Been a while since my last report and given the recent leg down, wanted to share a prior $MQ pitch now that it's less actionable.
Think this is consensus view but was my first look at fintech, probably still meat on the bone but less ideal risk / reward.
https://t.co/dy1S6yHsv5
To replicate marketplaces like $DASH or $AXP you need to replicate the demand and supply side. AI apps will do the demand side work for you, so to compete w DoorDash you just need to build the driver and restaurant network. The biggest competitor will likely be direct restaurant delivery vs using DoorDash.
Alap here, coauthor of the piece w @Citrini7. Imagine in October you can just prompt ChatGPT to re-order your favorite tan-tan noodles from your local ramen joint. The AI can price shop across DoorDash, Uber Eats, Grubhub and Instacart and directly to the restaurants site, choosing which ever offers the lowest price.
As AI companies start driving real volume to restaurants, they will start offering a demand auction that each restaurant and delivery app can tap into. Today DoorDash takes 25% of the total order, charging both the customer and restaurant. So instead the ChatGPT agent could transact with the restaurant's bid, with ChatGPT taking a 7% cut, restaurant only paying 7% (vs typical 15%) and the consumer saving 11% vs buying through DoorDash. ChatGPT could run a second auction with delivery drivers for fulfillment.
This works when AI apps have sufficient volume of orders, which is likely a 2027 event. Though Joe I do think this also partially answers your second question. As AI apps scale these auctions, they will build network effects in the same way Google did with adwords. Perplexity's auctions will have a lot fewer restaurant participants compared to Gemini and ChatGPT.
Cowork is now available on Windows.
We’re bringing full feature parity with MacOS: file access, multi-step task execution, plugins, and MCP connectors.
Have seen some theses around context graphs, aside from the data gravity hurdle think this ultimately requires deep behavioral changes from stakeholders to provide the info for agents to be useful ("context" isn't just sitting in some hidden repository), so in near-term we probably get early adopters (with inherently smaller GTM teams in this case) that adopt this new workflow but the largest enterprises (and sales teams to charge) are probably staying put for a long time on $CRM. Inclination is we get fragmentation in this market and upside for startups comes from capturing nascent companies that eventually turn into large enterprises.
@ChairliftCap Incentives matter - if you have a pile of cash and LPs are requiring deployment, it's going to find a home one way or another. It's a different game.