Can we agree to stop writing phrases like this: "Not only do I not think that the Alignment Problem is impossible/hopelessly bogged-down, I think ..."? The three negatives are a semantic mess. Some of still using our wetware here for decoding prose.
Perhaps "Not only am I still hopeful about the alignment problem, but I think" or even "I don't think the alignment problem is hopelessly bogged-down, and I think..."
My understanding is that 'blockchains' are, today, pretty centralized, especially effectively (i.e. in practical ways that matter). Any particular 'blockchain' also seems possibly 'lost' (e.g. abandoned), 'made obsolete', etc.. Isn't there any inherent and inevitable 'social expense' that has to be continually made anyways?
Not really. The blockchains themselves are decentralized, but the applications built on top of them are centralized. To my buying a movie example, suppose I have an app on my SmartTV called "stream1" that lets me buy movies and register them via NFT (cool). I would prefer that to buying them from say Amazon Prime which did not have NFTs, even if it cost an extra dollar or two. 5 years later, stream1 shuts down. I'm not saying I would actually go torrent the movie. What might happen instead is that a new company starts up called "stream2" which allows you to play movies you've bought in the blockchain, as well as buy new movies. Or maybe Amazon prime has added the feature by then and I can connect my licenses to their service. This is equivalent to a new company coming out that sell Blu-ray players. You already own the Blu-rays, they're offering you a way to play them.
I'm very skeptical that the companies that sell products like this would willingly offer terms like what you seem to have in mind, e.g. 'Pay us and then just torrent the movie whenever you want!'.
But this is precisely what we had before the internet. We used to have physical media that (theoretically) could be copied to multiple devices. You weren't legally allowed to, but the technology left you in control. I know gamers as a group want this control back after getting shafted by video game companies.
How would the person or organization requesting someone's diploma verify that the NFT was issued by the relevant college/university?
...
Surely some kind of centralized database is still required, right? Wouldn't the entity requesting the invoice need to have some kind of 'whitelist' of entities that are approved to issue these NFTs?
No not at all. The blockchain can do this. The blockchain contains information about who minted it. So a hiring company - say, Google - would know which wallet MIT used for minting their diplomas. If it came from that wallet, then it's legit. In the long run, applications could create a more user-friendly layer. So you paste in the relevant info to your job application, and Google's HR database automatically verifies it on the blockchain.
this just seems like a more convoluted way to 'paywall' a download endpoint.
Definitely it seems like a paywall because it is a paywall. Just a more balanced one. As it is, any time you join any kind of DRM or paywall scheme, you're completely at the mercy of the content host, who can remove your rights to all the stuff you "bought" at any time.
I think this is where there's been a beanie-baby craze with this stuff on the supply side - where some parties are eager to just sell more because there's a new channel. When in fact, what we have is a better channel, and we need to move old sales onto the new channel.
Sure, existing markets won't like it - because they lose control. But it's comparable to how Netflix disrupted Cable TV. People liked the way Netflix served the content better, and cut their cables. Then they became too big to ignore, and all the major content owners got on board. I could see this happening with NFTs too. If I went to buy some digital content and there were two competing marketplaces - I would 100% choose the one that let me really own my purchase via NFT, instead of one that requires me to just trust them to keep it for me.
Trust powers commerce, and I really believe NFT marketplaces implemented properly will drive more people to buy with confidence.
losing access to whatever 'wallet' the NFT is associated with
The process would be the same as if I lost my Physical diploma. Call the university and they have to mint an identical one for me, at a cost to me.
So there are lots of situations where receipts are pretty important, but increasingly there is a push to have them be digitized, either because it's just more efficient or the asset itself is digital. For example, suppose you buy a movie through your smart TV.
The problem is these digital receipts tend to either be just downloadable PDFs - easily forged and effectively meaningless - or they're central databases, which are under the control of the party that owns the database. They can be revoked, lost, made obsolete, etc. Suppose the company that holds my downloaded movie purchase goes bankrupt. Now if I torrent that movie I'm not really breaking the law, or am I? How would I prove that?
A decentralized blockchain system would solve those problems and could be used for proving all kinds of things. "Send us a copy of your diploma" - "here, it's an NFT", "send us your invoice to prove you are in warranty", "here's the NFT", "Can you sell me your old copy of Civilization 4?" "Yes, I'll transfer the NFT to you for $40".
The potential to store trusted receipts has so many useful applications, we're just not quite there yet in terms of building it.
This is my thought exactly. The digital Art NFTs that are selling now or a distraction.
NFTs are a decentralized receipt system which is totally awesome and which you cannot get rich off by collecting receipts.
I also play a few games of chess.com every morning and I have found it has the same use you describe.
I read this book by Rovelli. It's supposed to be the best explanation available, but I still don't really get it.
In the end he suggests that entropy is why we can remember the past but not the future. I'm not sure that's it though.
Still a good read.
https://www.amazon.ca/dp/073521610X/ref=cm_sw_r_cp_apa_glt_i_WS4HJZDE23GGEZ6MSQYK
It would help to know what genre of game you are making. You talk about exposition, "We need to keep the exposition of these ideas short", and I would take this to the extreme if I were you. Show, don't tell. If players don't learn the concepts from the gameplay, then try game isn't about those concepts.
For example, if you want to teach players that ai optimism is not a good default and alignment is hard, give them a chance to do an alignment task or make alignment choices, in which there are optimistic options, that end badly. Or make a game that's almost unwinnable, to emphasize how hard the problem is.
Have you played universal paperclips? I've found it a fun first introduction to ai alignment for people with no knowledge of the topic.