3 ms·

I don't get what this does or how it does it. It sounds like they have some mumbojumbo machine that decides who to dole out cash to and how much, but that's unc

I don't get what this does or how it does it. It sounds like they have some mumbojumbo machine that decides who to dole out cash to and how much, but that's unclear. Capital As A Service is, by definition, what VC is. You give capital as a service to your startups. I'm not sure (yet) what is novel about this approach, but if it's "shotgunning money into things based on a set of rules" they might as well just call themselves YC-lite.

jtmarmon 9y ago
Sounds like they're saying they're going to programmatically invest in companies based on metrics alone. Which is probably a bad idea, because there's almost always more to the picture than a few statistics can convey. And the relevant statistics vary by what the business does.
- fataliss 9y ago
  But it allows them to do it at scale, which may or may not reveal itself a good bet. The good thing about their approach is that their scale is only limited by the amount of capital they have to invest. Say out of a 1000 investment based on their metrics, they average a 5% return. Then their model allows them to scale to make it potentially very profitable. Now of course that's assuming that with a larger scale they'd come out positive at all. If not, they'll soon waste all their capital and this will be the end of it :P
  - cdoxsey 9y ago
    It also allows a clever "entrepreneur" to game the system. In 6 months you'll be reading about it here on Hacker News: "How we lost millions due to an algorithmic oversight"
rpedela 9y ago
If you click the two links, it explains in more detail the how. Essentially they have developed a set of GAAP-like metrics for early-stage companies, such as MRR, which they use for evaluation. It appears to be similar to investment firms using GAAP to evaluate public companies.
- SandersAK 9y ago
  thanks, that definitely makes it easier to understand!
tengbretson 9y ago
This seems like it would be very susceptible to manipulation by overfitting what the algorithm finds to be "investible" at the expense of actually being a sound investment. I suppose the human version of this is exploitable in the same way.

jtmarmon 9y ago

Sounds like they're saying they're going to programmatically invest in companies based on metrics alone. Which is probably a bad idea, because there's almost always more to the picture than a few statistics can convey. And the relevant statistics vary by what the business does.

fataliss 9y ago
But it allows them to do it at scale, which may or may not reveal itself a good bet. The good thing about their approach is that their scale is only limited by the amount of capital they have to invest. Say out of a 1000 investment based on their metrics, they average a 5% return. Then their model allows them to scale to make it potentially very profitable. Now of course that's assuming that with a larger scale they'd come out positive at all. If not, they'll soon waste all their capital and this will be the end of it :P
- cdoxsey 9y ago
  It also allows a clever "entrepreneur" to game the system. In 6 months you'll be reading about it here on Hacker News: "How we lost millions due to an algorithmic oversight"

rpedela 9y ago

If you click the two links, it explains in more detail the how. Essentially they have developed a set of GAAP-like metrics for early-stage companies, such as MRR, which they use for evaluation. It appears to be similar to investment firms using GAAP to evaluate public companies.

SandersAK 9y ago
thanks, that definitely makes it easier to understand!

tengbretson 9y ago

This seems like it would be very susceptible to manipulation by overfitting what the algorithm finds to be "investible" at the expense of actually being a sound investment. I suppose the human version of this is exploitable in the same way.