Background:
Bitcoin transactions are "programmable" in that they support a basic
scripting language. This allows for things like multi-signature transactions and timelocks. That is, the existence of this scripting language allows transactions to be formed which say: "for this bitcoin to be spendable in the future, the following TWO signatures are necessary"... Or: "This bitcoin is not spendable under AFTER this date."
Bitcoin supports a number of basic operations which allow things like that. Stacking these operations on top of each other can possibly yield some pretty complex and interesting behavior. But there is no looping construct. If you're a programmer, you'll understand that immediately (if not, probably need to look it up). Essentially, without looping, many types of things are not possible.
It's critical to understand that Satoshi omitted looping (and various basic operations themselves) very intentionally. Satoshi was concerned with simplicity, robustness, and security, and adding complexity to the transaction scripting rapidly introduces all sorts of very difficult issues. Here's an interesting quote on this from Ray Dillinger, one of the cryptographers who worked on the code with Satoshi before the genesis block:
https://bitcointalksearch.org/topic/m.9171295Ethereum:
Seeks to create a "Turing-Complete" transactions language; ie, one that allows for all the complexity which Bitcoin intentionally omits. It's unknown whether this is feasible without introducing potentially severe or fatal security issues. Gavin Andresen has noted these concerns, as well as the value-tradeoff in the first place:
http://gavintech.blogspot.com/2014/06/bit-thereum.htmlMy own assumption is even if Ethereum *eventually* works, it's going to be very hard to trust it with any real value for quite a long time. It's troublesome, because it won't get battle-tested until there's significant value to attack, and I for one, sure don't want *my* value to be the guinea pig. Bitcoin is experimental enough...
My other issue with Ethereum is that it tries to "fix" non-problems in bitcoin, like coin distribution and mining algs. It comes with it's creator's built-in notions of economic "fairness". It's partly possible to level that same criticism at bitcoin, except that bitcoin is very simple and largely takes a "the distribution will work itself out in the end" approach, whereas Vitalik has specifically tried to engineer ideas of fairness into it (an ultimately futile and counter-productive reflex often characteristic of youth).
tldr: Ethereum is technically interesting, but I wouldn't trust its security for many years after launch (many more years than it took for me to become confident in bitcoin).