Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

(I'm the post's author)

> The whole point of a startup is to make a bold bet on something not entirely proven and safe. That doesn't mean take DUMB risks, but this piece actually says you should try to get your business to the sort of low risk situations I list above.

I agree with you that startups are often making bold, unproven bets. The post was trying to say that yes, you're starting somewhere risky, but how can you de-risk your assumptions? How can you start proving the unproven? The proof might be a 5-10 year process, but it's important. The goal is not to reduce risk for safety's sake, but to validate that your end goals are reachable.

For example, if you came up with Snapchat 10 years ago, there would be many risks, including "do people want ephemeral messaging?" and "can an engineer build this product?" I would argue that the first risk is much more important to validate, but a lot of founders -- especially tech founders -- would focus on the 2nd risk. Using terminology from the blog post, the first risk starts out at a 1, and the second starts out at a 4, but too many people would focus on moving the 4 to a 5 instead of moving the 1 to a 3. A 3 still isn't a home run, but at least you know you're on the right track.



The risk is rarely CAN it be built. I think your point is to demonstrate it SHOULD be built by following de-risk steps.

Lean, right?


Yep!


To heck with SnapChat. Gads, it became a fad as young women stood in front of their bathroom mirror and used their smart phone to take imprudent pictures of themselves. No one knew that this would be a fad, but anyone could guess that the fad had a lot wrong with it, would get a lot of push back, and would not last very long.

Instead of all that nonsense, to get around the 'product/market fit' risk, here is how: Pick a problem where the first good or a much better solution will be a must have for a sufficiently good, large, dedicated, whatever audience of users/customers to make a business worth $10+ billion.

The classic would be a safe, effective, cheap one pill taken once cure for any cancer. So, have a problem, cancer, and the first good solution, the one pill, and it is a must have for a lot of people -- ballpark half of the population will die of cancer before anything else.

So, with that pill, have product/market fit so strong that just on a rumor the front doors of the company will be mobbed by people desperate for the pill.

That situation is what you want to aim for.

Well, everyone can see the problem of cancer, but so far no one knows how to make that pill. If making the pill were easy, then that problem would have been solved by now.

In information technology, can do much the same thing: The problem is already out there and fairly easy to see.

But, there is no free lunch here, no royal road, no ten easy steps. Instead, if lots of people know how to solve the problem, then it would have been solved by now.

So, need to attack a problem where not many people know how to solve it. In particular, need a problem that just routine software is not nearly sufficient for a good solution.

Next, when have the problem and the solution, the risk is already way, way down, but have yet to cover ANY of the risk reduction techniques in the OP.

So, the real key to risk reduction and high ROI for startup projects is to have good solutions that nearly no one else knows how to do.

For now, for information technology, the best approach to such solutions is original research in applied math, typically based on advanced prerequisites. An advantage is that such work can be reviewed reliably. And passing such review, now have a low risk project where the rest of the work to high ROI is routine. That's the goal, right? I just gave you the magic, golden, secret sauce, right?

THAT'S, for now, how to do information technology startups with high ROI and low risk, right from the beginning, that is, based on, say, a project proposal on paper with the math included.

The ante in that game is to be able to read, review, check, at least direct such work, in the math.

I doubt that there is a single VC firm in the US that can do that. The NSF can. So can ONR, and DARPA. So can high end journals in applied math. So can the applied math departments of leading research universities. But VCs? Nope. With at most a tiny number of exceptions: History majors from Williams College? Nope. Computer science chaired profs at leading research universities? Nope. Silicon Valley entrepreneurs? Nope.

So, we're talking something rare and exceptional. Well, we know that except for luck that is a necessary condition for $10+ billion. So, we have to be able to work effectively with things that are exceptional.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: