I can't believe we're having an argument about the usefulness of ad supported web businesses with barely any mention of Google. Dear everyone, where do you think their money comes from? If there were no ads on the internet there would be no google and no facebook and no twitter, to name just a few. Maybe you can live with that, but it's hard to say that they aren't adding value or providing services that people like. The truth of the matter is that hardly anyone cares about ads, especially if they aren't obnoxious interstitials or flash spam and every once and a while you see something you like which is a win win. Run adblock if you like, but there's just no case for a crusade against advertising based revenue models. Those are the revenue models that support that suppor the services we love on the internet.

I don't know where you work, but the complexity of the algorithms we're using comes up every day at work. The difference is that you can't just implement the "assume without loss of generality" version of the algorithm; you have to actually make the damn thing work. And you have to handle possibly corrupt inputs, and up and downstream failures, and your problems span multiple algorithms feeding into and through one another.

So yeah, its complicated, and there's time pressure, but at good development shops people think about quality. It's just that quality is a much higher bar when you're trying to develop supportable systems for complex and evolving problems. What looks like ugly code to academics is, often as not, just hardened code, code that deals with the mess of disk failures and language-level memory leaks and all the detritus that piles up around our implementations of the turing machine. The big challenge is that the particular crap around the turing machine shifts over time, and no one has enough time to pay down technical debt all the time.

I think it's worth thinking a little bit about what the incentives are in this situation. Google's expertise, their secret sauce, the thing they pitch to the companies whose ads they distribute is that they have a done of user data AND they know what to do with it. The "and we know what to do with it" thing is a huge value add. Google knows much better than their clients who is the best target for a given add. That's their competitive advantage. They don't WANT to sell the data to other companies because that hurts their business model. That's why we can be confident that the Today policies are here to stay.

Couldn't agree more

Not that this helps the root problem, which is that a research lab has to blow its budget on lawyers to defend their work, but I have to imagine that the original research group published a paper either about the method, or using the method, since that's the primary deliverable for most research. If that's the case, they've almost certainly got a strong prior art case.

I'm curious as to what sort of limitations the building mechanism puts on the structures. There wouldn't be any supporting superstructure holding everything in place until the building is structurally sound. Every stage would have to be able to stand up on its own power. I wonder what the impact would be on design? Would buildings that are stable at every stage of their construction be more stable upon completion? Regardless, this seems pretty damn cool.

Based on their experience with the finger-crosser, I'd say this is a stunningly effective anti-terror mechanism

I don't think we're questioning whether or not she copied excerpts verbatim. The question is whether or not the entirety of the work is going to contribute to literature in a novel way.