Prediction Markets as Conversations

Also: DeepMind's Weather Modelling, BasePaint, and How Running Feels

I don’t play prediction markets, but I find them to be very interesting to read: both the market data and the usual comments attached to them. If you’re unfamiliar with them, a simple explanation is that people can bet against each other with money (real or play) on the outcome of events.

For years, I would at least once a week (or during elections: every day) pop into PredictIt to see what the market thinks about various future events. It was a unique form of news and more people (like Joe Weisenthal) are seeing the same benefit, especially as social media is splintering and new platforms are popping up.

As Joe mentions, they do not just provide one form of a crowd assessment of an event, but they also incentivise forecasters to surface and share relevant information from across social media that backs up their beliefs and predictions. It’s a topic channel that also creates a focal point gathering space because there’s a betting game that exists in the center of it. Thus, it becomes a unique source for news because of the incentive to predict it.

I think there’s still decent innovation space here and Manifold is doing a lot of the right things.

While PredictIt had been hamstrung by the CFTC, Manifold has gained ground by skirting around the regulations in using play money. For what it’s worth, my gut instinct is that play money prediction markets would function close enough to real money prediction markets if there is still a cost at stake. Manifold makes this work, not just in allowing the play money to be donated to charities for real money, but moreso that people care about their reputation in being predictors.

In PredictIt, there’s no profile system like this and so people stake their reputations just based on presence, but Manifold really leans into promoting the social aspect of prediction markets.

In seeing them as conversation spaces, what Manifold adds is valuable because you get more context beyond just the market and the predictions. For example, if you’ve made a prediction, it adds and anchors your information/bias to your comment.

You can also follow other predictors, seeing what they bet and comment on.

While the sentiment over time is useful, I think an additional valuable contribution would be to add comments into the timeline much like how SoundCloud innovated adding comments over an audio waveform.

Merge this:

With this:

It thus captures both market and conversational sentiment through time.

By moving beyond purely being a betting and prediction system into *also* being a social system for conversations is something I think in time will continue to grow.

Social innovation around prediction markets can definitely introduce more people to it. It’s not a replacement for your other news sources, but it’s definitely unique in the type of news you get from it. It’s sometimes much quicker to drop into a prediction market to get the zeitgeist vs trying to shuffle between Twitter/X’s poor search, going to specific social media accounts, or hopping around mainstream media outlets.

While I’m generally less interested in them being reliable or accurate over them being centers for news, Manifold has been quite accurate. They do reasonably well in last year’s US mid-terms.

Ultimately, they don’t have to be the best and most accurate for it function as a meaningful place to consume information. That being said, I think the markets and information are still a bit skewed given the predominant market that are using them: nerds, rationalists, tech people, and forecasters. I enjoyed the NY Times take on the Manifold conference. I must admit, while I don’t generally enjoy rationalism, the conference looked fun. Maybe I’ll attend in the future. Maybe we’ll bet on that. 😅

---

Bonus Content!

DeepMind Weather Forecasts

Speaking of forecasting. I’m a bit of a closet weather nerd. I’ve dipped into reading and trying to understand weather models and how they differ. I sometimes sit on Windy.com or TropicalTidbits.com for far too long.

DeepMind introduced a new weather model that is both cheaper to run and more accurate than other weather models. It kind of blows my mind.

GraphCast is now the most accurate 10-day global weather forecasting system in the world, and can predict extreme weather events further into the future than was previously possible. As the weather patterns evolve in a changing climate, GraphCast will evolve and improve as higher quality data becomes available.

To make AI-powered weather forecasting more accessible, we’ve open sourced our model’s code. ECMWF is already experimenting with GraphCast’s 10-day forecasts and we’re excited to see the possibilities it unlocks for researchers – from tailoring the model for particular weather phenomena to optimizing it for different parts of the world.

Like, here’s a piece of news that I didn’t really see mentioned much on the socials, but has the ability to make massively useful impacts globally. I’m excited!

It predicts hundreds of weather variables, over 10 days at 0.25° resolution globally, in under one minute. GraphCast significantly outperforms the most accurate operational deterministic systems on 90% of 1380 verification targets, and its forecasts support better severe event prediction, including tropical cyclones tracking, atmospheric rivers, and extreme temperatures.

BasePaint - 100 Days Retro

I’ve been enjoying playing around with BasePaint. Every day, people draw a new pixel art canvas together and then it gets sold. You get a cut based on your contribution to the canvas. It’s a fun, low stakes example of a blockchain art project.

w1nt3r wrote a really great retro on their first 100 days! Fun read.

Why Scavengers Reign Feels Alien

I’m still vibing from having watched Scavenger’s Reign. I enjoyed nerdwriter’s take on why it resonated. It feels as incomprehensible as our own nature sometimes.

“There is no real harmony as we’ve conceived it”.

Remember To Tip Your AI

Turns out that adding the prompt of a tip to the query generates longer responses. Which is both sad, alarming, hilarious, and all kinds of wtf. What have we done?

Felipe Gordon - This Is How Running Feels

You’re halfway through a long run, just had a gel, your legs are warm and any aches have dissolved. You speed up a bit and this song starts playing. Pure bliss.

That’s all for this week folks! As always, enjoy a sunset. There’s always one happening somewhere in the world (ht Rian).

Simon

PS. Been thinking more recently on where I want to take the newsletter as I’m coming up to the one year anniversary soon and where my life will take me next. For now, I’ve deprecated the Weekly Zora Mint section simply because it’s taking me too much time per week to find art I want to share alongside the rest of life. The referral fees just also wasn’t worth the effort to do so for that alone. Love the art, but not enough time for all of it! Sorry if you enjoyed that a lot!