Startup Playbook
playbook.samaltman.com
playbook.samaltman.com
The way to have a buyer before founding is to have worked at an established, profitable firm that has that exact problem, cannot build a solution themselves, and get a signed agreement from them to buy your future product. If you leave that firm and have that document, you can raise money, then use that money to hire a team, then build the product and fulfill the contract.
Getting into such a position requires an established career. Therefore safe, reliable entrepreneurs cannot possibly be college drop outs - they need to be established careerists with good connections who have entrepreneurial ambitions.
There’s also the issue of trust. For some reason, managers tend to trust third parties more than their own employees. For example, they won’t think twice about using open source libraries with a single hobbyist maintainer, but often feel uneasy about using an open source library built and maintained by one of their subordinates. I’m not sure why this happens. There must be a name for the phenomenon. I suspect the same thing will play out when you try to pivot from employee to supplier.
Blissful ignorance. If we actually knew the library has only one single hobbyist maintainer, we would be worried. But often we intentionally don't look at the fact. We just tell ourselves "well it's open source so there must be multiple eyes on it..."
Anecdote warning: I'm a college drop out. Joined a 5-person YC company when I was 20 as an engineer. 2 years later I was halfway through building a product (my current company which I've been building for 8-9 years).
I took a very similar path as one you're describing, except not as formal. The company was a tiny startup. They became my first customer (while I still worked for them!) which validated the need for the product, but I only charged them maybe $40/mo.
Then went on to a startup accelerator, funding etc.
All that to say the spirit of your comment is on target, but the specifics can be executed in different ways. No college degree required!
It might seem safe, but you can still screw up.
Most succesfull businesses I've seen started in my close circle indeed came from the job they had before.
Most of them were not new products, but either were what they were already doing or a variation or improvement of that.
In that case you know your product and often already have it, know your market & know how to reach them.
And often your former employer either becomes a client, or he is upset :).
But that's maybe not really a startup.
Creating a new product is often what is most difficult. It takes time to figure out what to create, who to create it for and even if it's a great product how can you affordably get enough new clients. That's a lot of stuff to figure out.
So yeah it's way easier to expand on what you already know.
Just make sure you are not captured by your first buyer. A product built for one customer often has little value.
You can succeed in achieving everything in that playbook, and still fail.
Life can be brutal.
Appreciate the importance of timing & luck.
Stars need to align & lots of things are out of your control. Like an unexpected global pandemic, a persons health turning bad, or simply a key customer not paying timely - are all things that can tank your perfect business and more.
If you do nothing - you won't succeed. If you do everything - you may succeed.
Life is a wide-ranging buffet of risks, just like the stock market. You can choose anywhere from bonds (career, stable job) to penny stocks (startups).
According to web.archive.org this is from Nov 2015:
https://web.archive.org/web/20151108010026/https://playbook....
> Startup Playbook (November 5, 2015 — 1202 points, 232 comments):
https://news.ycombinator.com/item?id=10514729
> Startup Playbook (2015) (January 16, 2019 — 640 points, 55 comments):
Companies that succeed may not necessarily do all of these things. Companies that fail may do all of these things.
A person who'd be inclined to click this link is likely new to startups and wants some kind of structure, knowing how unpredictable things can be. In some sense, this is exactly "what people want" ala YC advice.
As long as they say a few milquetoast things that other established entrepreneurs cannot argue with, they have essentially used the (possibly random) initial successes they've had to build credibility in the eyes of potential founders.
There is no doubt that a large part of YC's successes are due to its network effects but you first need to get lucky in order to build it. That's the part they seem to leave out.
For example, when I was young I worried that ideas were precious and had to be protected in case I told someone and they stole it. Now I recognize the fundamental truth Sam expresses in the article: the best ideas sound bad, and the majority of people will just roll their eyes if you tell them what you want to do. I've directly had that experience with several successful projects so this isn't just nodding along to something that sounds vaguely aspirational.
Whilst OpenAI seems to violate a lot of what's in the article, in that way it's example of it. They're in their current position because groupthink within the AI space came to state that scaling up language models wasn't worth doing and academically uninteresting compared to developing more sophisticated neural architectures. OpenAI said, no, let's just try spending shittons of money and engineering on brute forcing enormous model sizes with an ordinary-ish transformer network and see what happens. Right up until they started getting these amazing results they were sort of out on their own in being so committed to that approach.
The best ideas sound bad. Really bad ideas also sound bad. Many good sounding ideas are also hugely successful businesses.
Maybe the etiology of success had nothing to do with how bad sounding the idea is?
Also in the case of OpenAI, being in the field, I'd strongly argue that it wasn't that crazy sounding of an idea as people might conjure in retrospect.
Google had been throwing resources at LLMs and been using them in production for several years before OpenAI started to do it. MANY papers written about scaling laws and how we haven't reached the limit of scaling data / compute yet.
OpenAI is only relevant because they iterated quickly on releasing a product. Absolutely nothing visionary about the technology honestly. The question that should be asked here is why OpenAI was the first to release a product a few other companies were capable of creating..
Everything OpenAI has done is low hanging fruit. The logical progression of LLMs is chat, zero-shot learning, multimodal, connect it to web / other knowledge bases, planning, personal customization, multi-agent systems, etc. There is a vast amount of research in several vectors in this area..
> They are decisive, which is hard when you’re running a startup—you will get a lot of conflicting advice, both because there are multiple ways to do things and because there’s a lot of bad advice out there. Great founders listen to all of the advice and then quickly make their own decisions.
The rule of thumb I follow is that if you're not sure what to do, follow whatever mainstream advice is out there. But if you're feeling confident that you should do X, then just do X and don't bother listening to anyone else
"First create a non-profit then make it for profit once it takes off."
And before anyone criticizes me for defending his viewpoint, let's assume this: if Musk's $100 million investment were to be replaced by 100k HN users investing $1k each into a non-profit focused on safe and "open AI", do you think they would be uniformly happy about this decision?
An open AI is safe inherently, since it means that it can be easily ripped apart and thoroughly studied for exploitable points, unlike some closed black box system. Having Open AI be some closed system does nothing to reduce the number of bad actors - they will all choose to exploit Open AI's system once given the opportunity.
Replace the word "AI" with "ultra deadly and contagious bioweapon" and it comes clearer why being "open" is itself a danger, for those who aren't able to zero-shot understand it.
And yet...
For those unfamiliar, YC now runs its own “cofounder dating thing”:
This post may not be the most in depth thing in the world, but I'm sure there are plenty of people who could learn something from it.
But I wish I had read the bits about cofounders and exec teams.
The number of pitched battles I had about management processes in my pre-revenue, pre-customer, barely-there startup took so much of my time, while I was just trying to get the product out the door so we could sign people up.
I definitely chose the wrong parters... and it definitely sunk the company.
So really the advice is, be born into the right family, or at least get into to the right university, join the Good Old Boys network, and fail upwards. The old "two dropouts in a garage" idea of a startup immortalized by Steves Jobs and Wozniak is very much a romanticized 20th century phenomenon. Nowadays it's more like the old East Coast financial world of clubs and school ties, albeit with a different uniform of jeans and hoodies.
I wouldn’t exactly consider being acquired for $43M a failure.
I have seen/worked a few enterprise SaaS companies where the customers could not barely use the product without handholding from a customer success rep, let alone love it. Love YC's advice for indie, first time founders but a lot of it falls flat if you know you know a guy, who knows a guy.
> The former works better for consumer ideas [...] and the latter works better for enterprise ideas (if a company tells you they will buy something, then go build it.)
> Specifically, if you are an enterprise company, one of the first questions we’ll ask you is if you have a letter of intent from a customer saying they’ll buy what you’re building.
If you're a new startup building a B2B product, how does that work in practice? It's difficult to imagine an unproven, early stage startup getting letters of intent from customers. How would you even go about finding those customers?
Neither are easy, but the latter is just the former but with more steps and you'll probably have to make a ton of changes to your original solution before anyone buys.
1. They think the letter has to say "I will buy this today" instead of "I will buy this if it does xyz" 2. People think someone seeing a product (e.g. build something great is all that matters) trumps everything.
Everyone has more problems than they can pay attention to, so mostly you are probably looking at people you have direct contact with. Competing in an industry you have no experience in is going to be a very different experience.
This goes to the team. If you want to build a startup selling to enterprise, it helps to have a co-founder that has the skills/experiance to do this.
If its a big B and you need credentials, you'll need to get that first. You can do it by looking and sounding professional, bringing in advisors who add some weight, getting an intro from the CEO or owner (you'd be surprised how many will pay it forward). If its a small B just wander around and say hello.
People generally want to help people who are making things that help them. And your not getting anywhere without understanding your user so you'll need to talk to them, emphasise. A lot of the earliest discussions will be more around you figuring out exactly how your product will help them.
Is that why YC went so heavily into cryptocurrency startups?
/s