> Part of me even thinks it’s a more straightforward “data science”-problem when you ask the user what they want instead of playing the guessing game with tensors. Instead of worrying about an optimal recommendation, maybe there’s something to be said to allow users to customize instead. But why would bol.com ever consider this? Allowing for customization might be something that users are interested in, but it may cause them to customize for products with lower margins.
Optimizing profits at the expense of everything else is a surprisingly large part of much dysfunction on the internet. Significant parts of the internet are sponsored by ads and this business model creates perverse incentives for the people operating the infrastructure that serves those ads [1]. This is most obvious in social media but it's starting to be the case for Google as well, the quality of their search is continuing to degrade and more of their results are sponsored ads or SEO optimized sites instead of actually useful or pertinent pages [2]. I now mostly look for academic articles when I want to do actual research because most sites matching the keywords I'm looking for almost never contain any semantically relevant content.
We're not actually and assuming that is the case is why we're in the mess we're in right now facing an existential risk from global warming and potential ecological collapse.
Yes, that's the usual assumption when working with Turing machines and proofs. But I guess you could also allow infinite inputs and it wouldn't make that much difference, e.g. computing exp(x) for some real x as input.
I mean by looking at the source code for the neural network someone can give an upper bound on how many steps will be required before the entire network halts and gives an answer and they can prove that their upper bound is really an upper bound.
Unrelated to the article but I ran across an interesting result recently that is related to AI (and the hype surrounding it): Let A be an agent and let S be a Turing machine deductive system. If A correctly understands S, then A can correctly deduce that S does not have A’s capability for solving halting problems. [1]
One way to interpret this is that all existing AI systems are obviously halting computations simply because they are acyclic dataflow graphs of floating point operations and this fact is both easy to state and to prove in any formal system that can express the logic of contemporary AI models. So no matter how much Ray Kurzweil might hope, we are still very far from the singularity.
Really nice analysis of the non-viability of wind and solar:
> Such realities expose the silliness of the oft-repeated claim that solar or wind power have achieved “grid parity,” meaning that they can produce electricity for about the same cost per kilowatt-hour as a conventional machine—when they’re running. To match the energy produced by one conventional machine each year, and for years on end, you need at least two solar/wind machines, plus the batteries. That combination puts the sun/wind/battery option at roughly triple the capital cost of grid-scale conventional power. Even so, the cost for 12 hours of storage at U.S. grid-level alone would be about $1.5 trillion, and that would still leave the nation episodically in the dark. The alternative? A conventional grid with about $100 billion worth of conventional backup/peakers.
The only really sensible option at this point is nuclear and then maybe eventually hydrogen and fusion.
Software engineering is a form of process automation which by its nature is a form of productivity improvement. So if some teams are not streamlining and consistently improving business processes then asking another team to help them do so with more software is kinda nonsensical.
For a while I worked on build tools and CI pipelines and I can honestly say that I never improved anyone's productivity. Most of my work was figuring out how to remove performance bottlenecks that unwitting software engineers would invariably add to the build and test process because a manager was breathing down their neck about delivering some feature yesterday. If business processes are dysfunctional then that's not a software productivity problem and can not be fixed by adding more software.
Optimizing profits at the expense of everything else is a surprisingly large part of much dysfunction on the internet. Significant parts of the internet are sponsored by ads and this business model creates perverse incentives for the people operating the infrastructure that serves those ads [1]. This is most obvious in social media but it's starting to be the case for Google as well, the quality of their search is continuing to degrade and more of their results are sponsored ads or SEO optimized sites instead of actually useful or pertinent pages [2]. I now mostly look for academic articles when I want to do actual research because most sites matching the keywords I'm looking for almost never contain any semantically relevant content.
1: https://niklasblog.com/?p=25416
2: https://www.washingtonpost.com/news/arts-and-entertainment/w...