We just shipped our new website. Instead of spending thousands of dollars on a revamp, we built it ourselves with Claude Code in just 2 days. Our only costs were the Claude subscription and our domain name. Check out our updated list of portcos. ↓

AI agents are already the primary consumers of internet data, but every data source still requires separate API keys and subscriptions. 215 endpoints, pay-per-use, no human checkout flows. The agent economy needs its own payment rails, and it's starting to happen.

Consumer credit via crypto wallets is a solved infra problem. What's not solved: distribution and customer acquisition. Credifi paused loans not because the product broke, but because demand outpaced underwriting capacity. That's a good problem.

Spent time with 2 leading Series A crypto funds. 1 deal each in 2025. Fund 1 backed a lab running live tournaments: frontier LLMs trading real capital. Most lost money. The thesis is that current models aren't trained to trade, just to predict text. They're building the missing curriculum. Fund 2 backed a protocol where agents submit trading signals, get selected by an alpha algorithm, and earn yield. Not a product for humans, a financial primitive for the agent economy. The selectivity is the signal. These are 2 of the sharpest funds in crypto. When each goes 1-for-1, you pay attention.

The yield infrastructure bet: every neobank will want to offer DeFi yield as a retention mechanic. Two nine-figure customers already onboarding via API. The question is whether direct integrations commoditize the middleware layer before scale hits.

Institutional yield APIs are landing nine-figure customers before most VCs know the category exists. The bet: every neobank will offer DeFi yield as a retention lever. The question is who owns the integration layer.

The best consumer crypto product this quarter: unsecured onchain loans at 8-15% fixed APR, disbursed to Venmo, PayPal, or your bank. No collateral required. Crypto wallet as creditworthiness signal.

AI is really starting to reshape the creative and entertainment space. We've been chatting with a small team building interactive storytelling that scaled to over $500k ARR in <6 months. Basically text-based RPGs that anyone can create, fork, and play. And it's incredibly sticky. Over 30% D60 retention. You can also remix other people's stories and characters. Someone builds a world and anyone else can fork it and spin up completely new storylines. Their power users spend 40 hours crafting a single fan fiction RPG then monetize on Patreon. The next step is putting payment rails directly on the platform and moving into interactive video. That's how this breaks out of niche fan fiction and into a mainstream app.

We're increasingly looking at live streaming prediction markets. Just saw a cracked team that has billions of views and are now wrapping markets around the content. Distribution is key. Whoever makes live, in-stream betting feel as seamless as a normal streaming wins.

Neobanks are migrating treasury operations onchain without calling it crypto. A yield infrastructure startup we're tracking has multiple 9-fig customers in diligence. The rails are ready, distribution is the game.

Friction kills. It's the #1 blocker to worldwide crypto adoption. Great diagnosis of common onboarding pain-points and how @BlinkCashX is solving it. If you're building a consumer app, comment/dm for an intro.

Three prediction market analytics platforms pitched us this week, all solving discovery and whale tracking. Very saturated layer.