Flow State #1 - Non-consensus bets, AI pricing traps, chat limitations, and more
This week I'm trialing a new format!
Welcome to Edition 1 of Superfluid: Flow State.
This week I’m trialing a new format where I curate the smartest takes on startups, AI, and capital allocation from the 75+ articles I read every week. I’ll give you the key insight from 4 of the best pieces that I consumed over the last week that you can immediately apply to your business (in <2 minutes).
Today's word count: 325
Read time: 1 min 37 sec.
Over the last couple of months I've been working closely with a number of founders on navigating the VC fundraising process and on a variety of 0 to 1 problems (GTM, Product, Operations).
It's something that I've always done as a VC during my nights and weekends, and now it's even better to be able to do it full-time.
Whether you're preparing to raise, stuck on positioning, or just need someone who's sat on both sides of the table to pressure-test your thinking, I'd love to help.
Feel free to DM me on Linkedin, hit reply on this email or book in some time via my Calendly below.
Multiple Axes of Non-Consensus
With AI, we are in an environment where everyone wants to build cutting edge technology. Unfortunately, in a competitive market, that’s not enough to stand out, or non-consensus enough.
writes that at the moment “capital is heavily overweighted toward tech non-consensus - leaving room for under-the-radar entrants that win through market, model, and cultural non-consensus, or better still, unique combinations.”It’s so clear that this is the case, that it makes no sense playing the same game as everyone else. Combine 2 or more axes of non-consensus (see table above) if you want to win.
The topic of VC-subsidised AI credits was always going to emerge as it has in previous cycles with subsidised Uber rides or ‘Unlimited’ pricing with MoviePass. Whilst I’ve spoken about the impacts this has on unit economics, in this article, Chris Paik goes further and questions whether Cursor even has PMF despite being valued at $10B+, or whether users only use Cursor for the subsidies on tokens. The current meta of pricing your AI product at $20 a month will eventually run its course. It’s better to think about your business model and pricing sooner rather than later.
AI UX needs to evolve beyond chat
Chat interfaces are the training wheels of AI, they’re useful for getting started and building horizontal use cases, but “sucks for workflow, depth and longitudinal data.” I really enjoyed this article as Yoni’s article reinforces (+ adds a lot more colour & examples) to what I’ve written before about the need to reimagine AI user experiences from the ground up.
This one was my favourite practical read this week. Kyle Poyar presents prompts that help founders use AI Deep Research to help with GTM research. Good prompting is centred around extensive planning, defining constraints and supplying appropriate context. Kyle’s guide is a masterclass in crafting prompts that do just that.
Forward this to a founder who needs better signal.
Thanks for reading! I’d love to hear what you thought about the new format. Feel free to answer the poll above or hit reply and let me know , I respond to every email.
Abhi