ddrum001

Blog post on how this was built: https://blog.insightdatascience.com/elastik-nearest-neighbor...

Curious how this would affect private ventures like SpaceX?

Fair, perhaps the cursive analogy is too strong, though I'm sure that cursive still has a role in research of historical documents. IPA still has an important role in specialized situations, but I don't quite think it's something that should be known for the general public (e.g. the average wikipedia reader).

I guess I'm an anomaly in that I read far more books than watching movies, but I agree that most consumers will want a variety of music and shows, unlike books. However, isn't the price point ($35/month), which is 4x more than Spotify, roughly in line with the price for a single book, which is roughly 4x an album.

I guess the main difference is not that the price ratio between subscription and individual items is off, but rather that most consumers don't want 1 book a month. It's a bummer that they don't give people an option, but I'm still a fan of Safari...and I hope this move will mean they drastically improve the Safari app, much like Netflix has doubled down on Streaming now that they don't do DVDs.

Having purchased many books thru O'Reilly, and now using Safari Online - I'm a huge fan of this move. Feels to me like going from iTunes to Spotify, or Netflix DVDs to Streaming.

This is a great feature that has been missing for so long. Can't help but feel that IPA will go the way of cursive once tools like this propagate.

Thanks, we've included it in the Unified Batch processing because of its ability to have large window sizes that enable batch processing.

Fair point - the set of technologies is based off the teams we work closest with, which admittedly have a bias towards open source and Linux. So far, our map is far from comprehensive, so appreciate the suggestions (exactly what we're looking for by show HN).

To that point, just added CosmosDB, and plan to add others soon.

We put Dataflow and Flink in the "Unified Processing" since they can handle batch and streaming (as opposed to tools that only handle steaming).

We might add them explicitly to Streaming as well though.

Thanks, we tried to keep it simple and streamlined. If you know of any features or common patterns that would be helpful, let us know.

There may be a paywall issue. Can someone please post another related story?

Is this post equating trolling with bullying?

I'm glad that Microsoft is throwing this "hail mary" - topological QCs could be much trickier to build initially, but they scale much better since they don't suffer from the same type of decoherence issues as others.

For me, the grid of faces is much scarier than each on individually...interesting.

We don't have a fixed syllabus since we don't operate like a school per se. Instead, the Fellows choose a project and have to make engineering tradeoffs about which tools to use. With that said, most of the past participants focused on distributed systems like Hadoop, Spark, Flink, Kafka, and NoSQL DBs. You can check out past projects on the blog (http://insightdataengineering.com/blog/) and the past Fellows are here (http://www.insightdataengineering.com/fellows.html)

Interesting as something to explore, but I'm curious if a network of self-driving cars can be optimized to recognize and avoid these situations altogether.

I'm from Insight - our program is kept free for our Fellows because the companies that we partner with pay for the Fellowship. Rather than classes, we believe that the best way for advanced engineers to learn the detailed nuances of these tools is to use them, so the concept is to learn them by building a data platform. You're right that it's really difficult to understand distributed systems in a few weeks, but our Fellows already have several years of programming, so they have been able to pick up a general understanding quite quickly.

Not to mention the great push-based features!

Very exciting news if it holds! Can someone discuss how close the quasi-polynomial complexity is to P - that seems to be the really interesting detail that gets covered less.

An interesting part that the article didn't discuss as much was how to distill all mistakes into the finished product. It's one thing to say artists who make 50 lbs of ceramic pots have a few good ones, but how often are those pieces presented as the best by the artist, or iterated on.

In other words, deciding which projects to continue pursuing and which should be scrapped entirely is non-trivial.

A little passive aggressive (better than the legally aggressive approach of Axel), but pretty interesting. Not sure if I would qualify the eschewing of an AdBlock work-around as Big Brother, but the legal precedent is worth discussing.

This seems very odd to me - isn't this regulated or standardized?

Indeed, the script works quite well for Spark - but we also wanted to provide a guide for those who would like to further understand the config files, especially for those who want to 'tinker' with these down the road (e.g. adding nodes to the cluster). Also, we find it helpful for learning to set up Spark on other platforms, and setting up other systems that don't have ec2 scripts yet.

Thanks...makes sense now.

I don't think there is much to discuss here, tau=2pi is clearly the more sensible constant for almost every use case. The only question is whether we give in to conventions, or pay the one-time technical debt of switching.

We (in the US) may never convert to metric, but the much smaller community who works with pi regularly should know better.

That was a great story, and it speaks to the fact that tropes like prison escape, bank heist, and marooned on an island appeal to some human instinct

Could the OP or the Github Readme explain why this is called Ex-Machina...

Agreed, that was my reaction as well. I suppose that's the tradeoff of a descriptive name for your project.

Very interesting post and seems to be the best way forward for "marketing" and creating the right culture at a start-up.

...or just use a language where a cookie cutter isn't needed e.g. macros in Clojure