I wish they'd stop referring to DAG as "decentralized", especially when years later there is no cryptocurrency in sight that relies on DAG that is also significantly decentralized, for exactly the reasons I explained... More than a year and a half ago.
I'd like to point out that in DagCoin, it is still only miners that vote. If you're a user that applies work on a transaction and broadcasts it, then you're a miner.
Work is a waste from the perspective of the miners. There is no incentive, or even a method to incentivize miners to do work on transactions in DagCoin. Competition would atrophy hashpower on the network in order to drive up transaction speeds, which is the product that the miner provides, users enjoy, and the only method of remuneration. Mining is just a cost that the miner wants to minimize.
This is just like if we removed the block size limit from Bitcoin and removed the subsidy.
That problem, the lack of funding for difficulty increases in order to slow down transaction speeds, isn't referenced in the white paper unless I missed it. So ultimately to get back to answering your original question "Why 1 sec, not 1 min?" the answer is I don't see any decentralized throttling mechanism that could actually work in DagCoin.
http://forum.iota.org/t/discussion-removing-peer-discovery/939/2IOTA requires that you manually assign peers because running on it's own it implodes from the bandwidth, because of course it does. It removed Bitcoin's throttling mechanism and now they want to figure out how to prevent these out-of-control blocks in this "blockless" blockchain. Sometimes the real world really is stranger than fiction.