Hacker News Reader

Comment by bolt7469 | Hacker News Reader

The classic YC advice is "make something people want." This video dives into how that looks in practice: https://www.youtube.com/watch?v=C27RVio2rOs

The guiding principle is simple, but in practice it can be hard and complex.

View on HN

> The classic YC advice is "make something people want."

Ok, I want to make compilers. People want them but don't want to pay for them ...

Make a compiler that supports a modern language for an obscure chip that is used in industrial machines that is also compatible with the dialect that the vendor's compiler speaks.

I'm a paying customer of the Intel Compiler Suite. And I have been for many years. And I'm well aware that Intel is f-ing over AMD with vendor flags. But there simply is no alternative to MKL.

So if you release something competitive to TBB+MKL for AMD, then you're in business ;) But be warned, it's like millions LOC and 500MB of pure compiled binaries to get to that level of functionality.

Generally, you can go and "invent" business. But if you really do struggle there is another route, not as glorious.

Make something cheaper, its also a good starting point for a lot of "inventive" startups (Spacex is spacex, cause they are cheaper than others). However you don't have to be inventive, just do similar/same thing but cheaper.

I think a more succinct point would be to identify someone with a hair on fire problem. That I feel is more concrete. They do talk about this a bit later but I feel this is the starting point.

Reply on news.ycombinator.com