“An Outside Context Problem was the sort of thing most civilisations encountered just once, and which they tended to encounter rather in the same way a sentence encountered a full stop.”
– Iain M. Banks, Excession
My last piece on where prediction markets break down made me realize I hadn’t written about outside context problems here yet.
Outside Context Problems are the most interesting situation one can encounter, when considering pure innovation.
TL;DR: if you expect an embryonic prediction market to tell you the result of a highly-politicized event when there are perverse incentives at play, and a significant disparity between amount-at-risk and expected outcome, you are going to have a bad time.
While discussing BlueSky’s custom feeds and composable moderation with a friend, he wondered:
I will need to ponder whether it is viable at scale for millions of users to put in the effort to customize their feeds in a thoughtful way […]
They don’t have to - they only need one person they trust (or are willing to take at face value for a test) to create a feed, or to add a feed to a starter pack.
But more importantly, I realized, any customization someone can do, thoughtful or not, is an improvement. It introduces some delta on the perspective, even while they are within a group, instead of putting them into the monoculture dictated by which news items are more clickbaity and likely to bring ad revenue.
It is the same way with genetics: even small mutations can help diversify the perspective gene pool.
There’s a discussion going on the Topology discord about what systems can be described as peer-to-peer.
I think the whole thing is mostly academic, and too focused on implementation details.
Sure, your particular choice of blockchain might have a peer-to-peer gossip protocol and consensus mechanism, but that’s rather irrelevant if people are mostly using tokens with a freeze authority, isn’t it?
And while Binance Smart Chain might have a P2P protocol for sharing information, that’s entirely irrelevant when it’s completely proof-of-authority and a single institution chooses who gets to validate.
If you are looking to describe a system, describe the kind of things it enables that no other approach does.
That’s something the local-first camp does right, and something other projects should strive for.
Over at BlueSky, Dell Cameron brought up that Facebook’s involvement in crypto, VR, and GenAI aren’t working out because they are only secondary to their main business purpose: surveillance. (link)
I disagreed. Cameron was right in that it failed because their main business is surveillance, but I suspect they screwed it up because it was such a great fit that they got too greedy.
Crypto would have been for them - if Libra had taken off, Meta would have gotten access to their users' transaction information, something they don't necessarily have right now.
They luckily came out the gate too hard and scared everyone off by making it too obvious it was a supranational currency.
— Ricardo J. Méndez (@ricardo.bsky.social) Jun 4, 2024 at 9:08
[image or embed]
Would it have been different if they’d seen it as a technology investment?
Two projects grew on the compost of Facebook’s Diem open source dump: Sui and Aptos. As of June 4, 2024, their combined fully-diluted valuation is about $10 billion. Facebook could have captured that value if they had spun it out from the start, even as a subsidiary where they control most of the token supply.
But Libra was never a tech investment. You don’t get to have an arms-length independent company and share data.
When all you have is surveillance, everything looks like a data trove.
Markets are a series of bundling and unbundling, and whomever does the cleanest job, wins.
A clean unbundling makes it easier for others to build upon you, to integrate, to build businesses around you. It’s why people used S3 instead of other more elaborate mechanisms.
A clean bundling makes your whole stack look solid, integrated, like what Apple used to be.
In both cases, you get the advantage of reduced cognitive overhead.
Back in 1999, David Bowie spoke about how music used to be subversive, but had since become merely an information delivery mechanism - it was the Internet that was subversive.
Fast-forward, and the Internet is an information (and disinformation) delivery mechanism - it is crypto and other areas that are subversive.
The thing that is subversive now is the what will make money in a few years.
TL;DR: I don’t see any practical path forward for decentralized stuff other than abandon the pretense of cross-chain value transfer and embrace balkanization, while letting the centralized stuff handle the cross-chain value transfer.
Some thoughts on a 1996 paper by Narayana R. Kocherlakota.
Fully private systems, like Aleo, will be mostly memoryless systems (memory will be provided only by an individual’s viewing keys).
These systems will allow for a class of application we haven’t seen online before - UTXO/records as rai stones that just happened to materialize along with a zkp.
Given that fully private systems will be memoryless black boxes with zero-knowledge proofs, one cannot verify that a contract path got executed in the exact same way that you would have wanted. Therefore, it makes no sense to cling to the notion of traditional execution, and one should focus on the result that one wanted to achieve.
This would seem to indicate that private systems lend themselves to intent-based approaches.