>I'm also currently doing a doctoral thesis focussed on smart contract security, so I do know a bit about this.
That to me is another really awful thing about blockchains. They're like candy for academics. It's very easy to get sucked into the illusion as well because if you take it at face value there are a lot of things to study and write about, the algorithms seem complicated and interesting. And they are, if you ignore the rest of the effects. I'm sure you could write 100 doctoral theses on smart contracts. I'm sure companies would pay you to do a lot of things with them. They are still an extremely bad and broken concept.
Edit: I'm at the point where academics have sent me so many papers on blockchains. I would go as far as to say the entire point of the "defi industry" is not to produce useful products, but to produce research papers. But I don't even know if I could say that because a lot of the papers aren't even good! It's so common for a blockchain paper to launch into a lot of math and algorithms without even trying to make a problem statement or explain what purpose it actually has. Please don't let your thesis be like this. If all you're writing about is how existing users of smart contracts are avoiding getting hacked then you've been dragged into the same trap.
>The whole premise of blockchain is execution of intent, in an untrusted environment.
The issue here is that this is a completely nonsense premise. There is no such thing as finance in an "untrusted environment". If you aren't planning to live alone in a bunker, then somewhere along the line you need to have trusted intermediaries. Decentralization does not and cannot ever fulfill this premise, it actually increases the amount of trusted intermediaries you need to deal with.
>The threat you are concerned about is collusion, and the best control against this is further decentralization and scale.
No, this is also incredibly wrong. When it comes to market manipulation, further decentralization does nothing to reduce collusion. It is already hard enough to track the source of bad behavior in the market without people intentionally designing algorithms to do that, we don't need to make it even harder.
>Proof-of-Stake (and many DAO's), do provide more assurance against malicious actors by validators risking their own stake if they try malicious activity
This conversation again is veering into territory where it has nothing to do with blockchains or smart contracts at all. You can rephrase this as "people are less likely to engage in bad behavior when they are invested and hold stake in something".
>Ultimately, I am confident that the "dark forest" will give us much more security and transparency than the current closed/centralised system.
As someone who has been doing open source for the last 20 years, I am not. Much like the Linux desktop mentioned in a parent comment, pretty much every blockchain is just a random open source project that is accountable to nobody. People remain interested as long as there is profit and hype in it, when that dries up they will disappear and any illusion of security and transparency will fade away. This doesn't change because it was applied to finance instead of OS kernels or whatever.
>What would it take to change your view? If the answer is nothing, can you really claim to be thinking about this critically?
Actually my view is the result of thinking about this critically for the last 10 years. I was like you at one point and I would have made the exact same comments that you're making now. But sometimes a spade is just a spade and we gain nothing from pretending it's not. If you want to change my view, it would help to think about how to bring smart contracts into traditional banking systems. Look at all the reasons why it won't work. Then go forwards from there instead of working backwards from blockchains and "decentralization" because those are held as some kind of panacea.