The trouble is that the reason PoW protects the network is that the mined block is an input to the PoW function. This means that if an attacker wants to change the network it needs to spend the same amount of processing power mining another block (and as the blocks are generated, it's gets harder and harder to catch up)

Can you think about a real world problem that also has the freshly mined block as an input?

Perhaps we can have a randomized problem like a monte carlo algorithm of protein folding where the random seed is set by the hash of the mined block. Or something.

I’ve thought about this on occasion. It seems that allowing arbitrary work units (eg user defined ones) breaks the POW model as it’d not use a predictable amount of compute. Miners or others could readily game the system.

Maybe it’d be possible to make the compute time unpredictable (or more predictable) using homomorphic encryption or similar techniques. Then a computation would operate over all the possible inputs. That’d be interesting, but would limit the types of problems.

Alternatively, perhaps something like QM simulations over pre-specified inputs (with a salting function based on the current block hash) might do POW and be useable. The inputs could be just defined over an increasing number of atoms in set pattern. Though finding problems sets in QM that are also NP complete might limit the usefulness. Many interesting QM calculations aren’t NP and don’t have any polynomial way to check answers.

If the work is significantly useful outside of the protocol then the protocol is insecure because the actors care less about the protocol than the externality.

> the actors care less about the protocol than the externality.

And why is this bad?

In other words: in classic PoW attacker needs to spend money on the attack, and if attack fails, they get nothing in return, they just wasted money. If PoW protocol is doing some useful work, then attacker can also benefit from that work, which means the cost of failed attack goes down to almost zero.

oh, indeed! at very least it weakens the PoW

Never say never, but I think it's probably impossible to solve specific business problems and still have some hypothetical PoW crypto be decentralized, trustless, secure, etc.

Is that necessarily a problem? To me, it isn't.

IMO, everything here comes down to a question of competing values.

> If you generally trust the government and our institutions to manage money, you'll probably view PoW crypto as unacceptably polluting almost no matter what.

> If you generally distrust the government and the ruin they can do to the currency, there's almost no amount of pollution you won't accept to save yourself from the misery of hyperinflation and the horrors that can come from that.

There's no reconciling these two worldviews.

The model that makes sense to me is what DYDX is doing with their decentralized platform.

They use the blockchain to settle and track finalized transactions and offload computations to the validator nodes (which are effectively servers dedicated towards supporting the security of the chain).

Regardless of how it is implemented, I don’t see the computations happening on chain. It makes the most sense to use the blockchain to buy and sell an open network of machines.

Akash Network are doing similar. OTOY RenderCoin are another. Flux are doing “proof of useful work” aka “the work does something valuable other than solving crypto puzzles.”

I think about this a lot because I was at NVIDIA : Intel : Arm looking at ecosystem and how web3 would impact silicon markets.

I have a podcast about this here: https://www.linkedin.com/feed/update/urn:li:share:6932815281...

And an article here: https://www.linkedin.com/pulse/machine-economy-coming-rex-st...

And a sub stack about it here: https://rexstjohn.substack.com/p/signal-versus-noise

And I track all these so-called “MachineFi” projects here: https://twitter.com/rexstjohn/status/1536565016320540673?s=2...

Short answer, no. Long answer, also no. Blockchain is inherently an inefficient way to compute. You should just trust a centralized authority and get waaaaay more utility for your compute. Your centralized authority could be the very cause or organization that you are computing for. If you trust them enough to run a random algorithm for them, you don't need a blockchain. You need to trust them as a central authority.

A public good project I work on (sigstore) penned an article on why a suitable application found it a challenge to leverage Blockchain https://blog.sigstore.dev/sigstore-blockchain-vs-transparenc...

My mining rig is an effective space heater.

This is the correct answer.

And also the main thermal infrastructure or entropic gradient of a https://en.wikipedia.org/wiki/Matrioshka_brain

So you mean useful Proof of Work? If so, then yes, and that is Handshake. [0] A decentralized blockchain based naming and certificate authority.

So far, that is the ONLY exceptional case and example of using proof-of-work that is useful and what I am willing to accept that makes sense to prevent a repeat of this [1] or this [2] or more TLD takeovers and DNS hijacking attempts like [3] and [4].

[0] https://handshake.org

[1] https://news.ycombinator.com/item?id=21611677

[2] https://thehackerblog.com/the-io-error-taking-control-of-all...

[3] https://cointelegraph.com/news/curve-finance-exploit-experts...

[4] https://cryptopotato.com/dns-hijack-compromised-ankrs-servic...