I can't believe I still have to say this in the year 2025, but let me explain once again why permissioned chains like Tempo (or Canton or ARC or any other corporate chain) are likely to fail.
Most people think Satoshi wanted to invent a new kind of money, but I think he was more interested in creating a new kind of payment system.
The most astute TradFi-focused observation in the Bitcoin white paper is this one about intermediaries:
"Completely non-reversible transactions are not really possible, since financial institutions cannot avoid mediating disputes. The cost of mediation increases transaction costs, limiting the minimum practical transaction size and cutting off the possibility for small casual transactions, and there is a broader cost in the loss of ability to make non-reversible payments for nonreversible services. With the possibility of reversal, the need for trust spreads"
TLDR: When everyone knows who is in charge, and that entity could plausibly be held liable in a civil court or by the government for facilitating crime, they have no choice but to censor, rollback some activity after the fact, and so on.
In modern crypto terms it means no true liveness, safety, or CR. Before Bitcoin, this is how every single settlement or payment system--save physical cash--worked.
What made Bitcoin unique was the way miners voluntarily opted into a protocol that was the final boss. They didn't have to identify themselves or be permissioned, and faced financial penalties if they didn't follow the rules.
Thus: when any Bitcoin transaction happens, we don't necessarily know who the miner is. Even if we did (much of the hash rate is known) it wouldn't change anything because they can't be held liable. If any aggrieved party sued a specific miner for a particular transaction, that miner could say "I'm just following the rules of the protocol here. If I didn't follow them, someone else would, and the same transaction would still get processed"
Thus: an open, permissionless and decentralized protocol is in charge of everything, and since you can't sue or jail or kill a protocol, we end up with liveness, safety, and censorship resistance.
Same goes for a permissionless PoS chain.
Or even a properly designed L2 with rock solid proofs and L1 defection/CR mechanisms. The sequencer can plausibly claim "I'm just following the rules here, if I try to force an irregular state change the proof won't be accepted by the permissionless L1 validator set. If I censor a particular user they'll force include via the L1"
Fifteen years ago one could wonder whether this would actually work. What if the government decided to throw some miners in jail anyway just to deter others from joining the protocol?
Well, they haven't. Probably because the US realized that going after American Bitcoin miners would achieve nothing other than making non-American miners (and the governments that tax them) richer. People like me have been telling them that for a decade.
In fact the entire arc of regulatory interaction with crypto now is to respect decentralization, and even encourage it.
Now, back to Tempo, et al:
A permissioned network does not have this form of plausible deniability. The permissioning entity has to KYC every validator - not only are they known, but their participation is not voluntary.
Given this practical reality, the protocol is not in charge, the gatekeeper is. The gatekeeper can change the protocol whenever it feels like it - and coerce all validators to go along, lest they be de-permissioned and kicked out.
In a permissioned chain, the protocol is more of a "best practice set of recommendations" than it is something inviolable.
This is a problem because it returns us back to the ass-covering hell Satoshi identified. Both the participating validators and the gatekeepers can be held liable because they have the power to violate liveness, safety and CR whenever they feel it.
And because they are all run by smart professionals, and employ conservative lawyers and GCs, they will definitely cover their ass. They will censor. They will roll back the chain if something bad enough happens. They'll even halt it if the government forces them to.
That's not a bad thing. But it's a TradFi thing. It means a permissioned blockchain is a lot closer to a database than it is a chain. A bad database, overloaded with cryptography and consensus it doesn't really need.
So why am I saying all of this? Because a really smart researcher who I respect responded to my critiques of Tempo yesterday by saying "Coinbase is a major Ethereum validator. So the same liability problem exists."
But he/she is missing a key detail: Coinbase doesn't control Ethereum transactions, the protocol does. And because the protocol is robust, and the validator set is both diverse, global, and in some cases anonymous, Coinbase can plausibly claim being a neutral participant.
Coinbase can't do that if it get's permissioned by Stripe or whoever to be a part of Tempo. In a legal preceding or government action, it could credibly be argued that "The CEO of Coinbase could have gotten
@matthuang on the phone and convinced him to get the other validators to pair back the chain, at the risk of being kicked out."
That can't be argued for
@VitalikButerin on Ethereum, or anyone on Bitcoin.
And of course the folks involved with Tempo/Arc/Canton etc etc know all of this, which is why they'll be buried with legal and contractual negotiations for the foreseeable future.
All of that wrangling is why every other attempt at building a permissioned chain, despite the honest attempts of really smart people and the investment of countless millions, has ended in total disaster.
Lastly, the claim that Tempo/ARC/Canton are "permissionless but public" is a fantasy. All permissionless chains will become private eventually. It's the natural order of things when a corporation, and not the protocol, is in charge.