That Ethereum seems superior is just an effect of marketing and language around their scene. It helps they have people who can quickly say seemingly smart things (think GPT-3) that no sane person can parse in a short amount of time to arrive at knowledge. The system is overly complicated for a single individual to understand and trust easily. It may well be very difficult or impossible to tell if a given contract is secure or not.
In reality, there are very few mainstream "killer" use cases being served by Ethereum, although they do entertain themselves by trying to "break" each other's contracts. The most famous case was when the authors of Ethereum rolled back the chain to get their own coins back from a clever "hacker".
The fact is, that individual followed the rules and still lost the coin they were due them by contract! That's not happened with Bitcoin, at least on mainchain.
Ethereum aspires to have fun.
There is a narrow gap where Bitcoin or Gold could potentially be useful if some kind of breakdown happens. But if it's global, you can forget that anybody will be interested in them.
Bitcoin as well as Gold are pretty useless in the zombie apocalypse.
It’s portable and you can trade it without infrastructure. Unlike Bitcoin.
It’s value during a collapse is based on the idea that there will be a future recovery of some kind, but we don’t have to recover very far before gold is good, whereas we need to get back to the level of silicon fabs before Bitcoin can be useful again.
If it’s really an ‘apocalypse’ that we can’t recover from, then you only need one bullet.
Even if Ethereum is at present better technology (I think it is), it's also trying to become even more (scaling and PoS and more). That progress has been slow, and has many competitors with great technology. Labeling Ether as "digital oil" gives one useful mental image, but that's certainly not as sexy as "digital gold".
I attribute the continued market cap dominance of Bitcoin over Ether almost entirely to Bitcoin being (at least for now) a better meme.
OK, so what's driving the buyers of bitcoin to be more comfortable with ever higher prices and is that feeling not permeating over to ethereum?
Bitcoin is a store of value, Ethereum is an universal computer with a Turing-complete language.
I personally use Bitcoin because it has a smaller attack surface, and it’s excellent for my needs. The bug in the multisig contract in one of the Ethereum implementations a few years ago was where I decided never to trust Ethereum as a store of value, as multisig is very important for me. Another important difference is that with Bitcoin I can verify the whole blockchain on my computer, with Ethereum it’s much harder. Also the hard forks in Ethereum means that it’s not yet ready to be a store of value.
Also for Bitcoin I myself was able to read and verify and understand the whole smart contract code base (a few hundreds of lines altogether in 1 file), which increased my confidence.
Bitcoin, on the other hand, builds these core crypto competencies into the base layer, so that scripting can exist purely in the application domain.
My opinion in the parent post to yours is based on my decade of experience with bitcoin, and having been there every step of the way in its evolution and the birth and erratic journey of Ethereum.
In short summary, Bitcoin seeks to provide a secure, stable, no-BS minimal base layer upon which financial contracts of all sorts can be written. The Lightning Network is an example of a smart contract on Bitcoin, one that has been quite stable and secure compared with frothy DAO, ICO, DeFi, or whatever the latest marketing term for smart contracts is in the Ethereum space.
Bitcoin adds small, targeted features to the base protocol after a truly intense amount of work goes in to proving their correctness and the way in which they interact with other parts of the base layer. Ethereum chases after whatever the fad of the moment is and crams everything into a gigantic toolbox of everything and the kitchen sink, and makes you use all that complexity for even things as simple as multi-sig and replay protection.
To stretch an analogy, Bitcoin is a minimal RISC architecture while Ethereum is x86, with all the history of legacy features and obsolete stuff included. Now your task as a cryptocurrency application engineer is to build a secure protocol that must survive without mercy from attackers. Would you rather do this on x86 or a minimal, proven-correct RISC platform?
The choice to me is obvious. Going into the details of why is way more than could possibly fit in a HN comment, however.