I saw this message a few days ago on group chat:

Have spent quite a bit of time thinking about this and would be happy to chat over DMs (especially if you are planning to write/speak about this topic). But some initial thoughts:

There are a lot of pros to issuing on Liquid, but the main con as far as I'm concerned is segregated network effects; even though there are some "impractical" (i.e. requiring trust and/or patience) alternatives, Liquid is obviously optimized to be used with L-BTC in place of mainchain/LN BTC.

Meanwhile, TapAss is clearly optimized for a user journey that's centred around an otherwise standard BTC/LN wallet "with extras" in the form of tokenized assets and the ability to swap between everything that can be stored in said wallet. The ease with which any Bitcoiner can get started with TapAss without making impractical compromises is an important pro that puts it miles ahead of Liquid in terms of potential for adoption and network effects...

But there's an important con also, which is that just like any approach to tokenizing directly inside Bitcoin instead of a sidechain, Taproot Assets commits data to the Bitcoin blockchain, competing with monetary (BTC) settlement txs for scarce and increasingly precious blockspace.

Every asset issuance, Lightning channel creation or rebalancing transaction for tokenized assets, and every transfer of an asset that has too little liquidity to practically be used over Lightning, will take up some blockspace, regardless of what you use. At the same pace as more and more financial instruments become tokenized and traded against BTC (an inevitable part of hyperbitcoinization), it's likely that Bitcoin's UTXO set will also be used less and less for anything other than settling large transfers of BTC. That's a problem which only a sidechain can solve - and also the reason why I'm working on Sequentia, a fork of Elements that aims to combine the best of both tradeoffs by providing a TapAss-like UX from a Liquid-like architecture.