1) Yes, proof of work is terrible long-term, mainly because its cost scales with the market cap of the cryptocurrencies it secures. Ethereum is switching to proof of stake, which uses a normal amount of electricity and forever solves this issue
2) There are also a lot of excellent ideas and projects. For example, Gitcoin, quadratic funding, quadratic voting https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3243656
3) It's true that there are many scams/crimes in crypto, and that many or most holders of cryptocurrencies are speculators. What makes crypto different from tulips is that many speculators believe that crypto will come to power large portions of the world's financial and economic infrastructure.
4) Whether a random crypto startup is likely to be a scam depends entirely on your filters and definition of "startup". If you limit your population to projects or tokens with some level of social validation, such as being top-ranked on coinmarketcap or backed by in-industry VCs, then there're hardly any scams at all. Ethereum now has hundreds of quality teams working on many different parts of the ecosystem. For example, the DEX space (decentralized exchanges, eg. https://uniswap.org/) is very different than the layer-2 scaling space (eg. https://optimism.io/)
5) [It's a wild west...] Respectfully, this is an unkind or perhaps bigoted statement. I have many friends in crypto who are thoughtful, kind people. They think deeply about the ethical implications of the systems we're building. The same is true of many leaders in the space.