Building on Polymarket has made me think a lot about who actually ends up owning the customer in prediction markets. There are two very different models forming. On the US side, Kalshi increasingly looks like centralized market infrastructure distributed through giant consumer platforms. Coinbase is already surfacing Kalshi markets directly inside Coinbase, and Robinhood routes prediction contracts across several regulated exchanges. Will prediction markets become another financial product embedded inside apps that already have millions of users? That model has obvious advantages. You get trust, compliance, existing accounts, fiat rails and distribution without asking someone to discover a new trading product from scratch. Polymarket is different. The core liquidity sits in one place, but the company has intentionally opened up a builder ecosystem around it so other products can own the interface, workflow and even the user relationship. The current builder leaderboard has Betmoar doing roughly $79M of monthly volume, traderline around $52M and Gate around $38M, with the top 16 builders collectively routing well over $300M/month. And those builders are all pretty different, i.e. terminals, wallets, bots, trading tools and distribution products. Poly is much closer to how crypto infrastructure has developed generally. Coinbase or Binance try to own the full stack and the customer experience, while DeFi and venues like Hyperliquid create something developers can build around. You can have the same underlying liquidity show up through a wallet, a trading terminal, a bot, a Telegram interface or some product that doesn’t even look like the underlying exchange. I don’t think the answer is that the open model necessarily wins. Centralized distribution is incredibly powerful. If Coinbase, Robinhood, DraftKings and other large consumer platforms can make prediction markets feel like another tab inside products people already use, they may drive far more volume than a fragmented ecosystem of specialist front ends ever does. But as prediction markets become more like active trading markets, I also think the value of specialized interfaces goes up. Someone trading 5-minute BTC markets wants different things from someone buying an election contract once every few months: multiple price feeds, order-book depth, batch execution, alerts, wallet tracking, automation and much tighter workflows around entering and managing positions. A single consumer interface is unlikely to be optimal for every kind of trader. So will prediction markets end up looking more like online brokerages, where the distributor owns most of the experience, or more like crypto, where the underlying market becomes infrastructure and dozens of products compete on top of the same liquidity? It may end up being both!

Kalshi definitely gaining share, but likely distorted due to notional volume being "boosted" by parlays, perps incentives, and possible wash trading

People are still sleeping Crypto Up/Down markets. They're basically ultra-short-dated binary options in a much simpler wrapper. Same directional view, but instead of navigating strikes, Greeks and option chains, you pick a side and the contract pays $1 or $0. It's going to be a much bigger category than people realize

One thing I keep coming back to after digging into crypto Up/Down markets is that the cleanest way to think about them is as ultra-short-dated binary options. They sit inside prediction markets today, but economically they look a lot more like a very simplified options product. A normal BTC option gives you a lot of choices. You pick the expiry, the strike, whether you want a call or put, and then the value of the option depends not just on whether BTC moves in your direction, but how far it moves, how much time is left, implied volatility and a bunch of other variables. That flexibility is what makes options powerful, but it’s also why they’re confusing to most retail traders. A BTC Up/Down market collapses most of that complexity. Every few minutes you effectively get a new strike around the current BTC price and a new expiry a few minutes later. If you buy Up and BTC finishes above the relevant level, the contract pays $1. If it finishes below, it pays $0. The exact settlement mechanics differ a bit by venue, but the economic exposure is basically a binary or digital option with an extremely short maturity. The payoff is also different from a normal call in an important way. If BTC starts at $100k and finishes at $100,001, an Up contract pays the same as if BTC finishes at $105k. A vanilla call keeps becoming more valuable as BTC moves further above the strike. With the binary, most of what you’re trading is the probability of finishing on the right side of a line over a specific period of time. I think that simplicity is a big part of why the product is interesting. A trader doesn’t need to navigate an options chain or understand Greeks to express a short-term directional view. The downside is known when the trade is entered, there’s no liquidation mechanic like there is with leveraged perps, and the market price itself is relatively intuitive because a contract trading at 60¢ is basically telling you what probability the market is assigning to that outcome. There’s also a broader trend here that makes the timing interesting. Traditional options markets have already moved aggressively toward shorter duration, with 0DTE now accounting for a huge share of SPX options activity. Crypto Up/Down markets take the same basic desire for short-duration exposure and compress it from a day into an hour, 15 minutes or five minutes. Polymarket creating 288 separate 5-minute BTC expirations every day would have sounded pretty bizarre as an options product a few years ago, but traders are clearly using them. Where I think this could get more interesting is when the same structure starts moving beyond crypto. Stock Up/Down markets are already appearing, tokenized equities are moving toward 24/7 availability, and once you have reliable price feeds there isn’t much structurally preventing the same type of contract from existing around equities, commodities or other liquid assets. You can also start imagining more practical hedging use cases around earnings, macro releases, overnight exposure or very specific windows of risk, rather than these being used almost entirely for short-term speculation. I don’t think these replace traditional options, because the fixed binary payoff gives up a lot of what makes vanilla options useful. But I do think prediction markets may have stumbled into a much simpler retail wrapper for one slice of the options market, and the more assets and expirations they add, the less useful it becomes to think of the category purely as people betting on events.

Finally taking the covers off @ZoqoTrade which we've been building over the last few months! tl;dr we launched no-code trading bots for polymarket starting with crypto up/down markets. Until now, only a sophisticated few have been able to roll their own bots / automated trading infra. We're giving these tools to everybody else. Now anyone can turn an idea or signal into a persistent, automated strategy without having to build and maintain the infra yourself. You define conditions, entries, exits and guardrails, and let the system run them for you instead of wiring together APIs, scripts and infra yourself. We’re starting with crypto Up/Down markets on Polymarket, where the combination of fast expirations, order-book trading and heavy automation makes the need very obvious. We're rapidly turning on deeper automation logic, more sophisticated position management and execution tools, more external signals, agentic trading, backtesting and much more. We’re live now in invite-only beta. If you actively trade prediction markets and want access, DM me.

We’ve spent the last few months building around prediction markets, and one thing has surprised me more than I expected: crypto Up/Down markets are starting to look a lot less like prediction markets and a lot more like a new derivatives market. Polymarket launched 15-minute crypto markets last year and they ramped quickly, getting to almost $300M of weekly volume by February. Then 5-minute markets launched and passed 15-minute volume within two weeks. At one point 5-minute markets were doing roughly $385M/week on Polymarket alone. Kalshi has taken the same behavior even further. Crypto did roughly $9B of volume there in September, and the large majority of recent crypto volume has been concentrated in its 15-minute markets. This is happening inside a prediction-market category that itself is growing much faster than most realize. Kalshi + Polymarket did roughly $51B of volume in all of 2025. They had already done around $257B through August of this year. The mechanics are almost absurdly simple. Every few minutes a new market opens around a reference price. Pick Up or Down. The winning side ultimately pays $1 and the losing side pays $0. But the way these markets trade is very different from something like an election market, where someone might buy a position and sit on it for months. BTC alone can generate hundreds of separate expirations every day. The same capital gets recycled constantly across new contracts, so volume turns over extremely quickly relative to the amount of money actually sitting in the system. You also start to see much more traditional exchange behavior underneath the simple UI: order books, market makers, benchmark settlement, maker/taker economics, professional traders and a lot of automation. On Polymarket, on-chain analysis has estimated that more than half of short-duration crypto volume comes from automated traders. I don’t think this stays confined to BTC or even crypto. Stock Up/Down markets are already starting to appear, while tokenized equities are moving toward 24/7 trading and crypto venues are expanding further into perps and other derivatives. It’s not hard to imagine these categories starting to blur together. There are also some interesting use cases beyond pure speculation as the market matures. A short-duration contract with a known maximum loss and a defined payout can be a pretty simple way to express or hedge a very specific view around an earnings release, an overnight move, a macro print, or temporary exposure to another asset. I’m still not sure how large this ultimately becomes. Maybe 5- and 15-minute markets remain mostly a very high-velocity crypto trading product. Or maybe this is the beginning of a much broader retail derivatives category built around simple binary payouts, short expirations and assets that trade around the clock. Either way, it feels like something materially different is forming inside prediction markets than what most people associate with the category today.

Nubank is the world's largest neobank. Their 140M customer base dwarfs Revolut's 80M. Yet nobody thinks of $NU as the global leader. That will change with their announcement yesterday launching the US in addition to 35+ other countries. Until now, Nubank has really been a LatAm story. Brazil first, then Mexico and Colombia. Now they're launching a full banking suite in the US, its first major developed market, while simultaneously launching Nu Global, a multi-currency account for saving, spending, and moving money cross-borders. Nu Global cleverly uses stablecoin rails. Deposits are converted into USDC/EURC. Customers can earn yield, spend through Mastercard or transfer money internationally with no transfer fees or FX spreads. This puts $NU in a completely different market. They're no longer just taking share from incumbent LatAm banks. Nu Global puts them directly against Wise, Remitly and Western Union for hundreds of billions of cross-border money movement. The US will be harder since banking is competitive, customer acquisition is expensive and Nubank doesn't have the massive incumbent cost advantage that it had in Brazil. But they do have two unusual advantages: a cost to serve of less than $1 per active customer per month, and a brand already known by tens of millions of Latin Americans. Can $NU become the largest truly global consumer bank?

Wow this blew up. The replies fell into three categories: 1) You're an moron for giving it your credentials 2) Relax bro its just a proxy 3) Proxy fine but WTF are they using Iran for?!?! FWIW a few trusted folks did DM me and mention they knew of folks who had compromised accounts such as LinkedIn, and in one case it was actually an Iranian

So I was using Instinct and gave it my Verizon credentials, and then I received this authentication request from Tehran. This is unbelievably concerning and ofc I deleted Instinct immediately. Can someone (or @noahrshinn) pls explain how this could happen because otherwise it seems that their servers are compromised by Iran. I messaged Instinct directly about the incident but no response. No consumers should be giving these personal agents sensitive login info.

Gen Z is AI-native. But Gen Alpha might be AI-dependent. They’re incredibly fluent on mobile. But many are less comfortable with keyboards, files and folders, terminals, and basic computational thinking. The frontier layer of AI still lives there.

This is massive especially since traditional mortgage lenders count your crypto as $0. They don't factor in crypto at all when underwriting!

Yeah right. Imagine the shenanigans that’ll occur if you have the whole row to yourself on an overbooked flight

CLARITY’s direction on stablecoin yield suggests issuers and exchanges can’t just park reserves in T-bills and pass through interest like a bank account. If rewards must be tied to activity, yield has to come from actual capital deployment like lending, collateral, liquidity, and trading. That’s bullish for DeFi lending markets and the front-ends that route users into them.

Ok so has the pendulum officially swung back to remote work again or no? Since most of our coworkers are agents

Getting a little tired of these $150k launch packages shilled by unmarked paid influencers

The games industry pushing back on AI is totally fair. Craftsmanship matters and the experience is sacred. But vibecoding is doing something interesting in modding communities. Players who couldn't write a line of code are now building. Since it's not a threat to the core game, the studios that figure out how to harness this by give players the right sandbox, with AI lowering the floor, will be massively successful. Basically Roblox / Minecraft / Warcraft 3 custom games on steroids.

NBA Top Shot just venmo’d me $83 as part of a $7m class action lawsuit (illegal sharing of user data). Crazy how many years later these payouts can happen

Karpathy is right, software is about to move at lightspeed. the interface layer of the internet gets sanded down and rewired as agents become the primary users. It'll go headless fast (API/MCP becomes the product). I see this happening in four ways: 1) Fully headless - high intent flows like travel booking, hotels, flights, ride hailing, delivery, reservations, commodity ecommerce (electronics, household goods), bill pay, insurance quotes, etc, not to mention most of SaaS. These UIs are mostly forms and funnels. Agents fill forms better than humans. 2) Thin UIs - amazon-style browsing, real estate search, job marketplaces, trading platforms, app stores and plugin marketplaces. Agents handle execution. Humans still want some oversight and discovery. 3) UI remains core (experience is the product) - social networks, games, streaming and sports fandom, fashion, luxury, beauty, dating, creator platforms, brand-driven commerce, high-consideration purchases (cars, homes). Taste still needs a surface. Then a new category appears: 4) agent-native networks - agents discover tools, negotiate pricing, route tasks, build reputation, coordinate with other agents. Humans mostly watch. Moltbook is an early glimpse. tl;dr a lot of websites become "inventory" for agents. Brands become trust layers on top of APIs