How to Start Google

I am glad he did not mention "Ensure you are born in California".

But topic is "how to start your own Google" not "how to run your own Google".

Still pivoting for monetisation is important.

That's not the comparison I was making. I said consuming a meticulously curated list of the best university lectures on any topic delivered by an extremely rare breed (unparalleled research professors who are also competent as educators) is a better way to consume university lectures and requires a phone and $15.99/month.

I furthermore said that Grant of 3Blue1Brown fame and others inspired by him (Richard Behial, Sean Carrol, and many others, this matters enough that I'll do the rigorous citation routine if people personified in this instance by you don't regard this as common knowledge) have used the lever of technology to increase the upper-bound on the efficacy of education in a way that dwarfs the technological lever that LLMs represent for being wrong at scale persuasively enough to convince anyone but an expert in a bad mood.

I explicitly acknowledged the dominant term in the kinds of outcomes pg is talking about, which is having and effectively managing relationships with people who either already are or soon will be with high probability powerful.

In 2013 I was shortlisted as a candidate for some of the top jobs at the hottest company in the business and on a polynomially steeper career ascent trajectory than someone who now sits on OpenAI's board: that person diligently managed relationships and had a keen sense for whose star was rising and falling, I produced 10^4 more marginal revenue (very conservatively) but gave the establishment (including too many of my colleagues who chose the other path on geek/climber) the finger and didn't manage so much as my personal life, let alone the establishment. Which is why that person is on OpenAI's board and I'm scratching out a living as a consultant, and even that is only possible because I'm finally learning twenty years too late to be at least thoughtful, reasonably polite most of the time, and well-cited when being skeptical of an establishment.

The most effective way to interact with cutting-edge researchers short of actually getting a PhD at Oxford or something is to work for or consult for public-private partnerships between top academic labs and private-sector spinoffs. Even as a marginalized consultant I've worked directly with people at top research universities who get whispered about as Nobel candidates, these companies are often under-resourced in the beginning and if you're willing and able to represent a bargain, you can work adjacent to some truly stratospheric levels of elite research.

You're not wrong that going to an excellent school that you or your parents can actually afford is among the best advantages one can have early in life, you're exactly right about that.

You're incorrect in demonstrable ways on the rest of the stuff you've bundled into one statement. This is the problem with bundling together tangentially related ideas into one atomic unit.

I've kicked this thing back and forth a few times because even on revision I'm finding it difficult to say "What you Can't Say" [1] with both of: concrete and documented examples and low scope for seeming rude or snarky. I don't intend snark, but pg wrote those essays for a reason: being right is offensive sometimes. So I apologize if I give any offense to you or anyone else, but all this is true and a lot hangs in the balance right now for the industry and civilization as a whole. And I regret that you personally wind up being the person who said the thing that many/most think that needs demolishing, there's absolutely nothing personal in it.

[1] https://paulgraham.com/say.html

Getting people to pay for something isn't the same thing as making the world a better place. You'd need to figure out all the effects and costs of the service to know if you're doing good or bad. Tobacco companies sell things that a lot of people desire, but the effect on society is surely negative. Same could be true for social media and many other types of services, but we don't know because we're not really trying to figure that out. The only question we ask when we are starting a business is "are people willing to pay for this enough to make a profit?". We don't ask if the effect of that service is actually good for society or the world.

This does make sense. I think it's also possible that most of the successes PG sees are people doing something interesting to them. So the argument would be more like: among success stories, true interest is more common than trend chasing.

Of course YC funds tons of ChatGPT wrappers so dunno.