OK, so none of the usual FUD means anything any more to anyone that is keeping a close eye, but here's at least some food for thought.
It might not work. There is a non-zero possibility that, even with all their talent and hard work and ability to hire people, they just can't make it work. Do you realise what they are trying to deploy here? Hundreds if not thousands of nodes, every node being an extremely technical ongoing process (ie, no hobbyist or "passive" nodes, every node operator has to be completely devoted to operating and maintaining a node). There is a chance (and, depending on you level of LINK fanaticism, maybe a big chance) that a decentralised oracle network in fact presents too much of a technical challenge on an ongoing practical basis. The team understands that if the ChainLink network 80% works then it doesn't work at all. For the sort of people they are selling it to, it can't "kinda" work. It can't work "OK". It has to work impeccably. That means that they have to get potentially thousands of extremely technical nodes up and running, and they have to be working impeccably.
This isn't just "not easy". This is so difficult that even with all the tech wizardry and willpower in the world there is a chance this whole project will fail. Factor in recent question marks about the security of Intel SGX and you realise the sort of thing that could take this down. Even with a successful deployment a catastrophic security failure in the secure hardware or some other aspect of the network could arise only after deployment, and could be so ingrained to be fatal.
You can be excited about ChainLink, sure, but don't count your $100 LINK chickens until they have hatched. Any major and insoluble problem that arises during deployment could take all of the wind out of ChainLink's sails and actually kill this thing dead.