Yes, personally I agree with this. This is also why artists will mint their very best work on Ethereum mainnet only. Also indeed noticed L1 fees are very low.

Some of the zillion L2 rollups are so flaky, centralised and with almost no users, I guess we're just waiting for them to halt one day or worse.

It's actually pretty crazy how you can now just pretty openly say that ... mayyyybe the L2 roadmap isn't as great as we previously thought it was

I almost feel enabled to even say: "hey you know what, I think the L2 honest node assumptions are pretty impractical for on-top-of-L2 protocol builders"

It's interesting where we'll actually end up going here because if a lot of the scaling ends up being too decentralized we might see a reemergence of demand of L1 blockspace lol

I'm not sure I get that, so please correct me if I am wrong. Mainnet has now about 1 Gwei in gas fees. Let's say some DeFi Interaction would need 200k Gas and take off (just a bit) on mainnet with thousands of people interacting with the contract. Probably not unheard of. AFAIK the block gas limit is still sitting at 15M (?) to keep the base fee or up to ~30M Max(?), that's just shy of 150 TX per 14 seconds on that contract that would fill up the block and increase the base fee (tremendously probably). Please correct me if I am wrong. And I sincerely hope I am.

Let's look at the same on base for example. 120M block gas limit and 0.05GWEI per gas and a 14x faster block time (afaik 1s). It would take both longer to raise the base fee whilst being cheaper along the way and the txpool is emptied faster. I haven't checked, but I assume on Arbitrum the picture is pretty similar.

I get the drawback of the L2 with basically being owned by an entity, but really, for heavy usage imho the mainnet is the wrong place. It's the right place for asset storage, but if you need to put these assets to work, why bother with a super secure chain that's being congested rather quickly.

OK, but if it‘s not decentralized then it‘s not decentralized and then having congestion is ok

No but in all seriousness, e.g., if you ship on multiple L2s then u might also have to back up multiple L2 sequencers etc., so it all quickly becomes not as easily to decentralize as people like to admit. And there are other complexities too

I do understand the gas limit calculations of transactions that we service on e.g. an L2 etc.

Rationally speaking, I think we'll have to understand what belongs where at the end of the day. At the same time there are pretty big technical limitations. E.g. aren't "you" regretful of launching your L2 coin when it reaches 1M USD mcap because once the L2 shuts down you'll have troubles exiting these funds etc.

E.g. the numbers you're currently quoting on L2 gas consumptions are therefore potentially also just plain unsustainable and wrong. If all of gas consumption on Base is related to people launching memecoins there then this is a pretty risky thing and its unclear whether those gas consumers even have a good understanding of their risk taking. So we also have to act responsible as part of the elite developers and we have to say, "hey maybe it's not good that your > 1M USD memecoin actually cannot exit the rollup" and that "this is creating a ton of risk," be that congesting the L1 or not, it's just a fact.

I'm wondering how many L2s would become like zombie alt-L1s. Given that most of them don't have any good GTM, and just copy DEXs and yield farms from other chains, it seems likely.

On the other hand, wasn't the whole idea to run these L2s like startups, where most of them will fail, but for 10 Blasts we get 1 Base?

BTW thanks @tomw1808.eth calculations, they put things in perspective.