We don't know how to program...

I have to say that this blog post is a fantastic summary of a lot of the challenges to concurrent programming past and present.

I sort of wish that there was at least brief discussion of things like software transactional memory or compare and swap

http://en.wikipedia.org/wiki/Compare_and_swap http://en.wikipedia.org/wiki/Software_transactional_memory

I liked the article, specially since it comes from someone with long experience in the subject.

But what it seemed more interesting is the company. Specially here in HN it seems that many start ups are focused on social networks or other consumer applications. This guys are basically selling technology. They seem to have found an interesting niche.

A great post. I remember the criticism of Java circa 1997. It has come a long, long way indeed.

Tangentially, this is a reason why many (very smart) programmers stick to "blub" like Java. In practical every day reality, the investments in VM improvements, libraries & tools play a huge role in productivity. As an example, most modern IDEs (really lots of macros in disguise!) take the pain out of most of the verbosity while writing code, and such verbosity actually is a plus for reading code.

> In short, the easy pickin's have long gone, and now we need complex tools and complex coding styles to get more performance from more cores using the existing languages.

I don't think more complexity is the answer here. All this massive multithreading just makes things impossible to debug.