Post-AGI: * Sunsets => beautiful * Dancing => fun * Doing hard things => rewarding * Calls home => meaningful * Fresh flowers => delightful * A baby laughing => unrivaled * Meditation => calming * Skiing/surfing/diving => thrilling The nature of happiness will be unchanged.

This is well worth a read if you're interested in how SaaS companies can turn their narratives (and metrics!) around in the AI era. @shanehegde has put on a masterclass in this over the last six months, and many companies would benefit a lot from the insights here

Growth investors today are either dragging right, exponential surfing, or supply-side investing. Each requires a completely different investing style and craft, and it's important to know which applies. Dragging right is traditional growth investing. You take a business with an established, predictable, repeatable go-to-market motion based on either consumer acquisition or enterprise sales reps/quotas, and the investors' job is to figure out whether it can scale. "Okay, this company has twenty reps- can it hire another forty? Will the LTV:CAC hold?" Greatness in evaluating this came down to understanding TAMs/market segments, the breadth of product-market fit, competitive dynamics (often gentle), and how to analyze spreadsheet data/customer calls for signs of potential weaknesses. This was effectively all SaaS investors had to do for the last ten years, and it was glorious. The entire growth investing industry as we know it today was built around this model of funding predictable g-t-m machines with a fairly narrow range of outcomes- and the skillsets that made someone great at it were all about g-t-m logic, TAM analysis, customer work, etc. Exponential surfing has been the dominant growth investing strategy of the last two years. Consider Anthropic twelve months ago- if you had complete access to the traditional customer data/analytics and spoke to all of the customers, you would find nothing to predict the massive acceleration it saw from December 2025-April 2026. It is literally a completely different game, and most (not all) of the AI companies with material revenue are playing some version of it- not only is their revenue growth hard to predict for investors, it is largely out of their hands as well, dependent on whether the latest models have emergent, high-value, fast-propagating use cases. Greatness here is still emerging, but as far as I can tell the two ways to have been early to the Anthropic inflection were to be using the products oneself (so you can feel the step-function in capabilities in real time) or obsessively tracking the coincident data (Openrouter, credit card panels, etc.) for signs of sustained acceleration. A strong silicon Valley practitioner whisper-network is perhaps the third way. What is definitely NOT great is using SaaS terminology metrics (Y/y growth, gross retention, LTV:CAC, ARR, etc.) to characterize this wave of companies. Yes, many of them are "growing faster than SaaS ever did," but also the base rate of them stopping on a dime (or even shrinking) is much higher. It's just a fundamentally different business model, arguably with no clear precedent. Supply-side investing has risen to prominence in the last 12-18 months. This involves investing in businesses that are pre-product, where the economic value of the product is crystal clear... if it can get built/produced at scale. Neolabs, semis startups, certain very ambitious agent starts, defense tech, robotics startups, etc. all fit this bucket. Greatness here is about understanding technical challenges and unlocks in granular detail, studying talent flows, gauging which teams have the right combination of research and commercialization chops, etc. The rewards for getting it right are enormous (see: pre-revenue Anthropic investors), but the diligence is, again, completely different. What is NOT great is talking to dozens of potential customers who all breathlessly assure you that they would love to buy chips 10x more efficient than Vera Rubins at half the price. That is about as useful as reading LOIs from those same customers to buy them by the billions- aka not useful at all. Referring to those LOIs as "CARR" is, well, fantastical. In supply-side growth investing, the demand side is obvious by definition. Again- there are great investors who will find ways to make money in all of these styles, but it has never been more important to understand which version of investing you're doing and to apply yourself to being excellent at it in a way that fits the game you've chosen to play, not to apply the techniques of yesteryear.

Definitely true that the existential threat to SaaS is greater now than six months ago. The opportunity for vendors who adapt is also massive, but from a technical perspective, the bear case on SaaS is running ahead of schedule.

I've long been impressed that @billmag kept Braze at the cutting edge through thick and thin. It's no small feat to build a $5b company that endures (and accelerates!) in a rivalrous market and still finishes strong, even when its sector is out of favor. Congrats to the team

Checked PostHog for the first time in a while and LogostoSlides has 350+ MAUs (!!). Pretty cool for something I vibecoded on my walk home from work one day.

amen

Great takes but I do maintain that in many cases NY is ahead of SF on practical AI adoption. Elite hedge funds/quant funds are driving it, but they aren't in a hurry to discuss how they use it, at least not publicly. :)

IDK who needs to read this, but if you work in tech and your net worth is 80%+ illiquid stock in a startup, traditional investing advice doesn’t apply. You are exposed to tech at every level, including career. Owning public tech stocks puts another egg into the same basket.

Ari is a national treasure. Sadly, he’s Canadian

Five emerging X accounts worth following: @jhleath - CEO of @archildata, agent infrastructure/sandboxing @RojaGarimella - Kubera CEO/MD, healthcare tech @bstaples - $GTLB CEO, AI for code @shanehegde - @airHQ CEO, AI's impact on creativity @AndrewGSugrue - Avenir founder

This is a work of art and completely accurate w.rt. Maine/New England