_{Brought to you by:}

“Info finance solves trust problems.”

— Vitalik

Crypto’s emerging use case: Better information

If you’ve been reading this newsletter for a while, you’ll know I’m a fan of financial markets, mostly for their ability to harness humans’ innate desire for profit and redirect it into economic productivity and societal advancement.  

Also, they’re very fun.

This is why I like crypto too; like Midas and gold, crypto turns nearly everything it touches into a market — sometimes for the better, sometimes for the worse, but always for the more entertaining. 

Most recently, I’ve expressed my enthusiasm for crypto-enabled prediction markets as a way to make the world better informed by financializing information (see here, here, here and here).

This is not new (betting markets successfully predicted the Papal conclave of 1503, for example), but adding crypto to the mix seems to have vaulted prediction markets into the mainstream consciousness.

The crypto industry is understandably excited about this, although I think we may be getting a little overexcited. People now describe Polymarket as a “truth machine” and claim that it’s made pollsters, TV news and the entire mainstream media redundant. 

As evidence, they cite Polymarket’s election-night prowess, which is being a little exaggerated and maybe even mythologized.

Yes, Polymarket was quicker to get Trump’s odds to 90% than the network news (by about an hour, maybe) and that was very useful (I got to go to bed early). 

But however much it feels that way, 90% is not 100%, as anyone betting on the 2017 Super Bowl could tell you.

Politics is more important than sports (I guess), so the networks were right to refrain from calling the presidential election until 3 am ET when Trump really was 100% set to win, even if Elon Musk was citing prediction markets to declare him the winner several hours earlier.

The rest of us should proceed with more caution than Elon because Polymarket might get the next election “wrong,” or it might appear to declare a winner far too early.

This is important to keep in mind because if we imbue Polymarket with magical properties of divination (all based on last week’s sample size of 1) it could cause real problems in the next election.

Imagine the conspiracy theories that would ensue if betting market odds swung from one candidate to another on election night the way they did for the Super Bowl in 2017 — and how these conspiracy theories, in turn, could sow even more doubt about the integrity of US elections and undermine democracy even further.

This could happen because prediction markets are not all-knowing “truth machines,” as they’ve been anointed after last week.

Instead, they are “slightly-more-accurate-probability machines” (not quite as catchy, I know) and they’re not magic.

Instead, prediction markets work by incentivizing people to find better information and then share that information with the world by betting on it. 

For most things, prediction markets won’t work because there needs to be some informational mechanism (like a French whale commissioning private polls) to make them more accurate than pollsters’ models or people just guessing — and even then history suggests markets will outperform models by only a little.

Markets also need liquidity. 

A prediction market like the one for the bitcoin strategic reserve, for example, with minimal liquidity and no obvious mechanism to discover the truth, cannot magically divine the US government’s future decision to buy (or hopefully not buy) bitcoin.

This may change, however. 

In a blog post this weekend, Vitalik rebranded prediction markets as “info finance” and explained why he expects this new category of crypto-incentivized knowledge to become “incredibly powerful.” 

In short, he expects AI bots to make prediction markets both smarter and more liquid.

With swarms of purpose-built AIs placing bets on every available market, “we could potentially get reasonably high-quality info elicited even on markets with $10 of volume.”

If so, this would get us much closer to a sci-fi future where “truth machines” really do seem to magically predict events. 

Imagine thousands of autonomous AI agents competing to win in betting markets not just by scrapping the internet to ingest all human knowledge on a subject, but also by emailing humans with questions and paying money to incentivize the creation of all-new information.

Could AI bots commission a private poll that would give them an edge in the next presidential election? 

As long as pollsters accept crypto, sure, why not?

But this may only be the start. 

“Many of the most interesting applications of info finance,” Vitalik believes, “are on ‘micro’ questions: millions of mini-markets for decisions that individually have relatively low consequence.”

Prediction markets have been relevant only sporadically because it takes a high-profile event like the US election (or a papal conclave in 16th-century Rome) to create enough of an incentive for people like the French whale to devote time and resources to it.

Given the rapidly falling cost of both training and running large-language models, it should soon be economic for AI bots to spend time and resources on solving questions such as: Will the latest Elon tweet be flagged as misinformation? Will my high school win its next football game? Will my flight be delayed by more than an hour? Will my house sell for above asking? Will Coinbase meet its quarterly earnings projections? Will Byron ever shut up about prediction markets?

These questions may seem too trivial for markets, but probably not for long. That’s because “info finance solves trust problems,” as Vitalik frames it, and “a common concern of this era is the lack of knowledge (and worse, lack of consensus) about whom to trust, in political, scientific and commercial contexts.”

In the coming age of AI, we are going to have a lot of trust problems.

By creating better information for everything, info finance (née, prediction markets) will help us solve them.

— Byron Gilliam

Brought to you by:

Navigating the crypto world doesn’t have to be complicated. With Uphold, you can buy, sell, exchange, store, and even earn on your assets — all in one simple platform.
Here’s how Uphold helps:

• Trade over 300 assets* — and be the first to access new tokens.

• Enjoy seamless swaps, automatic recurring transactions for dollar-cost averaging, and advanced tools for smarter trading.

• Secure your crypto with Vault, Uphold’s “Assisted” Self-Custody solution.

Exclusive to US users: Earn up to 4.9% APY on your USD with the USD Interest Account on Uphold. Turn idle funds into passive income while you plan your next trade.

1. Fed heads into December meeting with new president-elect, same economic concerns 

2. A closer look at BTC’s record-setting surge above $86K

3. How memecoins could bring back the retail investor in crypto

4. SOL crosses $220, eyes fresh all-time high

5. Remembering FTX: How the collapse shaped crypto’s future

Why Base Layer Wars Are Ending and the Race Against Ossification

Astria founder Josh Bowen and Jon Charbonneau join Mike to discuss the recent launch of Astria and its role in the Celestia ecosystem. They examine the economics of L1 tokens, the challenges of achieving true decentralization and debate the future of blockchain L1 differentiation. 

Listen to Expansion on Spotify, Apple Podcasts or YouTube.

Brought to you by:

Meet Kraken Institutional — powerful crypto tools with white-glove service.

Not investment advice. Crypto trading involves risk of loss and is offered to US customers through Payward Interactive, Inc. View legal disclosures at kraken.com/legal/disclosures.