Thread by @ivangbi_

Wake up honey! Governance is broken, again 😩

Every once in a while people like to sh*t on governance and remove it all fully. It is absolutely correct to do so if done based on proper arguments, and not just "it's all bad kill it forget it". Reasoning matters! A bit of a 🧵

Governance can be blamed due to:

1] Hard issue: objective intrusive flexibility (=control) over the protocol parameters and/or user assets
2] Soft issue: ineffective, lack of participation, vote rigging, & other subjective aspects

Scroll down to #7 for less obvious stuff.

1/14

Start with 1⃣

Governance has the following advantage: it can change things. Parameters, treasury spending, interest rates, LTVs, whatever else it is - depending on what admin powers it has. Changing things is super powerful, because you can make things better. You can FIX!

2/14

You can improve a product, add features... And that's BAD! Wait what, hold up fker, you just said it's good? - Well, for the OG ethos, is bad. Those hold $LUSD and love all things oracless, admin-less, and love Bitcoin bricks. I am partly in that camp too, as a user!

3/14

But it's not practical unless your product found great PMF. Because before you'd hit PMF, you essentially release a bricked version that you can't bootstrap or tweak in order to keep up with the market. You have no tools at your disposal. It's a cage you make for yourself.

4/14

So before that happens, you need to live with the fact that governance can change interest rates, remove or add assets, introduce new risks into a lending pool, and even make good features that can (due to faulty new code) - break an old, good product. That happened before!

5/14

You have to decide. Most go with the "let us have powers at the start" - fully agree here. Just put efforts into making such an environment safer nonetheless:

- high quorum
- timelock
- non-upgradeable contracts
- no actual user funds access
- audit every change...

6/14

If you don't have gov, you will feel sad when you can't add a simple feature while every competitor is adding it. Because the % of people who will value full immutability is low, and stats prove that. Same reason why barely any end user pays for privacy products. Reality.

7/14

NOTE 🤔 I am not sure the above thread stands for @Uniswap though. Somehow with fully immutable contracts (or rather, with super low admin powers, similar to @CurveFinance emergency multisig also only being able to halt & stop a gauge) - it's still able to add & improve.

NOTE [cont.] Could that be just because they are such an obvious >70%+ market dominant? And if you are 5% then migrations to new versions become much harder? Maybe. Any other outliers in this comparison at all?

Now onto the less obvious stuff, and what Michael actually referred to being 2⃣ the soft social aspects

Here, can't agree at all. There is no problem here that needs fixing to begin with. Voting doesn't mean you need 10000 individual voters. Is that a benefit? Not sure 🤔

8/14

Comparing to real world, is every director qualified? is every voter aware of all the data & context? Hell no, far from that. So why should a protocol have only fully sophisticated governors? You need a few (5? 10? 15?) qualified opinions, better a bit diverse, and done.

9/14

The point of DAOs, partly (and yes they are inefficient but good) is that many things are open and transparent. They can be assessed, judged, and discussed openly. Discourse is enough, that already means things are working.

10/14

If voter opinions are literally being ignored every single time, with a whale voting against f.e. - the value of such a system would drop, and/or capital will move away too. This freely accessible information flow already has its own causality, and it works O-K.

11/14

"But that is still imperfect"! Well, from the builder perspective, bricked no-governance systems are also imperfect. It's all pros&cons throughout the protocol designs. Lending P2P solves it? No, @AaveAave proved it otherwise. P2Pool? Not sure. Isolated? Too fragmented...

12/14

I will leave that thread idea to @apeir99n & @0xmikko_eth as they are more qualified to discuss that. "Security" aspect of bricked systems isn't even a full guarantee either, as we see that even after 1+ year an old bug can resurface. So it's only all marginal benefits 🫣

13/14

We can see people shifting goalposts now to admin-less and oracle-less but it's such a hard battle. You need to be damn sure your design is super solid and is THE best approach, or your building efforts are gonna be almost wasted. Cool, fun, but... is it practical?

14/14

PS: no beef, Michael is bigbrain, the above thread isn't a dunk at all (how could i). There are just different paths and approaches when it comes to such products, so everyone talks their own book (me included) based on what design choices they have made in their bag 🫡

Curious what @NathanVDH0x @euler_mab @lemiscate @StaniKulechov @gham1lt0n @spencernoon @pet3rpan_ @TokenBrice @delitzer (altho about Dan is clear haha) and other bigbrain sers think on this 🫂

@NathanVDH0x @euler_mab @lemiscate @StaniKulechov @gham1lt0n @spencernoon @pet3rpan_ @TokenBrice @delitzer On design of lending protocols 🧵