There is one crucial difference between the startup scene and the banking/financial industry - they operate under diametrically opposite risk models. SV's model results in wealth creation and localized losses, Wall Street's model results in debt creation and systemic instability.

In Silicon Valley, investors spread lots of little bets around, knowing that most will fail, a few will break even, an even smaller few will hit it big (like Heroku $200m exit), and a remotely small few will make billions (FB, Google, etc.).

If ~90% fail, 9% break even, 0.99% are Heroku, and 0.001% are Apple||Google||Facebook, investors make a lot of money. That's the driving incentive and motivation in SV.

Conversely, while Wall Street does that too to some extent, much of their modern money machine has come to rely on trading funded by creating or borrowing as much debt as possible and then using it to leverage returns on risky ventures, like prop trading. It's the house of cards model.

When SV investors invest more, they increase their odds of hitting that elusive FB or Google event. When Wall Street increases its leverage, it increases its odds of another systemic financial collapse.

It's been happening to the banking industry for all of history (with the exception perhaps of the ~76 years that Glass-Steagal curtailed too-big-to-fail) - picking pennies in front of steamrollers. Eg, chasing returns at greater and greater risk and fragility.

So yeah, there's a lot of greed in both places, but the SV model harnesses it in the service of wealth creation and innovation, even if there are periods of uncreative me-too social networks and whatnot. Aggregate losses are predictable and manageable, aggreegate gains are unpredictable but potentially unlimited.

But the Wall Street model completely inverts that, and puts greed into the service of a small cohort that profits from increasing debt creation, until the music stops, the debt collapses, and they lose more money than they ever made [see #2 below] and require massive taxpayer bailouts to save the entire country or world.

Gains are marginal and come at the cost of increasing systemic risk, while losses are unpredictable, potentially unmanageable, and system-wide (not localized to those who took the risks).

More succinctly, the Silicon Valley model inherently bets on positive outcomes, especially remote but inevitable extremely positive ones like Google and Facebook.

The Wall Street model inherently bets against negative outcomes, especially remote but but increasingly likely [the bigger the bet against it the more likely it becomes] extremely negative ones like systemic debt collapse.

Here are a few charts illustrating Wall Street's model [1]. Invert them and you'll have charts of SV's model:

1. Anatomy of a Blowup

2. Anatomy of IndyMac's '07 blowup

[1] Source.