Goodbye & Final Reflections 👋

We're closing + future of web3 opinion layers

Good morning friends —

After a lot of reflection, I’ve decided it’s time for Ideamarket in its current form to close its doors.

Our official product costs money to run properly, and after nearly 6 months of financial struggles, it’s become unfeasible to maintain it at this time. There’s a chance we’ll be back up and running someday, but it seems prudent to manage expectations.

Here’s what you need to know:

• The main website will remain live through February to make it as easy as possible to withdraw funds. After March 1, I can’t guarantee you’ll be able to use our website to access locked IMO, or to withdraw funds from earlier versions of Ideamarket (back to Feb 15, 2021). After March 1, you may need to call the smart contracts directly to retrieve funds. (Here’s the Guide to Selling Old Tokens & Unstaking.)

• The Ideamarket Discord will remain open. Token holders are also invited to discuss, coordinate, and develop new projects at https://commonwealth.im/ideamarket/

• I am seeking a new primary income stream outside of Ideamarket, so while I will be happy to advise community projects, I may not have much time, even though I still believe in Ideamarket’s approach.

• We’ll keep our docs live at docs.ideamarket.io, so you can read about our product and philosophy.

• Our code will remain open-source at github.com/ideamarket

A few concluding thoughts and predictions:

The future of the web3 opinion layer

An “opinion layer” of the internet, and of web3 especially, seems inevitable. The potential for use cases is massive, and the primitives required to build it are simple.

Here’s what I expect to see, and what I’d focus on if I started all over:

1. A web2 version

   1. Twitter or any other social network could easily do what we’ve been trying to do:

      1. Opinion graphs — Take people’s ratings and aggregate them, sort and filter, ad targeting, etc.

      2. Knowledge graphs — Link Tweets together to show inference and argumentation, so that the best arguments for any viewpoint get aggregated in the same place.

   2. I’ve written about this more at length in this recent blog post: How Twitter can fix misinformation, without fact-checking

2. Web3 opinion primitives

   1. Uniswap for on-chain opinions. The primitive should be as simple and composable as possible, so it can easily play the same role in many ecosystems and become the industry standard.

   2. For example: A dapp for rating any NFT. Opinion ratings can then be applied to any web3 social post. A separate website would display the most frequently-rated NFTs, reddit-style, so that people can see what everyone else is most interested in talking about.

3. DAO tooling

   1. Web3 opinion primitives would enable DAO communities to rate the value of each member contribution in order to distribute funds and rewards fairly.

   2. In order to decide whose opinions carry the most weight, a “delegated proof of stake” layer also seems necessary. If you don’t like an influential person’s judgments about the value of member contributions, you can delegate your votes to someone else. This seems to strike a happy medium between 1-wallet-1-vote democracy, and control by whales. This is what we used the $IMO token to do on Ideamarket.
      

4. Sorting & Filtering

   1. Sort opinions by other on-chain characteristics, such as NFT holdings, wallet holdings, DAO membership, or anything else.

      1. Filter by credential — e.g., “What do holders of [NeuroscientistDAO NFT] think about [topic]?”

      2. Community belief analysis — “What do holders of [BoredApe NFTs] believe as a group?”

      3. Airdrop targeting — “Whitelist all wallets who believe [x] above confidence level [n%]”

      4. Financial analysis — Discover how opinions correlate with on-chain financial statistics. See our experiment in this blog post: Belief-wallet analysis

      5. much more
         

5. Persuasion markets

   1. Unlike prediction markets, persuasion markets settle based on the on-chain opinions of a user or community.

   2. Example: I’ll bet 1 ETH that...

      1. [BoredApe NFT holders] will be over 50% confident that [Tether is a scam] by [date]

      2. [Balaji Srinivasan] will be over 50% confident that [COVID came from a lab leak] by [date]

      3. [All users who believe [x] above [n%] confidence level] will be over 50% confident that [y] by [date].

Indeceivable

We must build technology that openly acknowledges which information ultimately rests on personal opinion. (If you’ve never read my original exploration of the problem with “facts,” read it on Ribbonfarm.)

If we seize this opportunity, we might prevent all the problems that arise when we’re promised “facts” or “cryptographic truth” but actually receive “personal opinion” in disguise.

This is why it seems important to use terminology that announces this is a personal judgment, such as “Agree/Disagree.” By contrast, terms like “True/False” invite a power struggle, as they hide the underlying question: according to whom?

Above all, the opinion layer of the internet must not presume to be a fact layer.

If we build the future of knowledge around “opinions we’re allowed to iterate” instead of “facts we’re forced to believe,” we prepare ourselves to learn quickly, and to decide wisely who deserves our trust.

That’s my opinion, anyway. 😉

via SlateStarCodex — this image helped inspire Ideamarket’s founding

Systematic Gentleness

Many attempts at solving the credibility crisis seem to flirt with tyranny.

“Cryptographic truth” advocates say facts can be verified by looking on-chain. AI advocates say facts can be synthesized from a multitude of inputs. But even blockchain-verified facts must be interpreted and woven into narratives by humans, and AI fact-synthesis models must use parameters decided by humans.

Therefore, even if both groups succeed in producing facts, the decision whether to trust the humans involved will still rest on each member of society individually.

Against the cryptographic truth elites, people will argue “This may be on chain, but you’re wrong about what it implies,” and against the AI elites, people will argue, “Why do you get to decide citations of Alex Jones carries less weight than citations of CNN in your algorithm?”

Unless the public is able to represent their genuine opinions within the new sensemaking tech stack, it seems we’ll end up right back where we started. We might as well start saying, “No information without representation!”

Without such representation, I worry that advocates of cryptographic truth and AI will be left with a choice: admit not much has changed, or try to coerce agreement using the same tactics used today by governments and media corporations.

To avoid information tyranny, we must systematize non-coercion.

Changing our minds is hard. It tends to hurt. It tends to be unwelcome. Depending on the amount of emotional energy we’ve invested in our beliefs, changing our minds can feel like dying, and we often defend against it with the same desperate intensity.

I recently saw a movie about an autistic woman who revolutionized the cattle industry by designing slaughterhouses that keep cows in a state of calm and peace as they walk to their deaths.

To create a society that values truth, we must do this for our prejudices. We must build information infrastructure without prejudice regarding any voice or conclusion, so that as genuinely trusted voices and conclusions emerge, people walk willingly through the necessary death of mistaken beliefs.

What must be cryptographically verifiable is not “what the truth is,” but lack of prejudice as to what the truth is.

To verifiably care about finding the truth, we must verifiably not-care about what the truth turns out to be.

Final Thanks

I’d like to thank our team for their faith and determination over the years:

(Bold = team behind final product)

Sam Ratnakar
Alexander Schlindwein (CTO)
Bhanu Teja (Frontend)
Kelton Madden (Smart contracts)
Joshua Jackson (Frontend)
Mac Dziedziela (CTO)
Sai Dheeraj (Backend)
Braden Lockwood (Product)
Kamal Singh (Design)
Tuan Phung (Frontend)
Sergey Misyura (Frontend)
Garrett Dailey (Chief of Staff)
Sam Barton (Product)
James Ellis (Podcast)
Gritcult (Operations)
John Giuffre (Marketing)
Sam Hebda (Community)
Victor Valentine (Marketing)
Matt Byrne (Marketing)
Oceane Beck (PR)
Leah Elias (Operations & Enneagram Coach)

These are all stellar guys and gals.

Our investors

Hartmann Capital, Mechanism Capital, The LAO, Mask.io, and angels

Our users

And all of you, our friends, who have believed in us and helped us along the way.

Onward!
—Mike

PS — I am looking for a new source of income, so don’t hesitate to reach out. 😊 I may write further on these topics via my personal newsletter (mikeelias.substack.com).