Introducing Boring News
Our new daily news show based on Polymarket, built with AI
Welcome to the 267 newly Not Boring people who have joined us since last week! If you haven’t subscribed, join 235,645 smart, curious folks by subscribing here:
Today’s Not Boring is brought to you by… Speakeasy
Speakeasy recently raised $15M to revolutionize REST API development.
Even the most gifted craftsman needs the appropriate tools to build something they’re proud of. Unfortunately, the API tooling ecosystem is still stuck in 2005.
Speakeasy wants to make it trivially easy for every developer to build great APIs. To make that dream a reality, Speakeasy is building a platform to handle the heavy lifting of API development and unburden developers to focus on refining the business logic.
From fast-growing innovators like Vercel, Mistral AI, and Kong to hobby projects by solo devs, engineers use Speakeasy to accelerate their API development and adoption.
Hi friends 👋,
Happy Tuesday!
Over the past few weeks, we’ve been cooking something up, and I’m excited to share it with you.
Let’s get to it.
Boring News
Today, we are launching a new daily series, Boring News:
Every weekday, we’ll deliver three important stories from Polymarket. We’ll start with the odds that the market sets and use AI tools to try to explain why the odds are what they are.
We’re calling it Boring News because compared to sensationalism, tribalism, and endless BREAKING chyrons, simply taking the odds that the market sets and trying to explain why they might be that way is … boring.
We’re taking individual humans out of the process as much as possible, and replacing them with two forms of artificial intelligence: markets and AI.
The video you just watched pulls stories from Polymarket, runs their odds through news APIs, Perplexity, and Grok for context, turns that context into scripts with ChatGPT and Claude, and narrates it using my voice from Eleven Labs.
Then we share it directly with you via X, YouTube, and podcast feeds. Subscribe at YouTube, Spotify, or wherever you listen to podcasts to get it each morning.
To start, we’re selecting the markets and running them through an AI-powered pipeline. Over time, we’ll do market selection by algorithm, too – biggest movers, most liquidity, highest volume, closest odds – and remove ourselves completely.
Our early version is not perfect. We’re wrestling with the tools to try to get the most accurate and relevant information and the snappiest scripts without writing it ourselves. It feels a little bit like watching your kid do their homework and restraining yourself from giving them the answer.
But I think it’s both fun and valuable to experiment while the tools are still imperfect so that we’re in the right place as they improve. It’s a little like the Rox strategy – use today’s AI to build something useful and grow our audience, so that we can improve our processes and benefit from improvements in the underlying AI models. In short, Boring News will only get better from here.
So why are we doing this? And who’s we?
When Polymarket correctly called the 2024 US Presidential Election, I got the strong sense that news had changed forever.
Certainly, it won’t change overnight. It was a “The future is here, but it’s not evenly distributed” moment. The combination of unfiltered information on platforms like X, Substack, and podcasts with liquid prediction markets like Polymarket have the potential to deliver something closer to the truth to people.
Markets incentivize correctness, punish bias, and reward truth.
Prediction markets are news with numbers. In the future, news without numbers will seem archaic, a quaintly imprecise relic from a more biased era.
Personally, for almost a year, I’ve been checking the Polymarket home page to scan my news each morning before checking traditional media.
Now, whenever I see a headline, I check the Polymarket odds behind it to understand how real it is. This is, to be sure, a boring way to get the news, free of the hyperbole, hyperventilation, and hyperpartisanship that infects traditional news media. But it also seems to be more accurate.
That said, most people won’t check the charts every day. We’re used to neatly packaged stories that provide the context behind the headlines. If those packaged stories come with bias and sensationalism, well, that’s just the trade we make.
One of the things that technology is best at, however, is eliminating trade-offs. As I watched Polymarket during Election Week, and read stories like the one about the French Whale putting $30 million behind his correct call on Trump’s victory, the gears started spinning.
What if you could create a news show that delivers news based on prediction markets? You could pull from Polymarket, run it through LLMs, and deliver it practically autonomously. Self-driving news: insert market, get full story that changes as the odds do.
Technically, this is something that could not have been built until this year.
Prediction markets were too illiquid to be accurate. AI wasn’t powerful or internet-connected enough to provide the relevant context, turn it into a script, or deliver the script in audio format with a realistic voice clone. The tools will only get better, and we expect Boring News to continue to get better as they do (and as we do).
This is also not something that I could have built myself, even with AI’s help.
I keep saying “we.” I have a lot of ideas like this, and I execute on basically none of them.
But in late October, 19-year-old Indian cousins Aman Matreja and Sehaj Pasricha cold emailed me about a (different) project they wanted to work on under Not Boring. I liked the idea and thought they were excellent, so I asked them to join Not Boring.
After the election, I told them about the idea for Boring News, and they replied, basically, “No problem.” The next day, they had the pipeline set up and the first draft done. With very little direction, just some mad ramblings. Which is to say, this is more their doing than mine and I feel very lucky to be working with them.
Around the same time, on November 8th, I got a DM from Mike Wenner. We’d worked together when he was at Masterworks and Composer, and he told me he’d just joined Polymarket. A few days later, I told him about the idea for Boring News, and he told me they’d sponsor it if we did it.
In under a month, from a half-baked idea, we had a working pipeline, test videos, and the perfect partner. And now, it’s live.
The more we play with it, the more excited I am about the possibility that technology will both make news more accurate and restore trust in the media, even if “the media” looks a little different. Even if “you are the media now.”
One recent example really drove the point home for me.
On November 29th, the BBC published a piece on the Irish elections, Tight three-way race in Irish general election, RTÉ exit poll suggests. In an analysis, BBC News NI’s political editor, Enda McClafferty, wrote, “If the exit poll is right this time, then Sinn Féin will be pleased with its performance” before bravely noting that “the true picture will only become clear when all results are in.”
The true picture was already clear on Polymarket, where Sinn Féin had been languishing below 20% odds for over a month, and where bettors correctly predicted Fianna Fáil’s victory as early as November 25th. Today, Fianna Fáil is at 99.7%.
Despite my Irish heritage, I don’t normally follow Irish elections and I didn’t understand why the market favored Fianna Fáil when Sinn Féin was ahead in the exit poll. But when Aman and Sehaj ran the odds through our pipeline and asked it to explain why, it explained that “sentiment diverges from the exit poll's first preference results, highlighting the complexity of Ireland's electoral system, where vote transfers and strategic voting are crucial” and “coalition dynamics favor Fianna Fáil, as both they and Fine Gael have consistently ruled out partnering with Sinn Féin, making a renewed coalition between the two centrist parties the most likely outcome.” That is what ended up happening.
Starting with the probabilities and asking why they are what they are is a valuable flip.
Of course, daily news is just one small piece of journalism. There will always be crucial roles for investigative reporting, long-form analysis, and the kind of deep work that helps us understand not just what's happening, but why it matters. What we're trying to do is make the day-to-day news delivery more accurate and less sensational, freeing up human resources and attention for the kind of journalism that truly requires human insight and expertise.
To that end, Boring News is still very much a work in progress. We're going to try to put out a short, five minute video every weekday and continue to refine it based on your feedback. You might listen to some stories and think, “wtf that’s not what happened!” or “I don’t agree with that explanation!” First: blame the AI. Then: let us know so we can improve it.
If people like it, we'll work with Polymarket to put more resources behind making it as good as it can be.
I believe in David Deutsch's Principle of Optimism: "All evils are caused by insufficient knowledge." And I believe in Joy's Law: "No matter who you are, most of the smartest people work for someone else." Prediction markets leverage the latter to solve the former – they incentivize anyone with valuable information to contribute it.
That's what makes me optimistic about the future of news. Not because of any one project, but because the tools to find truth are getting better and more open. Going direct is open. Markets are open. Smart 19-year-olds anywhere in the world creating with powerful tools is open.
We won’t be the only ones who try to build something like this, and that’s awesome.
We're having fun experimenting with these tools to try to make the news a little more boring, and hopefully a little more true.
The news is dead. Long live the news.
Thanks to Aman, Sehaj, and Mike for building Boring News with me, and to Claude for editing!
That’s all for today. Give Boring News a listen and let us know what you think. We’ll be back in your inbox on Friday.
Thanks for reading,
Packy
Ha, I'd been working on an extremely similar concept, but without eleven labs. Its a work in progress as well but you can access it here.
https://polynewsdaily-40fee7e8b4b8.herokuapp.com/
Will this also appear natively on Substack? HUGE if so. Looking forward to adding to my morning audio newspaper.