A PREDICTION:
The trouble with ICO's based on bitshares, Ethereum and (probably) SMT: is low block hash speed.
Also these platforms have no parallel processing capability. Thus adding just these two deficiencies together; shows that they have inherent unreliable scalability.
Added together just these two facts will create uncertainties for such ICO applications when they come under pressure of any significant usage. That's when this market will rationalise and have its first major shake-out. Currently its all a feeding frenzy of greedy wishful thinking without much analytical thought or analysis of the technical capabilities of the platform upon which these ICO's are basing their future and that of investors upon.
Belief and hope and chance are very risky things to put your shirt on - so don't.
To build into these platforms, all the required modifications, they may all have to face a series of Hard Forks in the future to correct and re-set . So far there has been little indications of addressing such deficiencies, nor even accepting them. These are the ramifications of reality. Dan Larimer has already questioned the scalability issue of Ethereum with Vitalik Buterin in this video.
Most of the emerging ICO's are based on the Ethereum platform, you have to ask deep questions of people who rush to invest in a map of a possible gold mine, from a drunk in a bar with a stack of photo copies for sale.
Be warned, be careful.
Personaly I trust Dan and my best bet is for the ICO's that will emerge based on EOS.
Take a look here at all the live and up and comming ICO's:
Links to all ICO's
I predict a time when ICO failures will scare this market. Will developers on EOS be able to ride out the storm to show positive value based on a platform that does address the issues mentioned here? YOUR THOUGHTS PLEASE?