Every week, we like to thank our most thoughtful commenters from the last seven days by highlighting their responses.

When Andrea Peterson posted an interview with Valve's Gabe Newell on Friday, she kicked off a discussion in the comments about the company's creativity as a software firm. Wrote reader Adam Coate:

I don't find Valve to be all that creative of a company. The only game they ever created that was their own idea was Half Life. Otherwise they bought out other people who developed a game that was already proven successful such as Counterstrike, Team Fortress, Portal, Left 4 Dead, Dota, etc. For a company that always harps about autonomy and such, they don't really allow anybody to explore their own ideas. Of course I've heard a big factor in their being risk averse is that since their employee salaries are heavily royalty-based, they have to make games that are guaranteed successes (aka sequels). For all the money Valve is making they really aren't doing much that I find interesting, it's a shame.

joe b replied:

The article was about their work culture, not their track record of bringing completely original game properties to market. Also, I didn't see where anyone was "harping" about anything. I did read a lot about how deeply Newell thinks about how to structure a company that is agile (embrace change; avoid bottlenecks and pigeonholes) and profitable (listen to your customers), while treating his employees like people (with kids, lives) instead of interchangeable machine parts. Sounds pretty great. Also interesting.

I recently drew a comparison between Netflix's film categories and BuzzFeed's headlines, arguing that the increasing use of data to microtarget audiences encourages ultra-specific (and slightly ridiculous-sounding) titles. In response, reader wwbrannon raised another line of analysis:

The really interesting possibility this article doesn't touch on, IMHO, is the possibility of automatically generating Buzzfeed-style articles from this sort of data. I can imagine a learning algorithm that mines your Netflix history and especially your recent Netflix history to pull together a) a headline like this, and b) semantically related GIFs with some generated text to make a listicle. Buzzfeed-esque, but automated and personalized to you and your recent interests.
It'd certainly be fascinating to try. Would love to see how much of what Buzzfeed does can be automated.

When Timothy Lee explained how Bitcoin could provide an alternative payment infrastructure to banking and credit card networks, reader Rassah summed up the security problems of modern personal finances this way:

Current banking system is inherently insecure because it's based on a pull system instead of a push system. You give someone access to your account, and they pull what they need out. On the internet, and in bitcoin, everything is a push system, where you have to actually send information out to get it to someone. Bitcoin is basically like e-mail, where you send value to an address, while the established financial system is the equivalent of sending an email by saving the message in your inbox, giving the person you want to send it to your email login and password with full access to your account, and hoping that they will only read the message meant for them.