Comment by thisisit | Hacker News Reader

Comment by thisisit | Hacker News Reader

It's basically this kind of confused, bullshit reporting which gives false hope to the FOMO crowd or draws ire from bitcoin naysayers.

One, the confusing narrative around blockchain. Settlements don't require any kind of "proof of work" mechanism. It's basically a shared ledger where only permissioned parties can access/write. Some are calling it "Distributed Ledger".

Surely, someone is going to point out that "proof of work" is not the definition/requirement for blockchain. But, using it interchangeably and confusing the bank "blockchain" with the cryptocurrencies "blockchain" is causing the confusion and euphoria.

Second, settlements don't even need "coins". Cryptocurrencies use "coins" as an incentive mechanism. In PoW, you earn coins for securing the network. In PoS, you stake your coins to secure the network and earn fees.

But, if you have a permissioned, private blockchain there is no need of "coins" because there is no incentive structure. Will DBS earn extra coins for each transaction in the blockchain because of staking or something? If not, then these "coins" are just numbers to enumerate position sizes. In which case, why even call them "coins" and not "settlement amount in x currency"?

> some are calling it "Distributed Ledger"

Friend runs settlement infrastructure for a large bank. It's not sexy work. He has to fight for his budget. His TL; DR with a lot of this is it has made getting necessary infrastructure upgrade approvals easier.

Been needing to unify wire confirmations with some esoteric trade settlement system? Talk about Ripple. Adding real-time functionality to a net settlement system? It's a "distributed ledger". Boss gets a press release and maybe a glowing write-up, IT gets its kit.

I completely understand. And this is a problem with other hyped up technology too.

One of my previous companies, we had people working on 'far reaching ML and AI stuff'. What they were actually doing was calculating euclidean distance between users to build a recommendation engine. Data size was around 100-150 items but every user had access to max 20 items. Questions about - Do we really need a complete recommendation system to select from 20 items? Fell on deaf years.

Great press for the manager and IT gets to show off something new.

The problem with this kind of cryptocurrency hype is that Average Joe is putting money into bitcoin because "I read DBS is using blockchain so bitcoin will go up".

This. McKinsey/BCG/others makes presentations to banks executive boards saying Distributed Ledgers are the future and you need to get onboard if you want to survive. Now every IT department head will try to do nonsense distributer ledger stuff for his promotion/bonus and business department heads will be OK because they don't want to be blamed for missed opportunities.

> because they don't want to be blamed for missed opportunities

It sounds more rational than that. Blockchain announcements make headlines. And blockchain headlines are driving stock prices [1]. There are numpties in the mix, but I don't think it's them.

[1] https://www.bloomberg.com/news/articles/2017-12-21/crypto-cr...

Another phrase, perhaps along the lines of "probability of bet" would be a much better term than "proof of work" given that the reason the energetic commitment of work matters at all is that it represents a finite unit of value that someone, somewhere, allocated to a given blockchain in the form of risk. The amount of underlying value at risk is actually unknown because nobody knows what the cost basis was of the electricity used to "purchase" it. But using proof of work to describe the mechanism of committing a bet to a blockchain is rather like calling the ante in a card game "proof of chip toss". It's technically correct but it describes a relatively incidental aspect of the incentive construction process.

Tim Swanson: Enterprise Blockchain is in a "Trough of Disillusionment”

https://bitcoinmagazine.com/articles/tim-swanson-enterprise-...

Would a GIT repo work better than a blockchain for a large bank?

For what exactly? Banks used append only transaction tables with hashes to prevent row tampering/reorder ages ago.

Reply on news.ycombinator.com