That is not true, when you have billions of dollars at stake you need to be a lot more careful with the changes you're going to implement, otherwise, it would jeopardize the whole network.
There are many upgrades that have and are happening, some examples: segwit, schnorr/taproot, lightning network, etc.
1) TPS isn't the only thing that matters, 2) TPS on L2 such as lightning is both extremely fast and very cheap. Ethereum is also going to have an L2 because it makes a lot of sense.
It's pretty important for adoption to happen. Right now most Ethereum apps with complex contracts (i.e. Augur, Synthetix) can cost $50+ per transaction.
> 2) TPS on L2 such as lightning is both extremely fast and very cheap.
That's only true if there isn't high demand. If there is opening a channel costs $15 or more.
> Ethereum is also going to have an L2 because it makes a lot of sense.
Something which has been discussed since 2017. The architecture still isn't there. Most apps have to interact on the main chain. Not confident it is going to happen anytime soon even with startups like Matic.
>That's only true if there isn't high demand. If there is opening a channel costs $15 or more.
Sure, but lightning is (effectively) infinitely reusable. You can keep doing as many TXs as you need, and the fee is only to open and close the channel.
I have spent far more than the cost of opening a lightning channel on buying wallets for physical cash, and they aren't nearly as reusable.
They would argue that a chunk of the energy is renewable -- so that makes it better.