I've been predicting that the bitcoin value would shrink due to the growing transaction fees, and I've been quite pessimistic all since December 2016. In retro-perspective ... so far my prediction has failed, but anyway ...

Some things:

• gold does indeed have some unique properties and use-cases important enough that someone is willing to pay golden prices using it for industrial purposes. Bitcoin is mostly only unique by being first (network effect) and having a well-known trademark.
• gold has been around for a very long time, and has also been used as a payment vector for a very long time, bitcoins for some few years. It's pretty safe to assume gold will still be a relevant store of value a decade for now, it's not safe at all to assume bitcoin will be relevant a decade for now.
• if another cryptocurrency outperforms Bitcoin and a sufficient amount of people and businesses starts using it as a payment system, then Bitcoin will lose it's network effect. The coin value can only fall from there.
• gold probably gained most of it's current value by being used as a payments system. Bitcoin gained it's value for it's promise to become a useful payments system.

I do not buy the "bitcoin as a store of value"-proposition. I signed up for low-friction, fast payments. That was the value proposition some years ago, today Bitcoin transactions are slow and expensive.

I'm totally aware that blockchains does not scale forever, it's all fine to put coffee-size-transactions on the blockchain as long as the network is small and only used by nerds, but in the long run it can't work out. We do need payments on second-layers. Such a second-layer (like lightning) will needs years to roll out, and it will need to compete not only with on-chain transactions but also with other crypto-currencies. To keep Bitcoin great and valuable, we also need a block size increase urgently. segwit2x for the win!