dangoldin

Take a look at https://gitcoin.co/

No affiliation but listened to a podcast by the founder a few days ago.

Just yesterday I finished up a biography of Claude Shannon that I recommend to anyone that ones to learn more. It briefly covers this as well as ton of his other "toys" - https://www.amazon.com/Mind-Play-Shannon-Invented-Informatio...

Agreed but in this case I took a stab at downloading the buddy list and with a bit of tech know-how it's not terrible: http://dangoldin.com/2017/10/09/downloading-your-aim-buddy-l...

Years ago I took a stab at writing my own reader with the goal of looking at the top HN stories over the past 24 hours since I'm not one that checks HN every few minutes. Pretty limited use but surprisingly still works: https://yahnr.dangoldin.com/

All sorts of engineers | TripleLift | New York, NY | Onsite | US Citizen

We're looking for software engineers to come join us and improve the state of digital advertising. Our core product is a native advertising ad exchange that's running tens of thousands of auctions each second while having to deal with a variety of challenging performance and scale issues. The team is lean and likes to move quickly so if interested definitely reach out and let me know! dgoldin@triplelift.com

This reminded me of a book I read last year that takes about the role of work and poverty in a poor, anonymous town in California. Definitely worth reading if you're interested in this topic: https://www.amazon.com/gp/product/B0043GWQ22/ref=oh_aui_d_de...

Seems similar to what happened with Larry Summers - also affiliated with Harvard. http://www.thecrimson.com/article/2005/1/14/summers-comments...

I actually converted my blog (http://dangoldin.com/) a while back to be entirely in AMP. In this way I'm using the framework without caring about the CDN element.

After reading this I may just bite the bullet and remove AMP and just cut out a bunch of crap to get it optimized for lower bandwidth devices.

A bit of a plug but I wrote about this a week or so ago describing SQL as the perfect interface. Databases change and evolve but since they all wrap the underlying engine in SQL it becomes very easy to use new technologies under the same interface: http://dangoldin.com/2017/04/11/sql-is-the-perfect-interface...

It makes sense for the simple case but I can see it getting a bit hairy if you're doing a variety of joins.

Then it would be FROM X join Y on .. join Z on .. join (..) as T SELECT X.A, Y.B WHERE ..

I guess it's jarring but that's due to me not being used to it. Some people prefer doing joins using there where clause in which case the FROM ends up being in closer to location to the WHERE.

Why not write about the company in the abstract and not mention names?

I wish his behavior led to some side effects on the justice system.

Very cool. The partitioning reminds me of the way VoltDB does it. By splitting it across a set of keys it's able to ensure that queries can be run concurrently on their own node.

Are you guys familiar with it and I am understanding this properly?

Definitely respect that but a bit too extreme for me. Have you never had a vendor reach out with a potentially useful service?

Occasionally I'll find something useful. And sure I could have found it on my own if I was searching for that specific problem but occasionally something new comes across that piques my curiosity. Even then I end up looking at competing products (love relying on Google's autocomplete to do "product vs" and seeing what shows up).

The spam game is getting better and better too - I wonder what they're going to look like in a few years with some help from ML. At that point they will look personal.

I lead an engineering team and am on the receiving end of a ton of vendor pitch emails. Almost always they go ignored but nearly every single one will have a few follow ups and surprisingly (but rarely) I may end up responding at the 4th of 5th attempt.

I realize that they're likely automated but sometimes you just need to get me at the right moment in which I'm more likely to respond.

Yea - it does seem a bit high. We use Spark for our adtech data pipeline and we're handling tens of billions of events a day in less time. It may be a function of how much data they're pulling in from other systems or dumping the data back into a variety of systems. Spark itself is parallelizable so in theory can be sped up just by running more nodes.

Really interesting read. We've been in a similar mindset and it really comes down to understanding what's driving your costs and being creative around ways to minimize it. As others on here mentioned bandwidth and data transfer is a huge cost so if there's a way to make your caching more aggressive it's a big and easy win. Beyond that the low hanging fruit is in using reserved instances - there are a few options there but are significantly cheaper than paying the on demand price.

A shameless plug is I built a small visualization for the AWS detail log to help understand the cost drivers. All it is loading the data and visualizing it across a few dimensions but might come in handy for others - https://github.com/dangoldin/aws-billing-details-analysis

That was one of my first forays into coding. I remember doing the basic examples and then finding some of legitimate worlds and having my mind blown.

The good old days!

Definitely true! You also have the luxury of tweaking every line of code for a small system to get it perfect.

It's also the reason for trying to break down larger systems into smaller, elegant components.

One of my favorite simple examples of what clean and elegant code can be is Peter Norvig's spellchecker: http://norvig.com/spell-correct.html

It's in Python and a single file but it comes with a wonderful description and shows how a complicated task can be broken down into a few small and powerful functions.

TripleLift | Software Engineers | New York | ONSITE, REMOTE, http://triplelift.com/

We're looking to hire for a variety of roles ranging from fullstack engineer (PHP, Angular), to data engineer (Spark, Java) to a web server engineer (Lots of Java + Netty), to a data scientist (Spark + streaming), to an SRE, to a QA. We're exploring the idea of remote workers but are still figuring out how that would work but my initial gut is that we would want someone to be fairly experienced for us to go ahead with it.

A bit about the company: TripleLift is an adtech company specializing in native advertising. We're trying to get publishers to move away from the idea of commoditized banner ads so that they can focus on content first and monetization later. We do this by giving them the ability to drop a JS tag wherever they'd like an ad (typically in the main feed) and then rendering an ad that matches the look and feel of their site. There are quite a few technical challenges getting this to work at scale and I'd love to share more so please reach out - dgoldin@triplelift.com

I had some issues getting the Wallabag app working on a smartphone. Have you used that successfully?

Very odd and surprising it's taking that long to recover. I got the following email from them this morning which seems to indicate they will only be fully normal a week from now.

As we shared earlier today, Instapaper is experiencing an extended outage. After 31 hours of downtime, we were able to rebuild a database instance to get Instapaper back online! In the interest of coming back up as soon as possible, this instance only has the last six weeks of articles. For now, anything you’ve saved since December 20, 2016 is accessible. We’re working on getting the rest of your articles restored, but wanted to give you access to your most recently saved articles in the meantime. Again, there hasn’t been any data loss. The full database exports are taking much longer than anticipated. We’re aiming to get the full archives available by next Friday, February 17 at the latest. We appreciate your patience and understanding in the interim. If you've got any questions or thoughts, just reply to this email or let us know at @InstapaperHelp on Twitter. - Instapaper Team

Shameless plug but I did something similar in 2014 and used PhantomJS to analyze the content of the top 1000 Alexa sites: http://dangoldin.com/2014/03/09/examining-the-requests-made-...

Part of this can be explained by Snapchat not having a self serve product (AFAIK) so a lot of the search results might be individuals/smaller businesses looking for help/advice to get started. I know the major brands are spending fairly significantly on Snapchat and doubt they contribute much to the Google trends data.

It's nice if you're looking for an alternative to a paid columnar database (ie Vertica, Redshift). I used it on a side project and it performed surprisingly well for large queries (think sum/group by across tens of millions of rows).

Not sure how it works on a production system.

Same here. My bear case for Uber is just that self driving will commoditize the market and turn a two sided market into a one sided market. At that point it will be the cheapest offering that wins. Wrote about it in more detail here: http://dangoldin.com/2016/05/28/uber-and-self-driving-cars/

It's true. Google has AMP and Facebook has Instant Articles. They're both about keeping the content inside their walled gardens that they can both improve the performance of (especially on mobile) but I'm sure at some point use their own monetization solution.

Yea - I was inspired by seeing how well AMP worked on some sites and actually converted my entire blog to use AMP for every page. It's incredibly fast and I'm happy with it although I'm sure I could have achieved most of the results with some highly optimized CSS and HTML: http://dangoldin.com/

The pickup and delivery costs don't make sense given the high cost of labor. In India it's significantly cheaper so good to prove the concept out there - at least until drones arrive.