aaronlevin · HackerTrans

aaronlevin

no profile record

Submissions

Show HN: An album art-based Wordle clone called UNCVR

uncvr.it

18 pointsby aaronlevin4 lata temu8 comments

SpecialGuest: Weird Marketplace for Zoombombing Artists

specialguest.best

2 pointsby aaronlevin6 lat temu0 comments

Leveraging iOS Frameworks to Speed Up Development at SoundCloud

developers.soundcloud.com

8 pointsby aaronlevin9 lat temu0 comments

Statement on LambdaConf 2016

statement-on-lambdaconf.github.io

15 pointsby aaronlevin10 lat temu7 comments

Thoughts on LambdaConf 2016

alissapajer.github.io

4 pointsby aaronlevin10 lat temu0 comments

Want to fix MySpace? Buy Bandcamp

aaronlevin.ca

1 pointsby aaronlevin12 lat temu0 comments

comments

4 lata temudiscuss

Alice Coltrane rocks.

4 lata temudiscuss

Something I’ve noticed that I didn’t expect is that people are able to guess a lot of albums in only one or two squares. Some of this is luck and image layout paradigms being surprisingly consistent (titles at the top!), but we also may have better context-based image memory than we think. There’s an obscure setting to increase the difficulty, which makes the squares smaller (or larger) to combat this.

Interesting idea about the reveal! I’ll consider adding that (or some cheap version of placing the image in the success modal at least).

4 lata temudiscuss

Author here. I was inspired to make this by thinking about how Wordle, a game where a word is slowly unraveled by guessing letters, could apply to images (and specifically album art).

Perhaps clone is a poor word here and inspiration would be more accurate.

5 lat temudiscuss

The main objection to computer-aided proofs is not that they are more difficult for humans to understand.

The main objection is that most proof-assistants use a different logical foundation than modern mathematics. Modern mathematics is built on ZFC[0] whereas most proof assistants, such as Coq, Isabelle, Agda, etc use different logical foundations, such as the Calculus of Constructions[1].

Many important results in modern mathematics are not easily stated or proven in systems such as CoC[1]. For example, Brouwer's _Fixed Point Theorem_[2], a pretty bog standard result in Topology that is useful to proof many things in Functional Analysis, has a clear statement in ZFC, but does not, to my knowledge, have an equivalent in CoC (and if it does it will be stated radically different).

[0]: https://en.wikipedia.org/wiki/Zermelo%E2%80%93Fraenkel_set_t... [1]: https://en.wikipedia.org/wiki/Calculus_of_constructions [2]: https://en.wikipedia.org/wiki/Brouwer_fixed-point_theorem

9 lat temudiscuss

SoundCloud employee here.

If you are hiring, please reach out to people from SoundCloud. The decisions of who to let go were not based on performance. A lot of amazing people, both attitude and technically.

9 lat temudiscuss

It's a combination of:

- getting dark around 4/5pm - extremely dry and cold - winter starts in October, lasts until late April - massive urban sprawl so the city core is dead during the winter, and therefore the city feels lifeless.

That's particularly bleak and definitely biased, but it's not far from the truth.

9 lat temudiscuss

University of Alberta has a very strong pure mathematics department (especially in Functional Analysis and Algebra). Pairing this with a strong Computer Science department could eventually bring it to a level to compete with Waterloo.

On the other hand, as someone who grew up in edmonton, I don't ever want to live there again. It's a tough city to live in, though it has its beauty when you look hard enough.

9 lat temudiscuss

One of the people involved in the reproducible builds project is a NixOS committee. Fairly certain they're aware of nix/guix

9 lat temudiscuss

SoundCloud | Senior Data Scientist, Machine Learning | Berlin, Germany | ONSITE, FULLTIME, RELOCATION OFFERED | https://soundcloud.com/jobs/2017-01-31-senior-data-scientist...

We are looking for an impact-driven Senior Data Scientist to solve exciting problems in recommendations, search, and personalization using Machine Learning. As a Data Scientist at SoundCloud you will work closely with other data scientists and engineers to find opportunities and build, tune, instrument, and evaluate machine learning models.

Some projects you might work on include:

   * a personalized playlist recommendation model to empower curation at SoundCloud
   * offline evaluation to improve the search experience of our users
   * prototyping deep neural networks and demonstrate their benefits in powering recommendations for our users
   * iterating on our Spark-based locality sensitive hashing implementations to improve recall.

Please see the ad for more details! https://soundcloud.com/jobs/2017-01-31-senior-data-scientist...

10 lat temudiscuss

I can't speak openly about the specifics, but we're definitely not in the latter category.

10 lat temudiscuss

Backend Engineer (Scala) | SoundCloud (Discovery/Recommendations) | Berlin | Onsite | Full-time | Relocation

I'm an engineer on SoundCloud's Recommendations team. We are looking for scala engineers to help users find lit tracks on the world's largest audio platform. You'll work alongside data scientists and engineers in a stack based on Spark, Scala, Finagle, and Cassandra.

SoundCloud is a wonderful place to work and Berlin is an amazing city to live in. I relocated from Canada and haven't looked back.

Empfehlungen sind die Zukunft!

job link: https://soundcloud.com/jobs/2016-08-04-backend-engineer-disc... blog: https://developers.soundcloud.com/blog/tracks

PS - ping me on LinkedIn (https://www.linkedin.com/in/aaronmblevin) so I can refer you!

10 lat temudiscuss

If you're interested in making deep-jazz more discoverable, consider applying to our Search team! :)

https://soundcloud.com/jobs/2016-02-19-search-engineer-berli...

11 lat temudiscuss

For those struggling with this, the last three paragraphs sum up the author's thesis pretty well.

It's actually more of a critique of current leftist ideas (from a Left perspective) than anything.

quote:

Getting less work seems unlikely to come about without the fight for solidarity, the chief intellectual achievement of the workers’ movement, and one that none of the accelerationists see fit to mention as an ideal worth preserving or even renovating. This is despite the fact that automation—or, more broadly, the increasing precariousness of labor through technologically assisted means—has always been dialectically connected with it. What tech enthusiasts call“disruption” is in fact almost always directed at forms of organization that preserve a modicum of workers’ control over knowledge and the products of labor. Because London taxicabs are controlled by people who have built up impressive maps of one of the world’s most complex cities in their brains, they ought to be replaced by self-driving cars operating on Google Maps. Because high school teachers have professional systems of accreditation and training and have unions to protect these, they must be replaced by lower-paid short-term Ivy-League graduates and cyber charters.

Automation isn’t a neutral, inevitable part of capitalism. It comes about through the desire to break formal and informal systems of workers’ control—including unions—and replace them with managerially controlled and minutely surveilled systems of piecework. An entire political and legal infrastructure has been built up to make these so-called tendencies seem like the natural progression of capitalism, rather than the effects of fights—sometimes simple, sometimes violent—to deprive people of whatever sense of control they have over their work. The only reason such work has ever not been totally shitty is that some attempt to preserve such control was made. This — not some implausible notion of a fully automated postwork future—still remains the surest of utopian impulses, the one most likely to deliver the things we want.

Though the times are bad, the accelerationists presume that the state of the left will get better simply because it can, in principle, get better. An imaginative (if implausible) account of the future, accelerationism is a weak account of how anyone might be persuaded to get there. We recall Lenin’s comment about how communism would be “Soviet power plus the electrification of the whole country.” The accelerationists seem to be telling us: forget about the Soviets, the electricity will do that work for us! But politics can’t get done by machines.

11 lat temudiscuss

This sounds very similar to warp, the Haskell web server that is known for its high-performance on many-core machines. The Web Application Interface (WAI) is also coroutine based (I believe): http://www.aosabook.org/en/posa/warp.html

11 lat temudiscuss

> Also it's hard to take an article seriously when it cites Marx, whose theories obviously don't work in the real world, as humans are competitive to varying degrees by nature.

Marx's "theory" was a method of analyzing history by focusing on class conflict as the main driver for history. By applying that analysis to capitalism, he predicted that the proletariat would inevitably come into conflict with the bourgeoisie. Das Kapital was a deep dive into how and why this conflict comes about (both economically and philosophically).

As for what happened next, those are "implementation details" which Marx didn't write too much about. We've seen a few implementations (Bolshevism, Stalinism, Cuba, Maoism) and a few modern contradictions (current China, North Korea, etc.) and most have been flawed.

However, not every implementation of capitalism has been successful, either. There are plenty of fascist and repressive states that are capitalist. Nazi Germany and Italy during WWII, Argentina during Pinochet, hell, you might even say ISIS is capitalist.

So, I'd argue that many implementations of communism have failed in the same way that several implementations of capitalism have failed, but there is a large space for experimentation as there has been with capitalism. For example, democratically planned economies with analytical input is a space that hasn't been explored too well beyond very tiny experiments in sectors within Latin America.

11 lat temudiscuss

Armin Ronacher wrote about this back in 2013. [My Favorite Database is the Network](http://lucumr.pocoo.org/2013/11/17/my-favorite-database/)

12 lat temudiscuss

"He is..."

12 lat temudiscuss

I did not say it was impossible to write Haskell with bugs. I simply stated that you are not guaranteed that a particular implementation of a design pattern adheres to the patterns semantics, therefore introducing a bug. With good usage of Haskell's type system, you can wholly avoid such bugs, and a great example of this, as I stated, is the associativity laws within the pipes library.

Semantics! :)

12 lat temudiscuss

twic: yes, if you take my statement literally you are correct: it is possible to introduce bugs in a Haskell program.

12 lat temudiscuss

You can encode a specification within Haskell's type system and get guarantees that are very difficult in other languages. Gabriel's pipes library is a great example of this (see a Pipe's associativity laws).