass all the test cases for certain questions.

Merchant sets their price in dollars (for example) and the price in Bitcoin is calculated when a transaction is made using the current btc exchange rate.

Yep, also enriching uranium to anything near weapons grade is a cumbersome process.

Also a good read on the subject http://paulgraham.com/procrastination.html