What was the aha moment for Scott with crypto?
- Scott was aware of crypto early, but very skeptical. Saw that everyone was trying to decentralize just for the sake of it. But decentralization has costs. Centralized entities are often more effective in coordinating interaction.
- His opinion started to change with examples of use-cases where tech was obviously more powerful.
- First lightbulb moment: Diem.
  - Decentralized transaction layers that worked like the internet but for transactions. This was very important because of financial inclusion and reducing transaction costs.
  - This is a place where the degree of decentralization actually contributes to the product because there shouldn't be just one point of failure
- Second and more powerful lightbulb moment: rise of NFTs and other consumer crypto apps that make it possible to have a market in digital assets.
  - These markets existed before, but we didn't have the concept of ownership for digital goods.
  - With NFTs, we started seeing people able to transact in assets that were previously almost impossible to transact in. In the past, you could build markets within video games, but they would have to be extremely small and niche. Or you could buy/sell digital art, but it was difficult and the only people who you could trust with certification or provenance were galleries.
  - But now, with crypto technology, all of the information can be transparent, and the transactions are easy to engage in.

How do you see the progression of marketplaces to the point where we now have crypto marketplaces?
- What do marketplaces do? They provide rules and/or infra that address sources of market failure. Market failure is when a market isn't creating as much value as it could. For example:
  - Complexity from heterogeneous preferences or products. To address this, marketplaces might create more nuanced search tools, or tools that do the matching themselves.
  - Hidden information e.g. lemons in the used car market. To address this, marketplaces might engage in verification, certification, insurance, etc.
- Online marketplaces initially started out as just a place for supply to meet demand (e.g. Craiglist). Soon after, marketplaces realized that they had high-level info that they could add value to the transactions (e.g. providing insurance)
- In the next wave, marketplaces have been building huge amounts of infrastructure to provide value to both sides of the marketplace, and to address issues like disintermediation (e.g. Gamer Sensei provides excellent tools for their coaches).
- With web3, the trust certification in marketplaces can be managed through the web3 tech rather than the platform. So it's much easier to enter and start to become engaged. As long as activity is on-chain, you don't need marketplace platform brands to be managing the trust process.
- There's much more interoperability now. People are building marketplace layers on top of underlying blockchain infra layers.
  - Tech is interoperable so if you're launching a new NFT platform, you can immediately leverage the existing NFT content.

Your essay on reputation systems says that they should be designed with a two token system: one for membership and one for liquidity. Why is the liquidity piece important?
- A lot of web3 platforms are trying to establish a reputation system that folks can convert to some reward. Incentives matter for most types of interactions, so if you want to encourage people to participate, having some way for them to share in the value that is created in their actions, it aligned their incentives.
- So giving users a degree of ownership is improtant. Simple idea: if you are a part owner, you are more incentivized to help create that success.

The moment you introduce a financial reward, people are less interested. Could financial incentives actually detract from our use of various platforms?
- It varies from context to context. Some people do things because they want to do them for free. And financializing it would actually make them less happy. E.g. some creators opted out of Quora financialization system.
- You have to know your audience and users and think hard about what makes them want to participate.
- Rewarding people in a way that is consistent with their motivations. Yelp reviewers may want to use their platform to build social capital.
- It's hard to keep people engaged without a direct feedback loop. It's difficult to get people to keep editing Wikipedia.
- If you had a version of Wikipedia where you could see how much the things you had written were being read, that alone could be a powerful motivator.
- Offering rewards doesn't have to be in conflict with people's desire to do things that they would enjoy doing anyway.

If you have a single token that is the mark of reputation and the fungible currency that people trade,
- If it's a really good marker of reputation, people will buy a lot of it, which means it's suddenly not a great marker of reputation.
- Reputation asset should be a non-tranferable thing that is attached to a person. But it could emit tokens that people can trade.
- The reputation can also do things. If you're a top poster on Reddit or Quora, you might share that reputation with friends.
- To the extent that you want to let people earn from reputation, then you might have a utility token that's used for sending extra messages, adding cool badges, etc. Then the benefit is that you could give the token to the high reputation people as a liquid component of the reputation reward.

What gives the liquid tokens value? Are they governance? Are they tied to revenue or cash flows of the businesses?
- Very important: people assume that if you add a token to the system, then suddenly you've added a bunch of value. Tokens aren't a source of value. The source of value is what people actually do with the token. A token has to have a purpose and a value framework before you decide how many to give out. ICO boom and bust is a good example of this.
- The use of tokens varies. You could imagine a variation of Instagram where all of the advertising were using an internal platform token. The platform would do this to control the amount that people do this type of boosting.
- Similarly, tokens are attached to levels of engagement. Twitch gives you points for watching a stream actively. Then you get micro-rewards.

LinkedIn Deep Dive: how would implementing tokens for LinkedIn work?
- There's a lot of information on that platform, but hard to determine accuracy.
- Let's think about skill system. If you want to be certified for having some skill, you ask someone in your network to certify you for that skill. But there's no opportunity cost of saying one person is good at a skill vs. another. Also no one knows if the giver is worthy to commend someone for a skill.
- You could imagine having marginal cost to give attestation. Maybe I accrue reputation to give feedback on certain topics. Down the line, if my reputation helped a hire to happen, I could gain reputation. So there could be a reputation loop.
- You could imagine attaching cost through native token to sending InMails. If you build reputation for evaluating skills, you accrue tokens which allow you to use the platform better.
- But why web3 for this stuff? Why not just use web2 methods?
  - LinkedIn could implement this with internal web2 'credit' type of token.
  - The web3 benefit is much more about interoperability. Enable people to build on it. People might want to collect it and give out e.g. universities giving out tokens to students.
  - Trustless nature of web3 is important also; if you're a small platform, people might not trust your internal credits (this isn't so much the case for LinkedIn, of course)

How do you think about defensibility for interoperable marketplaces from a business perspective.
- The main value add of web3 isn't always interoperabitliy.
- Web2 platforms have largely been about owning data and locking in users as their sources of competitive advantage. Web3 won't dismantle this overnight
- But there is a huge amount of value in building technology that is aligned with your users. You give us surplus in the short run, but you gain a lot in the future.
- Once you turn your user base into a community that is invested in your platform, a technical platform gain the benefit of a brand. Users see themselves as part of the entity. They invest a lot of energy into growing the platform. This can be an even stronger competitive advantage in the long run.
- Uniswap and Sushiswap are essentially the same, but people identify with one or the other.

What are the implications on the size of companies going forward? Some say they will be smaller because they will be infinitely forked. Some say larger because they're interoperable and better aligned.
- Hard to say. One thing is for sure: many more things will be growing bigger with smaller teams.
- We'll still see very large platforms. Potential for vampire attacks means that consumer have much more power relative to the platform compared to in web2 ecosystems.
- Platforms will differentiate based on more consumer-focused things; design, consumer facing features.

There's a lot of promise in the space, but having hard time to point to transformative DAOs or communities that live up to this promise. What's keeping us from taking advantage of that value?
- I don't want to generalize. But the simple answer is that it's still very early days. Tech is very new. We're not done with the part where we're taking existing transactions and making them more efficient.
- Hard to say how it's going to change our core interactions before the basic things get solved (e.g. event tickets).
- I'm confident we'll see it do cool stuff that we haven't seen yet.

Do you think the most powerful platforms will be web3 native? Or will we see platforms building two-token systems on top of web2 data.
- It varies from use case to use case.
- We'll see web3 native platforms become very big and do very well. Maybe some will replace their web2 counterparts (e.g. creator platforms).
- They will also be at different levels of decentralization.
- We'll also see web2 platforms where that technology is right. There will still be some advantage to having a centralized product.
- But interoperability will become a competitive advantage so it will influence all companies more into that direction.

Sari Azout • Market Design Masterclass ft. Scott Kominers