Mention Oceans event

Hey y’all — Paul Graham is a legend.

As the co-founder and primary founding face of Y Combinator, he’s overseen countless unicorns emerge from their earliest days.

He sat down with Vivian Shen, a current visiting partner at YC, for a talk on ambition, and where he sees the startup world being at in 2026.

Watch it here, and read my takeaways below.

Btw, as we head into the fall, I’m looking to do more IRL private/curated dinners in NYC, SF, and elsewhere. Let me know you’re interested in joining here.

I took an AI class in college. This was 2013-2014.

We built (now what would be considered basic) software for little robots so that they could play soccer.

Most of what we learned, and what had been the dominant assumption in AI research up until recently, was that AI would “arrive” through building very narrow systems. Predictable environments that it would gradually get better at navigating, and then expand to larger and less predictable environments.

The opposite happened.

Instead of going from a mouse in a maze to a cat in a house, what we got was something that’s broadly human-level but full of BS.

AI is a confident undergrad bluffing through their research paper.

PG calls this the “jagged frontier” where AI is already better than us at some things, but laughably bad at other basic ones like factual reliability.

PG has been at the forefront of hyper-successful startups for well over 20 years now.

What he’s observed is that most of what matters when starting a new startup is the same regardless of whether you’re building with microprocessors, AI, or consumer apps.

Technology changes, yes, but the hard parts about building (finding something people want, shipping fast, dealing with obstacles, and not quitting) do not.

In particular, even with powerful AI tools many startups still ship too slowly in PG’s view. Idea generation and judgement still matter too. The main new operational difference is that large AI bills can now rival human headcount cost.

He noted how YC itself has changed remarkably little.

Yes, cohorts are bigger now and there are more of them per year, but he looks at this as “the same thing but more of it” and notes that there have been criticism the program has jumped the shark since 2008.

This was an interesting one.

PG commented on nostalgia for “early YC” where companies like Reddit, Airbnb, Coinbase, and others got their start.

He flagged Reddit as “valuable, but not world-changing” and compared it to recent batches that include genuinely extreme ideas like intercontinental ballistic cargo, multiple approaches to curing cancer, etc.

Logan Roy would likely say that YC is now full of “serious people” moreso than the early days, and PG says this is a good thing.

Founders need that ambition (which PG says is mostly internally born rather than brought out of them in YC) and should be “formidable” (PG says this is the key founder quality) to get through those grueling parts.

Talent or “beautiful ideas” are not enough, and the next trillion dollar companies will come from the right founders more than from a particular idea.

I’ve also noticed this in my investments. My biggest wins and biggest misses have all been founders who no one would hesitate to call formidable.

PG commented on the increasing trend of high-achievers treating YC like a career credential like getting into Harvard.

He said these people typically romanticize founder life and usually don’t understand how brutal it can be.

We see bigger and bigger rounds happening these days even if a product is very early (Instinct is still in beta but recently closed a $2.5B valuation round).

But PG says you can still start with relatively little money by adjusting the scope of what you build to the capital you have, regardless of what niche you’re in.

He also expects inference costs to fall dramatically over time.

Even though YC started somewhat accidentally (it was intended as an “angel firm” in PG’s words, where the batch model just proved to be so powerful that it became the core), it has obviously become a huge force in the startup world.

And those advantages are compounding and structural, in PG’s eyes, due to three key reasons:

Advertise in my newsletter to get in front of 60,000+ founders.

Founders: treat yourself for once.

A few of our partners are offering $$$ for taking a quick call with them (that’s it).

Honestly, block off a couple hours, knock them all out, and walk away with some cash to get off the ramen diet for the weekend.