Hacker News Reader

torbjornon Ask HN: What maths are critical to pursuing ML/AI?

3blue1brown is great.

I also recommend Siraj Raval's Youtube course the Math of Intelligence: https://www.youtube.com/watch?v=xRJCOz3AfYY&list=PL2-dafEMk2...

torbjornon Storj: Not a Dropbox Killer

You can only assume these people are more about $$$ than what they are building.

Bingo.

torbjornon Statistics Professor Just Banned by Google: Here Is His Story

using google's "takeout" feature you can back up all of your gmail/google acct data.

torbjornon Store your zsh history into a sqlite3 database

anyone know of something like this for bash?

torbjornon Inside Darpa’s Push to Make Artificial Intelligence Explain Itself

You are right there is a large gulf between the deep learning theory hype and the measurable improvements in physical quality of life that the hard sciences have proved themselves capable of.

However I submit a notable exception! The vision system of self driving cars is a neural network that is currently changing the world and there is no lack of explication, we know exactly how it works. These "convolutional neural network" infer object boundaries from differences in intensity among matrices of pixels by scanning different patches of the pixel matrix and cross comparing these patches.

But you are right there is a lot of hype. I think part of the issue is the collective idolization of "algorithms". The limiting factor is data. And we don't have the data required to development models that are agent like in structure. Self interested agent systems, that employ subject-object language, have only been developed once. And they needed a training dataset measured in millions of years of evolution.

torbjornon Inside Darpa’s Push to Make Artificial Intelligence Explain Itself

I don't mean to come off as dismissive, and yes my comment is a sassy one. Machine Vision systems are proof that neural networks and other machine learning techniques are revolutionary. I really believe that.

But the notion of "Explainable AI" just reeks of upper middle management folks performing what they see as visionary risk management. When in reality these people don't know what they are talking about. Explaining exactly what's happening inside "the AI" is the domain of linear algebra. We can't be as exact in plain english but the generalization "it's fitting the training data" speaks volumes more than any _just_around_the_corner_ "explainable AI" secret sauce algorithm ever will.

torbjornon Inside Darpa’s Push to Make Artificial Intelligence Explain Itself

this is a pretty cool idea actually... if we just spent some time manipulating their chemical reward systems such that their amoeboid population densities opened and closed circuits on a piano key we could make them produce music. and i think the transference of this pattern would be enough to imbue them with the fire of thought! they would seek to control their collective amoeboid destiny! their will to be would drive them to express themselves, they would form an amoeboid super structure, spelling in english letters: "s e n d n u d e s". And AI is born.

torbjornon Inside Darpa’s Push to Make Artificial Intelligence Explain Itself

this is a fool's errand. today's statistical inference based "AI" is a probability distribution that squishes itself through cracks in data to "learn" a path of least resistance. subject-object reasoning systems, such as English, are cut from a different cloth. semantic language is a discontinuous function.

Asking asking an entity that can only discern its environment through continuous functions to explain itself is like asking an amoeba to play piano.

Machine Learning is a statistics renaissance, it is changing the world, but "Explainable AI" is a notion that is totally divorced from where the field is at atm.

torbjornon A practical explanation of a Naive Bayes classifier

I just did an jupyter notebook on Naive Bayes for Siraj Ravel's Math of Intelligence YouTube course.

https://github.com/NoahLidell/math-of-intelligence/blob/mast...

I used naive bayes to classify raps from Biggie and 2pac.

torbjornon There they go again, again [pdf]

Early adopters of technologies, platforms and financial products tend to benefit more than later adopters. Is everything a pyramid scheme?

(Assuming the product or platform succeeds rather than fails)

torbjornon Ask HN: What tools do you use to manage your life?

Every day I write 750 words on 750words.com, it keeps my content generation muscles strong and keeps me journaling. I have a script that maintains local copies of what I write.

torbjornon 100 days of algorithms

How did you enumerate this list of algorithms to do in your self challenge?

torbjornon How I learned to code in my 30s

mother earth is pregnant for the third time...

torbjornon Ask HN: What Podcasts are you listening right now and why?

Another great podcast from the Marketplace team is Make Me Smart with Molly Wood and Kai Ryssdal

torbjornon Ask HN: What Podcasts are you listening right now and why?

The History of Rome - An excellent 191 episode series covering, well, the history of Rome by Mike Duncan.

torbjornon Ask HN: Why TensorFlow instead of Theano for deep learning?

tensorflow has tensorboard, a great application that allows your to explore your models in depth. it makes neural networks less of a black box.

torbjornon Bitcoin is eating itself

How many millions around the world are lost in the form of dropped pennies no one will ever bother to pick up? Is cash eating itself?

torbjornon Ask HN: Anybody using Amazon Machine Image for AWS Deep Learning?

Siraj has great material.

torbjornon Thiel says he’s trying to get entrepreneurs to go after bigger problems (2014)

I completely agree.

The human tendency toward partisanship is holding all of us back.

torbjornon Ancient Romans depicted Huns as barbarians – their bones tell a different story

i nominate you to run for office

torbjornon Show HN: Detailed data on every cannabis retailer and producer in Washington

Americans are innocent until proven guilty in a court of law.

torbjornon White House Opens Door To Crackdown On Recreational Marijuana

They only care about states rights' when they don't control Congress or the presidency

torbjornon The Long-Shot Bid to Put Crispr in the Hands of the People

Is there some kind of letter to your congress person type thing that can be done for this? I would gladly sign a petition or write an email as part of a campaign to demonstrate public support for making access to crispr as open as possible.

torbjornon New NIST password guidelines: don't require character types or rotation

>Unicode [ISO/ISC 10646:2014] characters SHOULD be accepted as well. So, can passwords contain emojis under these guidelines?

torbjornon Debian Security Advisory: nginx security update

sudo apt upgrade

torbjornon Debian Security Advisory: nginx security update

So if you have nginx 1.7 or higher you're ok?

torbjornon We Got Phished

To use google authenticator you do not need to give out your number. Most sites will have a qr code you can scan with your phone and the google authenticator app uses that to generate 2fa codes that are valid within a certain time frame.

torbjornon The Myers-Briggs Personality Test Is Pretty Much Meaningless

I am always amazed by how many discussions of MBTI go on without any consideration of the Jungian cognitive functions that the 16 MBTI types are based on. Without seeking to understand the function stack associated with a MBTI personality type it's hard to use MBTI usefully/as it was intended to be used.

torbjornon Ask HN: What do you want to learn in 2016?

Go, machine learning, and dip my toes in various introductory computational biology/biochem topics.

torbjornon Amazon Provides DIY Echo Plans for Raspberry Pi

This is awesome. I am strongly considering getting setting this up as I just purchased a fresh raspberry pi.

The only limitation appears to be you have to click a "start listening" button to get it to start recording audio. You can't simply say "Alexa" to get the raspberry pi + alexa web service to listen for your query.

Anyone have any ideas for a work around/ solution to this?