I think you're asking "what happens when someone captures 51% of the voting shares?" - This is actually the #1 problem with these types of systems, since the majority could just vote to put 100% of the funds into their own bank account.
The slockit-authored DAO in the article addresses this through two failsafes:
1. Shareholders can vote to split the company in two at any point to extricate their funds from majority control.
2. There is an extra layer of "curators" (kind of a misnomer) who are trusted members in the community who make sure payments aren't being funneled through an intermediary to a majority shareholder.
Whether these safeguards will work remains to be seen.
Anyone can 'split' the DAO at any point of time, which renders the 51% attack moot. A 51% attack works only when the attacker can get more than what he puts in.
Consider this obvious attack vector: Let's say today is the last crowdfunding day, and the DAO has raised $150 million so far. A wealthy attacker can immediately put another $150 million, effectively controlling 50% of the voting for the DAO, and then vote to send all the money to his own Ethereum address. This is perfectly possible, since this is all written in Ethereum contracts anyway.
To avoid this, the rest of the people can 'fork' the DAO, effectively leaving the attacker with 100% control over his share of 50%. The attacker gains nothing. The other 'fork' can now continue on.
This forking described above is also going to be used (I think) when it comes to funding new projects. I am sure some people will disagree with some proposal, and they would fork the DAO away to not invest in that proposal even though the majority agreed they would. They are free to do so.
Well, I'm not sure I'd use such absolute language. For one thing, people can still perform a 51% attack and try to influence decisions covertly, without triggering a split.
Also, the spectre of a 51% attack required for some major compromises to be made in the DAO design- Having a company that can arbitrarily split at any moment is likely to have a major impact on the future operations of the DAO.
You're right, I shouldn't have used an absolute language. There is also the issue of voter apathy so yes, a 51% attack might be worthy of a try for a wealthy enough adversary.
In terms of the split, if the community broadly agrees that there is an attack, then I don't see it having a long-term detriment. However, as you said, if it is more 'stealthy' then it would be harder to detect and correct.
All that being said, I think there is a theoretical safety mechanism in place. Whether that will work in real life or not, we'll wait and see. I am sure some good lessons will come out of this in either case and the next generation of applications can improve upon some shortcomings.
>2. There is an extra layer of "curators" (kind of a misnomer) who are trusted members in the community who make sure payments aren't being funneled through an intermediary to a majority shareholder.
If there are humans involved that can exercise this level of control over The DAO, is it really autonomous?
I'm not 100% familiar with it yet, but there is some code where anyone on the losing side of a majority vote can opt to fork the DAO and retain their original tokens with all investments, instead of going along with the group (similar idea to a Bitcoin hard fork). This prevents someone with 51% ownership voting to send 100% of the assets to themselves.
