pacala

48 ms·

451.
by pacala 8y ago
> Google isn't trying to control or limit what people think. Advertisement is precisely trying to control what people think. Specifically, influence their economic and political decisions.
452.
by pacala 8y ago
Names have a small, but pervasive, influence on economic outcomes. This was/is well known and name anglicization was/is a mass phenomenon. See for example, http://ftp.iza.org/dp7725.pdf or https://en.wi
453.
by pacala 8y ago
Also consider alternative perspectives on IAT, for example https://www.psychologytoday.com/us/blog/culture-conscious/20... . Inline excerpts: > The test-retest reliability (repeatability) of the Race IAT is
454.
by pacala 8y ago
Regrettably, there is no such thing as post-scarcity society. The exponential nature of biological reproduction ensures that no matter how much carrying capacity one builds, it will be filled to the brim in a geological blink of an eye.
455.
by pacala 8y ago
My fear is the scenario is all too sustainable. North Korea is at the third generation of Kim's and its political pyramid appears to be quite stable.
456.
by pacala 8y ago
For whom it may concern, would be nice to consider autobatching [0] of DyNet fame when designing AD systems. Makes the world of difference for NLP applications. [0] https://arxiv.org/pdf/1705.07860.pdf
457.
by pacala 8y ago
pylint + typing might do the job.
458.
by pacala 8y ago
Of course they will.
459.
by pacala 8y ago
Fortunately [0], it doesn't quite work, [0] https://en.wikipedia.org/wiki/Censorship_of_images_in_the_So...
460.
by pacala 8y ago
Fantastic! > The code and pre-trained model will be available at https://goo.gl/language/bert . Will be released before the end of October 2018.
461.
by pacala 8y ago
Google had a very reasonable money printing machine as its core business model: Search Ads. * You tell Google what you are looking for. * Google returns a handful of ads, alongside world-class organic search results. * Merchants bid to have
462.
by pacala 8y ago
> supposing I see this image, what are some likely scenes (positions and shapes of objects) that might have produced it? Sounds like there exists a device where I can write some program, run it over some inputs, pass the program and the
463.
by pacala 8y ago
There is barely any data in the article, none stating the number of white CEOs, and none stating the evolution of said number over years.
464.
by pacala 8y ago
Exceptions are for error reporting. Pass them through all the way to the top, and log them. Enforce module preconditions proactively. if(!isInstalled(Foo)) { purchase(Foo) } else { start(Foo) }
465.
by pacala 8y ago
A ball of iron the size of Earth. I'll show myself out now.
466.
by pacala 8y ago
In other news, there goes Snap’s market value. Time to short.
467.
by pacala 8y ago
It is definitely cool to mine asteroids. Judging by https://en.wikipedia.org/wiki/Asteroid_mining#Potential_targ... , it is also a bit weird to setup such an sophisticated enterprise to mine for elements that are readil
468.
by pacala 8y ago
Go was designed for data munging via gRPC.
469.
by pacala 8y ago
It depends on what the meaning of the word "equally" is.
470.
by pacala 8y ago
Fair enough. N:M surveillance is worse than N:1 surveillance from a personal perspective. Though unclear if that also holds at scale. They are still both evil and much worse than no surveillance.
471.
by pacala 8y ago
While it is technically correct that ad networks don't sell personal data, it is also irrelevant. It's not the bits that matter, it's what one does informed by the bits. Ad networks sell _actionable insights_ informed by pers
472.
by pacala 8y ago
The big ones are in the business of surveillance. Do not want any of them anywhere me or my car.
473.
by pacala 8y ago
Jupyter code is draft code. Draft code rarely makes into clean code, because it requires extra tedious effort. Most of us are too overworked to have the time or energy to perform it. Alternatively one can write exploratory code as unit test
474.
by pacala 8y ago
We could blissfully ignore that Little Green Footballs / Daily Kos / etc. ever existed and go on with our lives. Enter Facebook / Twitter and their hunger for "engagement" and all of a sudden inflammatory political
475.
by pacala 8y ago
The pytorch graph model is not as dynamic as you think it is, because batches. All batch elements must be processed by the exact same code trace, limiting per element dynamism. The dynamic aspect of pytorch is an implementation detail for d
476.
by pacala 8y ago
IP
477.
by pacala 8y ago
> Knowing where you are is the only way this feature can work. I'm 100% confident that IPhone weather app updates automatically the weather for my home town. It doesn't need my personalized location, just the location for my ho
478.
by pacala 8y ago
Can we rename it to "<Bernard L. Madoff Investment Securities LLC> product poorly named"?
479.
by pacala 8y ago
For weather, I use 2 mechanisms: a. I setup IPhone weather app to my home town location. Works 95% of the time, I'm not traveling that much. b. Search for "<town> weather" when I'm traveling. Browser autocomplete m
480.
by pacala 8y ago
Read the article. What are the "clear descriptions and robust controls" to turn off the following Google location tracking behaviors: > Google stores a snapshot of where you are when you open the Maps app > Automatic weather

More ›