What the @#$@ are you talking about?  "blockstreams 2017 roadmap"?!?!?! Where can I get a copy of this?

And "potential"? The fact that a miner could create contrived high cost blocks is nothing new, right now miners can create blocks that take something like 20 minutes of cpu time to validate. But this sort of intentional bloat-block is almost equivalent to just delaying publication of the block.  What matters for the long term system cost is the average size of blocks, not the potential for a single unusually costly one. Without segwit the unusually costly risks are just different (instead, they take the form of utxo bloat blocks or signature hashing bloatblocks). If your metric of risk is the worst case time it can take to transfer and process a block, segwit does not increase this over the current state: It does allow more data to be sent in the worst case, but segwit transactions have massively faster worst case validation. The end result is that the worst case block with segwit is exactly the same as the worst case block without, you can't use segwit transactions if you want to make a block that propagates very slowly.

Segwit has been public for a long time... Block 0 on the current segnet has a date in January, and segwit is live on testnet now.

Where is Bitcoin Classic's segwit implementation? Hell, where is their mediantimepast and CSV implementations?  Even when they can just copy code from core they can't seem to keep up. If you're going to complaint about something not moving fast enough, it's not core you need to look at...