laxatives

You can eat an avocado with almost anything and it its delicious.

All the complaints about the uncapped ICO aside, the biggest concern for me is the compensation structure. Breitmans et al make out big immediately. So much so that they have relatively small incentive to ensure the success of their coin. They have also demonstrated that they are both toxic PR people and incapable of responding to even the most innocuous line of questioning without going into personal attacks or sarcasm.

Its too bad. I first heard about Tezos when one of their devs posted an amazing interview question here on ycombinator that spun up an enormous discussion that was being refreshed all day in my browser. But their leadership is suspect. Their statements regarding their employment timelines (fulltime vs parttime dev work, 1 year vs 4 years) doesn't really add up either.

Hopefully they prove the FUD wrong, but I avoided the ICO because I can't accept that kind of risk.

I never said it starts off bad. DFW just makes a point to make the book inaccessible. It isn't meant to attract a wide audience, but many good things are initially dismissed because they aren't easy.

FWIW, I didn't like it for the first 300-400 pages either. But it very quickly became one of my favorite books. Its worth sticking it out to the end.

Inverse company marketing aggression.

Months and dozens. No leverage, no shorts. It just goes long and reallocates every couple of weeks. So it hasn't really sustained a serious bear market, but its also never had a drawdown more than 0.5% less than that of the S&P500.

I have a very simple algo (like 30 LoC) that is currently performed 9x S&P returns on >100k in initial capital. Granted, this is over a relatively short span of time and it doesn't observe the requirements to be part of Quantopian's contest. However trading live, it is doing significantly better than any back-testing I've done.

There's nothing preventing designers from putting their own money in. Contest winners get profits in addition to whatever they invest themselves, on the condition they share the algorithm (I think closed source?) with Quantopian. That being said, Quantopian probably has enough information to reverse-engineer at least the main ideas of anything you share, given they can observe the orders made and probably also your network requests.

Maybe a similar idea would be running those bots as evaluation tasks. I suppose voting would be a decent proxy objective when you are dealing with very qualitative results and don't want to spend a ton of time evaluating them yourself. Spin up a bunch of bots with different ideas and see which ones score the best.

There's a massive brain drain to look forward to in the next couple decades, especially in the US. Pretty much none of my best peers went in to a PhD program for many of the reasons you described. One of my roommates got his PhD in optical physics and is essentially living the life of a monk. Its so much easier to cash out for a pretty basic engineering role and get a lifetime (hopefully) of easy, stable income. Many skilled jobs are leaving the US (ie CGI leaving Hollywood for Canada and elsewhere) and political instability (ie immigration) is cutting out a big chunk of the upper quartile as well.

In addition, the average individual has most of their life decisions made for them via simple, accessible tools and the lower half has a fraction of the competence of their predecessors. I'd guess maybe 5% of my friends/peers spend any time reading. Maybe 60% over my friends outside of tech are seriously struggling financially.

Almost all of my coworkers are immigrants on O-1 visas and there's really nothing here for them long term, especially if they feel unwelcome. The moment another country provides a better offer, I bet 95% of them will go there instead. Probably many that are here now will go too.

I'm by no means an expert, but I believe you should ask for a 1099B with proceeds at $0 and cost-basis at either $0 and marked as not reported, or with cost-basis equal to your expense. If the former is the case (not reported), you report the cost-basis yourself.

This seemed inevitable. People lamented the growth of fake news and niche circles reinforcing extreme attitudes on the internet. Ads exist because they effectively target increasingly small and specific groups of people. Its concerning enough when the motive is profit, but the same techniques can be applied to much greater effect as propaganda. The real concern is when those reinforcing themes are applied to individuals, instead of large swaths of people.

Using Facebook, Google, LinkedIn, and other data aggregators exposes you to manipulation. It can be convenient when it shows you relevant ads and the motivation is clear (money). But by the time you realize you're being manipulated for greater causes, like electing officials to power, its far too late to revert those changes.

Strongly disagree. Maybe thats the case the a huge company, but most small organizations I've worked with are extremely top-heavy, filled with STEM PhD's who are very capable, but require 1-3 years to get a useful result and aren't often familiar with programming best practices or how to turn their results into a product. You need a larger team of engineers to make that happen and if there's a large overlap between engineers familiar with machine learning, that transition is much easier.

Furthermore, there's a number of practitioners that expect their data to be ready for them in some perfect state. Probably a majority of the task is create a pipeline for acquiring data and labeling it appropriately if necessary, which may require developing some ontology or classification with rigid guidelines such that someone in India can delegate the task to a large team. Then the practitioner spends an inordinate time optimizing some heuristic that has a meaning that drifts over time, or is completely inconsistent with the goals of the product. These are both problems outside the realm of domain knowledge or experience.

Whoa I thought John Markoff retired? Is he still active? He's been one of my favorite NYT writers.

The article consists of a list of algorithms with almost no other information. Its posted because it has a link to the OP's own blog.

I've had very similar experiences sailing. Sleeping in a bunk, you get accustomed to the swooshing sound of the water running against the hull and the drone of an engine. After a few weeks, I was certain I could hear the radio on and knew the tune being played. I asked the crew members who brought the radio and they thought I was crazy. There was no radio. It was uncanny, i thought I heard the song word for word, almost note for note, but it was all in my head.

Are patents really a great indicator of success? I have a handful, and honestly its a much stronger indicator of networking/coworker relationships/legal force of a company. There a number of key people missing on many of the patents I am co-inventor purely based on these relationships. These patents also would never exist, if not for the legal team at the purchasing company. If the organization went for an IPO or alternate exit, the odds of patents are probably close to zero.

This is nonsensical if you don't already have some existing notion of a relationship. These are vectors, not concepts. There a loads of operations that don't have any intuitive notion, yet are well defined in the context of vectors and most likely yield useless results.

9 women can make a baby in a one month. I used billions of data points to train my model.

LinkedIn's incredible algorithm was reading your contact book without your permission before those kinds of permissions existed or had any sort of granularity between none and complete access.

* Fiction *

The Alchemist -- reread one of my favorites of all time

Candidate -- reread

Infinite Jest -- takes 300 pages to "get" David Foster Wallace's style and 500 pages to enjoy it, but well worth the investment. Probably one of my favorites of all time now.

The Brothers Karamazov (in progress) -- Have been wanting to read more Russian Literature and apparently Infinite Jest borrows many plot points from Brothers Karamazov. I just started it a few days ago...

* Non-Fiction *

String Theory -- another book by David Foster Wallace containing a few essays on tennis. Even if you don't like his fiction, I doubt anyone would argue that the eponymous essay isn't great.

Open -- Andre Agassi's autobiography

Inner Game of Tennis -- reread

Winning Ugly -- Read this twice in 2016 (reread in progress). My mental game is volatile to say the least.

My System (in progress) -- Aron Nimzowitsch's Chess study

In a Sunburned Country -- Bill Bryson's description of his trip to Australia

Sailing Made Easy -- ASA 101/102

Coastal Cruising Made Easy -- ASA 103

Buying Your First Sailboat

Open Water Diver Manual -- PADI Open Water Certification

Enriched Air Diver Manual -- PADI Nitrox Certification

Adventures in Diving Manual (in progress) -- PADI Advanced Open Water Certification

Shadow Divers -- Diver's in the 90's discover a German U-Boat just dozens of miles off the coast of New Jersey

Triple Your Reading Speed -- It definitely works, but its only applicable to simple texts like popular fiction and maybe news articles/blogs. I guess it probably does still benefit denser stuff though.

The Origin of Conciousness in the Breakdown of the Bicameral Mind -- This is quite possibly the work of a crank, but there's some interesting ideas that seem not impossible. A lot of his arguments are clearly cherry picked and loads of sentences like, "The statue had a big ear, obviously an example of bicameralism".

Piano Handbook (in progress) -- Been working on the exercises in this and started learning the piano as an adult, two years ago. Just two chapters of eighteen to go...

Jazz Piano (in progress) -- Will focus on this once the above text is done

The Wine Bible (in progress)

Porsche: Origin of Species (in progress)

Millionaire Next Door (in progress) -- In strong conflict with the ideas presented in the text above...

* Technical Non-Fiction *

Functional Programming in Scala

Pro Git

Learning from Data

Optimization Models; Califiore, El Ghaoui (in progress)

Technical Analysis of the Financial Markets (in progress) -- Not great. Very little motivation to finish this one...

Had the goal of 30 books, at least 8 of which had to be textbooks. Not quite there, but there's some really dense stuff this year, so I'm not too worried. I had no idea how little fiction I was reading; will need to adjust that next year.

Maybe I'm ignorant, but reading the abstract/introduction,I immediately got the sense that this guy (Gwern) was a crank. At the very least, I figured it was some tangentially related philosophical quote, not part of the main body.

Mostly agree with you, but those apps can be handy when the page is formatted with 50 word lines, instead of say, 10 word lines.

OpenAI doesn't attempt to solve problems that even remotely resemble what you are describing. Hardly anyone claims to be working on AGI because AGI isn't a well defined problem. Its something cranks and the media tout because its excellent clickbait, along the lines of the singularity.

All the efforts of ML solve problems with a concrete input space and a concrete output space. This means they solve very discrete problems and are not generalizable, unless they have similar looking model inputs and the model has the same limited range of "moves" it can make.

There's an article about the writers hearing anecdotes they had to tone down for TV because it was too unbelievable or ridiculous. One that stands out was an angry exec storming out of a meeting on rollerblades, not being able to angrily slam the door without falling.

edit: whoops, businessinsider is shit. Here's the original article: http://www.newyorker.com/culture/culture-desk/how-silicon-va...

I tried like 8 queries and none of them yielded a single relevant result. Most of them didn't yield a video at all. Granted, they were difficult/obscure, but didn't make a great impression.

If this needs some special syntax or keywords, you should let the user know.

Manna is an interesting story, but its incredibly poorly written. It starts out great, but the last few chapters feel like they were written by a high school student realizing their deadline is approaching and is still 5 pages short of the minimum requirement.

I think its a strange distinction from most west coast folks. I remember at school it was pretty common for people to have huge courseloads who would generally downplay their stress/effort. On the flip side, my friends from the east coast thought it was some kind of rite of passage to talk about how their heavy workloads and how hard they were working all the time. Similar for coworkers from France.

Stuxnet would play AC DC Thunderstruck over the PA system at odd hours in Iranian uranium enrichment facilities.

Here's my favorite essay from String Theory that really sums everything that book is about: http://www.esquire.com/sports/a5151/the-string-theory-david-... This version is a bit of pain though since the footnotes are collected at the end and you would have to swap back and forth to read them; the book allows space on each page to make it much more convenient. I think I'm in the minority, but I really wish Infinite Jest handled footnotes in the same way.

If you have a turing-complete machine, you can execute anything you could on any other turing-complete machine (given sufficient resources).