jjn2009

8 ms·

31.
by jjn2009 7y ago
Seems like that information is already largely public just not linked publicly to your phone information.
32.
by jjn2009 7y ago
Seems like the only thing uploaded is your picture. I'd be curious to see what data it actually collects from your phone otherwise but just a photo feels pretty innocuous
33.
by jjn2009 7y ago
It really depends on how much control on the life cycle of state exists between the user, program and operating system. More to the user, easier to press reset and clear the state. More to the OS then it needs to have some clear guidelines
34.
by jjn2009 7y ago
Its partially because they are in a war of attrition.
35.
by jjn2009 7y ago
These issues are entirely caused by NPM in the first place. Anyone who starts a node project ends up with a massive web of unaudited code which can only be dealt with my moving off NPM and adding packages to your own private repo after youv
36.
by jjn2009 7y ago
I was forced this morning to have Comcast charge me $40 to install a cable to set up internet at my new place by an authorized technician. I hate Comcast SO much, and from what I can tell I don't seem to have another option, supposedly
37.
by jjn2009 7y ago
Understanding the uncertainty of a decision is important to make good decisions however, software massively benefits from first mover advantage so there is a reason why absolutism (even when usually wrong) is actually beneficial. For every
38.
by jjn2009 8y ago
> it is valuable to me that my company will not hire someone who doesn't think our business does something important Is it not possible to identify those people without paying them less to prove it? I'm not really sure what you
39.
by jjn2009 8y ago
It depends on how well the problem is understood. If the problem is your standard MNIST dataset then sure it could very well be a waste of time to sit around and serialize your manual hyper param search. For any new datasets which may or ma
40.
by jjn2009 8y ago
"hyperparameter descent by grad student" is a lot more efficient at first, much of the time the loss function has caveats which make it easy to fall into parts of the search space which don't actually accomplish the task (for
41.
by jjn2009 8y ago
Nice. I've been using hyperopt with random search, I'll definitely check out your work.
42.
by jjn2009 8y ago
It sounds like an easy way to increase performance but really exploring the hyperparameter space is likely more efficiently done manually first and only automatically when you have figured out how to distribute the work.
43.
by jjn2009 8y ago
Measuable growth. If something is working it will be self evident if youre looking in the right places. The airbnb way is probably nonoptimal.
44.
by jjn2009 8y ago
I agree, as long as you do out of sample testing correctly then using this as a benchmark for finding the correct transformations to perform should reduce overfitting. That said I can see how if done incorrectly you could easily end up over
45.
by jjn2009 8y ago
>whole sample to determine bins, including validation sets. well thats not random or arbitrary just wrong. >And in the end you want to invest based on the model. I'm not quite sure what you mean here. Either you determine a model
46.
by jjn2009 8y ago
It's arbitrary if you make it arbitrary. Depending on the data you might target various time intervals to find the optimal way to trade on that information just like any hyperparameter and the bins are a function of that time, on avera
47.
by jjn2009 8y ago
With financial data you might be better off separating out the side of the bet from the size of the bet and thus it makes sense to just make three bins. The features to use for direction and volatility are different.
48.
by jjn2009 9y ago
I can understand the way people feel about blockchain but a backlash against what boils down to mathematics is a futile exercise.
49.
by jjn2009 9y ago
I would guess these types are projects are ones done in 20% time at google.
50.
by jjn2009 9y ago
by solidity do you mean EVM?
51.
by jjn2009 9y ago
2017 was the biggest year for blockchain by a huge margin by many metrics (see https://coinmarketcap.com/charts/ ). there is now 200 billion dollars in this space, give it another year or two. Some things will shake out
52.
by jjn2009 9y ago
I see, that makes much more sense. I wasn't entirely sure what was meant by that statement.
53.
by jjn2009 9y ago
>In Morgan’s view, the key part of the Howey Test in terms of ICOs is whether the expectation of profit relied on the effort of others This makes sense and is precisely how its been done, nothing new here but.. >What [the SEC are] try
54.
by jjn2009 9y ago
For the time being its going to be more of a "I'll know it when I see it and so will you" type of situation. Preventing scams and things advertised as investment opportunities appear to be the focus without trying to pin down
55.
by jjn2009 9y ago
This seems pretty reasonable and is the same conclusion which the DAO investigation came to. I think both ICO creators and participants will benefit from this new information now that we have a clear line in the sand as to how these things
56.
by jjn2009 9y ago
I'm telling you in plain mathematical terms that the way you go about hiring effects the distribution of candidates skill if you introduce any bias into the system regardless of whether or not its a literal or figurative bar or a bias
57.
by jjn2009 9y ago
Lowering the bar for a cohort is going to create different distributions of skill in that cohort, this person wasn't the first and won't be the last to come to this conclusion. I think damaging organizational cohesion is never the
58.
by jjn2009 9y ago
Extending the chain of trust down to the source of information isn't exactly going to happen at a snap of a finger. If you trust TLS and trust what the server gives you then today that is a fine option, the server is signing a piece of
59.
by jjn2009 9y ago
this is actually incorrect you can verify a tls connection was made to a server and a particular output was given in that tls connection with oraclize it.
60.
by jjn2009 9y ago
Its computationally difficult to calculate every possible move a character might do in the occluded region. This requires a long attention span and running simulations of possible opponent actions based on previous 'frames'. Alpha

More ›