ogreyonder

And when they give you their feedback, listen. They may only tell you once.

Already done:

https://news.ycombinator.com/item?id=6660089

http://i.imgur.com/mmPAztu.png

I didn't get much interest so I didn't bother building it :)

Yes; for people participating in an employee stock purchase program, it is called a Disqualifying Disposition. It puts you in the short-term capital gains bracket (with about 15% higher tax).

However, it's still a VERY good idea to do if you plan on holding stock.

I know stories of several people who were exercised options on 7 figures of stock, only to see the price collapse before they were able to sell. The taxes they owed because of that eclipsed their net worth several times over.

Agreed. His argument is essentially that you can't meet deadlines with Scrum, because you must always be in a 'shippable' state at the end of each sprint.

I'd argue that if you can't meet deadlines with scrum, there's no way you'd meet them without it. His definition of 'shippable' is very narrow--like you said, just because it doesn't meet console certifications doesn't mean that it's not "doing scrum".

I wouldn't put the app I've got after the first dev sprint on the app store, either. It's not done.

I wouldn't... I would live like a miser for 2-3 years and retire to do whatever I wanted. Maybe 4 if I felt like travelling the world for a decade. :)

On the contrary: in a world filled with a cacophony of malicious and intentionally misleading statements, the silence caused by the criminalization of lies seems welcome.

And Ebay was the first mover in the online auction space. I'm sure we can cherrypick counterexamples all day, but in general the first mover enjoys an advantage those that come after do not.

Agreed.

Just apply the usual headline rule: If the title is a question, the answer is "No."

5+ days for Fidelity

MtGox should absolutely have known about this issue. Everyone building anything related to BTC should know not to make assumptions about the protocol, and to treat every input as hostile.

You'd think they'd have at least one guy dedicated to nothing but breaking their software. They make my salary every day with transaction fees (well, maybe until recently) so you can't say they're unable to afford it.

Edit: Rereading this, it sounds more accusatory than I intended. I think your clarification was perfect, but at the same time that MtGox is at fault.

Is there any interest in a Chrome plugin that would passively report anonymized download speed for things like YouTube or Amazon? I can imagine that kind of info would be useful.

Yeah, I was confused about this too. They hook up your github then...don't even use it?

In regard to working on side projects:

> Just ask. The worst they can say is "No."

And that's where I have to ask: what do you do when they say no? I followed this advice a year back or so and it essentially put the brakes on any serious side-projects. I work for a big company that you've absolutely heard of.

Yes, the idea you espouse. Yours, given that you promoted it. Or do you want to revise your statement?

+1 to parent post. It looks awesome, the fade-in is kinda distracting. It also re-fades the main area when you switch between categories.

To the coders: Nice job on using multiple URLs to allow pages to be viewed without JS enabled. Having to reload each page when it is requested is not trendy, but it's nice to know Mailpile will work on locked-down browsers. Kudos!

Thanks! This looks really cool, I'll check it out.

A hybrid system sounds reasonable, but I'm more interested in how they think that Dagger is ASIC-proof. I can't tell without a real description of the algorithm, but what is presented doesn't seem hard to parallelize.

I'm interested in seeing a real explanation of the Dagger algorithm.

In brief: the only source I can find (that page) just contains a few sloppy lines of pseudocode. Where is real description and proof of its properties?

https://news.ycombinator.com/item?id=7115825

http://wiki.ethereum.org/index.php/Dagger#Algorithm_specific...:

You sound knowledgeable, so perhaps you could clean up their algorithm for me? There are a number of problems with their description that make it hard for me to evaluate. Having tried designing a memory-hard algorithm, I'd like to see what their key insight was, but there are a number of problems with their article. They use confusing operators, the wrong terms, and seemingly random constants.

How did they come up with the numbers 2, 3, 11, 2^21, and 2^22, for example? Is D the hash function or the underlying data? Or is 'data' the underlying data? Does + mean addition or string concatenation? What about "++"? They never even use "||", which they defined as string concatenation... Where does the nonce N come in? Is that actually supposed to be 'n'? How do they justify that the optimal algorithm is the naive one? There is essentially no proof of this claim in the article.

In short, I'm very suspicious of their "memory-hard" algorithm. It took a fairly dense, multi-page whitepaper to explain Scrypt, and yet their 'superior' version is just a couple of sloppy lines of pseudocode with no justification.

Here is what they wrote, for reference:

Let D be the underlying data (eg. in Bitcoin's case the block header), N be the nonce and || be the string concatenation operator (ie. 'foo' || 'bar' == 'foobar') . The entire code for the algorithm is as follows:

0: D(data,xn,0) = sha3(data)

1: D(data,xn,n) =

2: with v = sha3(data + xn + n)

3: L = 2 if n < 2^21 else 11 if n < 2^22 else 3

4: a[k] = floor(v/n^k) mod n for 0 <= k < 2

5: a[k] = floor(v/n^k) mod 2^22 for 2 <= k < L

6: sha3(v ++ D(data,xn,a[0]) ++ D(data,xn,a[1]) ++ ... ++ D(data,xn,a[L-1]))

For a throwaway, you are remarkable.

This is excellent. You're on to something, why not write more?

Normally I'd downvote the doge on HN, but this is particularly apt.

The article abuses the word "derp" to sound hip, makes weak shots at people and ideas left and right (without ever arriving at an argument), and generally rambles on for a whole page.

What a waste, how did this make HN top 20?

Well, which is better: the illusion of privacy (which you can do nothing to further secure), or the known fact that you have no privacy in an environment where you can take steps to protect yourself?

Your bank offers the illusion of privacy. Sure, a random person can't just walk up and see your balance. But a teller can pull up your account and look around. Sysadmins can certainly see your money pop up in server logs. A particularly motivated individual might even be able to pull data from disk images or memory if they were so inclined, so how safe is that information really?

I'd argue that knowing everything is public is better. It motivates people to come up with solutions like Zerocoin and CoinJoin.

I looked that logo up thinking it probably came from some sort of 70's science fiction book about a dystopian future.

Nope, National Reconnaissance Office.

http://en.wikipedia.org/wiki/List_of_NRO_Launches

I mean, at least they're being honest!

Yup. I've gotten 1 parking ticket in my life: I parked fully within a clearly marked, paid parking spot in downtown Austin. A stop sign nailed to a wooden stand showed up while I was at dinner along with a $100 ticket for parking to close to said sign.

Or a technology already exists :)

"UC Berkeley scientists have developed a system to capture visual activity in human brains and reconstruct it as digital video clips..."

http://gizmodo.com/5843117/scientists-reconstruct-video-clip...

While I appreciate the effort that went into writing this thought experiment, that's not what I'm interested in.

I would like a historical example demonstrating the problem of the gold standard. The world worked this way for centuries, so surely there must be some documented case of one of the doom scenarios you proposed.

After trying several times to write this reply without sounding snarky, because I am actually being sincere: can you link me to reading on historical problems directly caused by the gold standard? Or maybe some search terms?

I've heard this assertion a lot in regards to bitcoin/deflation. However, everything I can find winds up in the same place. The gold standard was fine, but countries abandoned it so they could print endless money to fund World War I.

> if you can't do the time, don't do the crime.

That is entirely the point. He may well have been willing to do "the time". But "the time" is wildly disproportionate to "the crime". This is bad for the same reason that we don't execute everyone convicted of breaking the law.

Furthermore, how does your logic apply to the NSA? A group of someones very clearly broke the law hundreds of millions of times after being told to stop and being kicked off networks, then finding ever-more invasive ways of stealing our emails to grandma. Remind me, who is threatened with jail for the next billion (!) years under the CFAA?