I'm not personally against them ideologically, but the article kind of proves why we shouldn't do them. As you note, the really important smart contracts get audited and then coded into the blockchain, and as you also note, applications like steemmonsters (and d.tube and dlive.io etc.) exist without them. "But more decentralization must be better" is not a compelling argument to me. If it was, I wouldn't be involved in a DPoS chain. As with every decision we make it comes down to developer resources and scalability. There's a reason why we have more transactions per day then every other blockchain combined and are capable of supporting those txs. While some might think that Smart Contracts are unrelated to scaling ... nothing is unrelated to scaling.
Some of us have maintained from the beginning that Smart Contracts are overvalued, and that Bots+Private keys can get you to all the same places. And if your application really needs Smart Contracts, then there are plenty of protocols out there for you. We are an application-specific protocol. We are the best protocol by far for organizing and incentivizing content. This "simple" design is key to scaling. General purpose protocols are far more difficult to scale, especially when they are open source.
I've yet to see a compelling counter argument and the staggering lack of scalable Smart Contract blockchains as well as serious applications on such protocols is not something that should be ignored. Ethereum blew my mind when it was announced and was a serious trigger for me getting involved in the space (after Bitcoin), but that was because I wanted to see DApps. From where we're standing today Steem remains the only blockchain with real DApps and it did that without "Smart Contracts." Or rather, it did it because it has the right Smart Contracts built into it.
I think this is a good conversation to be having (it's not the first time it's been had) but for me personally, I don't think we should even be thinking about Smart Contracts until Communities and SMTs are out there and then we can reevaluate.
I have to agree with @andrarchy here. Steem + smart contracts would just be a worse EOS. Steem is an application-specific blockchain. It's like an ASIC miner vs a GPU miner - GPU miners (Ethereum/EOS/etc) can be made to do anything, but they will never be as good or efficient at the one specific task of an ASIC miner (Steem).
Yes, it means that Steem Monsters cannot be a truly decentralized application (though it can be come surprisingly close using the methods currently available), but it also means that it will be the overwhelming best option for services trying to monetize/reward content publishers online in a decentralized manner.
I think this can be summed up with a quote (I don't remember who said it):
Great points, the ASIC miner analogy is perfect.
Not totally sure who originally said "Do one thing and do it well", but I believe that's the UNIX philosophy / mantra, isn't it?