5 ms·

> It is great to see effort being put into researching cleaner alternatives. Unfortunately, there is no working PoS-based system that does not rely on some degr

by baby 6y ago

> It is great to see effort being put into researching cleaner alternatives. Unfortunately, there is no working PoS-based system that does not rely on some degree of centralization in its security model. This is false. See Algorand, Cardano, Mina, etc. Proof-of-work is very much a technology of the past. It's the fossil fuel of cryptocurrencies.

tylercubell 6y ago
Additionally, Algorand has pledged [1] to be carbon-negative. [1]: https://www.algorand.com/resources/news/carbon_negative_announcement https://www.algorand.com/resources/news/carbon_negative_anno...
encryptluks2 6y ago
Algorand isn't currently decentralized. They have a list of approx 100 relays required for the network to function that they control tied to a centralized single domain.
- baby 6y ago
  Do you have a good source on that? This is not the first time I hear this. Btw I’ve studied the consensus protocol and it fulfills promises of decentralization so I’m wondering why they are using some single points of failure pieces (it is decentralized from a safety perspective: you can’t double spend, but it’s possible that some pieces are centralized). My guess is efficiency.
  - encryptluks2 5y ago
    https://algorand.foundation/faq https://algorand.foundation/faq Who manages the list of relay nodes? What about decentralization? > Currently, the Algorand Foundation manages the official list of relay nodes... > We are working on a model where the decisions on relay nodes will be done in a more decentralized way. As for the centralized domain, you can look it up in the source code but essentially it gets the list of relays from a SRV record on the algorand.network domain: _algobootstrap._tcp.mainnet.algorand.network
    - baby 5y ago
      Ok I see. In practice this reminds me of the issue of downloading the client and of seed peers that every cryptocurrency has: - everyone downloads the client/code of a cryptocurrency from the same place - that client also hardcodes a list of peers that you first connect to in order to discover more peers The first issue is more problematic, and the second issue usually matters less. My guess is that algorand rely on relay nodes because they are targeting a high throughput, and thus need quick communication between nodes.

by baby 6y ago

tylercubell 6y ago

Additionally, Algorand has pledged [1] to be carbon-negative. [1]: https://www.algorand.com/resources/news/carbon_negative_announcement https://www.algorand.com/resources/news/carbon_negative_anno...

encryptluks2 6y ago

Algorand isn't currently decentralized. They have a list of approx 100 relays required for the network to function that they control tied to a centralized single domain.

baby 6y ago
Do you have a good source on that? This is not the first time I hear this. Btw I’ve studied the consensus protocol and it fulfills promises of decentralization so I’m wondering why they are using some single points of failure pieces (it is decentralized from a safety perspective: you can’t double spend, but it’s possible that some pieces are centralized). My guess is efficiency.
- encryptluks2 5y ago
  https://algorand.foundation/faq https://algorand.foundation/faq Who manages the list of relay nodes? What about decentralization? > Currently, the Algorand Foundation manages the official list of relay nodes... > We are working on a model where the decisions on relay nodes will be done in a more decentralized way. As for the centralized domain, you can look it up in the source code but essentially it gets the list of relays from a SRV record on the algorand.network domain: _algobootstrap._tcp.mainnet.algorand.network
  - baby 5y ago
    Ok I see. In practice this reminds me of the issue of downloading the client and of seed peers that every cryptocurrency has: - everyone downloads the client/code of a cryptocurrency from the same place - that client also hardcodes a list of peers that you first connect to in order to discover more peers The first issue is more problematic, and the second issue usually matters less. My guess is that algorand rely on relay nodes because they are targeting a high throughput, and thus need quick communication between nodes.