Right, this sentence made it unclear to me whether they only used the popularity features to compute past-performance, or whether they included past-performance in addition to other popularity features.
To your question, other work on success prediction of tweets [1, 2] demonstrates that past-performance is indeed much more predictive than the typical content features. This way of looking at success of "cultural products" assumes it depends to varying extents on both inherent "quality" (measured by content features), and the social processes of sharing (which are much harder to understand ahead of time, as the paper I referenced in my parent post shows).
[1] Martin, T., Hofman, J. M., Sharma, A., Anderson, A., & Watts, D. J. (2016). Exploring Limits to Prediction in Complex Social Systems. Proceedings of the 25th International Conference on World Wide Web - WWW ’16, 683–694. https://doi.org/10.1145/2872427.2883001
[2] Bakshy, E., Hofman, J. M., Mason, W. A., & Watts, D. J. (2011). Everyone’s an Influencer: Quantifying Influence on Twitter. Proceedings of the Fourth ACM International Conference on Web Search and Data Mining - WSDM ’11, 65. https://doi.org/10.1145/1935826.1935845
Did I miss something or does this project include popularity measures as features?
In the section on dataset features, they include "popularity" (calculated by Spotify) as well as Billboard chart stats like weeks, rank, and a custom-made "score". To me it's not clear whether these features were hidden from the train/test sets or whether the popularity features were only used in their "artist past performance" measures.
If they included these popularity features, it's like asking "can we predict whether a song is a hit just by looking at how popular it is?" If it is the case that they peeked into the future and observed ex-post song popularity, obtaining just 89% accuracy hints at how unpredictable song success truly is. Check out [1] for a famous study of song success which experimentally demonstrates the unpredictability of song success.
[1] Salganik, M. J., Dodds, P. S., & Watts, D. J. (2006). Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market. Science, 311(5762), 854–856. https://doi.org/10.1126/science.1121066
I was just thinking that something similar to this would be really interesting for reading/peer-reviewing scientific papers. Scientific knowledge is built atop of claims made in previous papers, each claim having its own statistical strength. It would be awesome if each footnote in a paper was not just a citation to another publication, but a link to the first time that claim was made, with its p-value and sample size prominently displayed.
Totally agree, this is just about employing affirmative action, how, and why. And yes, I get the criticism but I haven't heard anything convincing as of yet. The only baffling thing to me is that its critics believe that pure meritocracy is still the flagship solution, when the flaws seem so obvious. I think the author does a great job explaining why for this case.
This blog post is good evidence of the quality of thought that exists on the political left, and revives my belief that the left will eventually win the misguided attacks on “political correctness” culture. The comment section reinforces this even further.
This is really vague. Javascript is necessary for building rich web applications. What are the uninteresting problems you see as Javascript's main purpose?
Normally I would agree with you completely, indeed AI is (very) far from perfect. However, I think we're hung up on the home property example. As Vitalik suggests, this proposal works a lot better in situations where the underlying property is more fungible, and the participants in the market are more or less equal. The property example I like from the article is radio spectrum licenses.
I'd also add that using an AI also doesn't necessarily mean something opaque. A hand-made decision tree would be preferable to a neural net, for example. The decision tree could prompt you for your human decision if it encounters an outlying, uncertain instance.
Recall there are two goals:
- allocative efficiency (property belongs to whomever can derive most value from it)
- investment efficiency (people have incentives to invest in property, build Gigafactories, etc)
If I understand, your argument is that self-assessed taxes would eliminate investment efficiency.
The property owner should factor transaction costs into their own value of the building. This would increase their assessment of the land-value, and increase their taxes. This works because, keeping land valuations constant, the one with the highest transaction costs is the one who can derive the most efficiency from the land.
I think the property/personal home example is hard to defend. Consider industrial property, such as a factory or warehouse. The most reasonable application of self-assessed tax would be to any property that can be valued on purely economic/objective bases.
"... people are not experts at property valuation, and would have to spend a significant amount of time and mental effort figuring out what self-assessed value to put for their house, and they would complain much more if they accidentally put a value that’s too low and suddenly find that their house is gone."
The solution: users would choose an AI to generate real time valuations.
In the property/personal home context, I think this is a bad idea, because if the AI gets it wrong, and you're forced to sell your house, that is bad. However, this idea could work very well in other types of property markets (this was all discussed in the article).
Part of the introduction really reminds me of Nassim Taleb's Antifragile.
"As artificial agents enter the human world, the demand that we be able to understand them is growing louder."
"Let us stop and ask: what does it actually mean to “understand” another agent? As humans, we face this challenge every day, as we engage with other humans whose latent characteristics, latent states, and computational processes are almost entirely inaccessible. Yet we function with remarkable adeptness."
In the chapter I'm referring to, Taleb describes situations where practitioners outperform academic theories. Practitioners are driven by results, they develop heuristics and mental tricks with experience, but they don't always have a good understanding of the underlying complex system that they're predicting about, and they can't usually explain their heuristics and tricks.
Academia (and broader society) tends to demand an explanation for why things behave a certain way. Usually this means developing some nice-sounding narrative that may or may not be true. I think it's fair to say Taleb is skeptical that it is at all possible to learn the underlying mechanisms of a complex system to the point where we can actually predict the future behaviour of that system.
That's a great point, and the nature of the firm is something worth thinking about especially in this context.
However, the "firm" structure creates lots of efficiencies. Individual contracts are costly to enforce and having one type of contract (the employee contract) allows hiring to scale a lot better. Employees (and usually not contractors) are exclusive to an employer which is also desirable for employers. Employees also get benefits, which a lot of people like.
There's also efficiencies gained by people working in the same room. Any type of creative work for example can benefit greatly from a quick in-person chat, or just bouncing a quick idea off of your colleagues. This is even true for some developers (myself included). Some types of work truly rely on communication between colleagues, and that would be really hindered by latency issues due to low frequency async communication. Coordination is also very important and much harder to do with async communication.
Tribalism is an interesting lens, and maybe it started out that way, but the firm is certainly valuable.
Hopefully someone can expand on these comments since I'm not an expert, but I know a little.
D-Wave [1] is shipping quantum computers with thousands of qubits. The difference seems to be that D-Wave qubits are not generalized qubits, but rather they are simply designed to solve problems using quantum annealing [1]. Annealing is just one algorithm which efficiently (but probabilistically) looks for the global minimum of a function, but it is prone to getting stuck in local minima. Quantum annealing exploits quantum phenomena to do this better and probably more efficiently [2].
On the other hand, we have "general" qubits: these can implement and carry out any arbitrary quantum computing algorithm (for example, Grover's search algorithm or Shor's factorization algorithm). It seems to me that researchers at Alphabet, Rigetti, Intel, IBM, etc. are trying to build these general qubits, since that's where we can finally unlock the full power of quantum computing. It's also much harder to build these general qubits, and that's why we only see 17 qubits announced here, or Google claiming to be working on 49 qubits [3].