Take Solana for example. It takes a powerful computer with 128gb+ of RAM just to run a Solana node. You can run the ETH software on a raspberry pi.
High hardware requirements means fewer people can afford to join the network. This makes Solana more centralized. They also do some sketchy things. For example, in proof-of-stake algorithms the nodes vote on the next block. Solana counts votes as transactions which inflate their transactions-per-second numbers.
The Solana network has also been taken down by a software bug. This is much less likely to happen with Ethereum since there are multiple, independent development teams writing Ethereum client software. A bug in one client will only take down part of the network but not the whole thing.
Let's take another popular competitor Algorand. With Algorand they have two types of nodes relay and participant. The relay nodes take care of consensus. However, you have to send in an application to the Algorand foundation just to be allowed to run a node. There is not a lot of info on relay nodes but apparently the harware requirements are non-trivial.
Cardano I don't know as much about. However, there has been some debate over whether it's PoS algorithm is too similar to DPoS with similar vulnerabilities. Also, the founder of Cardano, Charles Hoskinson, is a known psycho and liar. This isn't just my opinion, just google him.
Polkadot has parachains which is kinda like sharding. However people are also developing Layer 2 rollups for polkadot. This helps prove my point that layer 2 isn't just an ethereum thing.
I don't really have issues with Bitcoin on a technical level, although I think they are too slow to change and adapt. Bitcoin also has multi-dollar transaction fees just like Ethereum which helps prove my point. They are also looking into layer 2 with the lightning network which, again, helps prove my point.
Overall, I really don't see any compelling alternative to what Ethereum is doing. Other chains are starting to look into similar Layer 2 solutions. Other chains are cutting corners to boost their transaction-per-second numbers. And other chains are run by questionable folks.
Thanks for writing up these viewpoints of different chains. I haven't had the chance to check out Cardano, Algorand, or Polkadot, but in regards to the downsides of Solana vs Ethereum, I'll state why I personally am choosing to create applications on Solana compared to something on the Ethereum ecosystem (despite me really liking the tooling that some projects like The Graph provide) -
I'm an applications developer - to me the most important things are lessened complexity, performance, and usability. You can certainly use chains like Polygon, which I might do in the future. However, from my viewpoint, the more complex the ecosystem becomes, the less likely it is to achieve widescale adoption unless there is a lot of masking of the underlying complexities. I've been interested in crypto for years, and been in the space as an engineer for about 6 months, and cannot keep up with how complex the Ethereum ecosystem has grown and how these chains scale. If I cannot keep up, there is no way I can expect for people who've never even touched crypto to be able to wrap their heads around it. I do know that there are efforts to mask crosschain EVM interactions, and personally haven't tried them out too much yet beyond standalone bridges like Synapse. If these projects work out and a year or two from now the ecosystem's complexities are masked, that is fantastic.
For me, I will happily take the tradeoff of making it more difficult / costly to become a validator (in the case of Solana), with the huge performance improvements it provides compared to EVM. While the cost of hardware does mean it is more centralized to entities that can afford to run a Solana validator, to me it is not a concern. It's still decentralized, albeit not as much as Ethereum. I was skeptical of Solana at first until I used it as an end user and am quite happy with the experience, there's a gigantic difference between a 1-5 second transaction time vs 30s - 1 minute (or higher, if you submit a transaction while there's a sudden dip in gas price). I've had transactions stuck on ethereum for an hour and a half because the gas price metamask suggested was too low, there was a 5 minute window I submitted my transaction where the average gas price was super low. Same with Polygon, I've had a transaction stuck in a processing queue for half an hour.
These interactions to me have driven me to look into non-EVM solutions because it truly leaves a bad taste to have to wait helplessly on ETH to process these transactions. Hence, Solana's throughput and promises that they will not need to rely on L2 solutions to achieve scalability (whether that is true is to be tested by time I suppose) are incredibly attractive. It being down for a day was bad, yes, but the mainnet has only been live since last year and is still in beta, and has only happened once to my knowledge. I will still take that downtime given the incredible performance the chain provides. And if it is true that Solana will not require L2 solutions to scale in the same way ETH has, then it's even more attractive to me as an app dev. I think it's likely to be a lot easier to convince the average person to use a platform like this, compared to a complex ecosystem which is multichain only to allow for scalability.
I definitely understand that to a lot of people, decentralization in ETH is the number one driver. However, I do want to emphasize there are certainly reasons to use a chain like Solana, despite having a lessened degree of decentralization.
Take Solana for example. It takes a powerful computer with 128gb+ of RAM just to run a Solana node. You can run the ETH software on a raspberry pi.
High hardware requirements means fewer people can afford to join the network. This makes Solana more centralized. They also do some sketchy things. For example, in proof-of-stake algorithms the nodes vote on the next block. Solana counts votes as transactions which inflate their transactions-per-second numbers.
The Solana network has also been taken down by a software bug. This is much less likely to happen with Ethereum since there are multiple, independent development teams writing Ethereum client software. A bug in one client will only take down part of the network but not the whole thing.
Let's take another popular competitor Algorand. With Algorand they have two types of nodes relay and participant. The relay nodes take care of consensus. However, you have to send in an application to the Algorand foundation just to be allowed to run a node. There is not a lot of info on relay nodes but apparently the harware requirements are non-trivial.
Cardano I don't know as much about. However, there has been some debate over whether it's PoS algorithm is too similar to DPoS with similar vulnerabilities. Also, the founder of Cardano, Charles Hoskinson, is a known psycho and liar. This isn't just my opinion, just google him.
Polkadot has parachains which is kinda like sharding. However people are also developing Layer 2 rollups for polkadot. This helps prove my point that layer 2 isn't just an ethereum thing.
I don't really have issues with Bitcoin on a technical level, although I think they are too slow to change and adapt. Bitcoin also has multi-dollar transaction fees just like Ethereum which helps prove my point. They are also looking into layer 2 with the lightning network which, again, helps prove my point.
Overall, I really don't see any compelling alternative to what Ethereum is doing. Other chains are starting to look into similar Layer 2 solutions. Other chains are cutting corners to boost their transaction-per-second numbers. And other chains are run by questionable folks.