Stop promoting other coins please. Don't turn these comment sections into an alternate version of "crypto Twitter".

I'm trying to understand this part about "as long as less than 1/3 of the nodes are evil"part. That seems like a very big thing to gloss over. Bitcoin is secure precisely because the cost to "do evil" is more than the value you would get by "doing good" (i.e. mining). That the brilliance of it. If I understood him correctly, you could attack this system for a very small cost just by creating cheap nodes. What am I missing?

Only one problem.. It't not secure in a public environment.

Chris: start here https://www.youtube.com/watch?v=wgwYU1Zr9Tg plus check out some of the associated videos, it's worthwhile, once the penny drops, the true brilliance and potential of this technology will be revealed... MR Petter - this technology is THE MOST secure of DLT, so on a relative basis it IS secure. On an absolute basis, it is VERY secure, however not 100% secure (but pretty darn close).

What might be left out of his "as long as less than 1/3 of the nodes are evil" is the additional phrase "and coordinated." 1/3 would have to be not only evil but lying the same lie.

Apple = form over function. Looking forward to the coming demise of it and Silicon Valley at the hands of the Chinese in Shenzhen, just like how the Japanese obliterated US color TV & electronics half a century ago to the shock of the complacent naysayers.

at low scale, every DLT is vulnerable to bad actors. the way to ensure that the public network is as secure as possible, it should start as a permissioned network which then converts to public... I don't see a problem with that.

Great. If I create a closed DLT and only give myself permission to change the network it will be totally secure against Sybil Attacks. All the users have to do is TRUST that I am a benevolent dictator and will use my control of the network only for good. Is that a potential problem? Does that change if there are multiple people with permission? Who chooses the entities that get permission first? What if the permissioned entities decide to change the rules after letting you in the network? etc. etc. What is the difference between a permissioned (controlled) DLT and the current financial system? Does giving certain network users but not others control seem like a problem?

Or let's look at it a different way... Isn't the point of a DLT network that all we have to do is trust the code and not trust the users??? All we have to do to bootstrap a public hashgraph network is to give permission (trust) to a certain set of users in exchange for security. Is that contradictory or am I missing something?

You are right... the objective of DLT is to eliminate the risk of trust. I.e. trust is achieved via the consensus algorithm. Blockchain seeks to achieve this one way, hash graph seeks to achieve it differently. The main issue seems to be how participants obtain voting rights. Per the discussion paper you refer to, there are a number of methods currently deployed in the market, each of which have their own risks / benefits. It all boils down to incentives. Obtaining voting rights should incentivise good behaviour. Inversely - behaving badly will cost you more than you will probabilistically benefit. MR

Nope, you are spot on. People massively underestimate the importance of security and censorship resistance when it comes to cryptocurrencies. Creating a fast and low-cost cryptocurrency is not difficult. Creating a secure and censorship resistant one is. If money is on the line, and there is a trust issue in the system, be sure that the trust issue WILL be exploited at some point. And since hashgraph only works in a permissioned enviroment you would not be able to trust it in an open setting. This means it won't compete with Bitcoin.

Chris, I think you picked up on some of the words, but didn't totally understand what Leemon was trying to say. Hashgraph is the-most secure distributed ledger technology. Relatively more secure therefore, than others on the market which use proof or work or proof of stake. If a certain proportion of bad actors collude (lets say 1/3) then it is 'possible' to mount some sort of attack. But it is the type of attack which also matters. DDOS - not possible. Firewall - not possible. So whilst attack is possible, the type of security risk is very important and the probability of success also matters. Petter - your interpretation of hash graph only working on a permissioned network is not accurate I don't think. There are plenty of online resources which explain this technology including the white paper and some online lectures which Leemon (and others) present the technology in different ways and conduct Q&A sessions: https://hashgraph.com - start here... MR

Marten - I just read through this document: http://www.swirlds.com/downloads/Swirlds-and-Sybil-Attacks.pdf In the document it is stated that you can choose among several different consensus algorithms in Swirlds: - Proof of Stake - Proof of Burn - Proof of Work - Permissioned - Hybrid - Trivial The first 3 options are no different than what current cryptocurrencies use. So nothing new there. Out of the 3 remaining options only "Permissioned" and "Hybrid" are resistant to sybil attacks (the risk of someone trying to trick the system into thinking he/she is more than one person/node in order to get more votes in the system). In a permissioned system only a pre-determined set of participants can join. In a hybrid system a group of "founders" can decide to invite more people to the system. These new members can in turn decide to invite even more participats (to whom the invitee shares his/her votes). So even though the hybrid system is somewhat open, it is not permissionless. You need to get an invite from someone in order to join. I obviously can't say that Hashgraph will never succeed in some way, but it is definately not built on the same open and permissionless innovation principles that are the foundation of Bitcoin. The fact that Hashgraph is closed source software built by a private company (afaik) is a huge red flag to me.

Petter, Thank you for the follow up. This is an interesting discussion. The distinction between hashgraph and other DLT is crucial. For the whole audience here is the link to the paper to which Petter refers: http://www.swirlds.com/downloads/Swirlds-and-Sybil-Attacks.pdf I think that there is confusion regarding the consensus algorithm and the method by which a member of the network gets a voting stake. The consensus algorithm, is hashgraph, within which we have the gossip about gossip and deemed voting breakthroughs/ innovations. On top of this 'platform', app developers can build stuff. It is they, who determine how a 'member' obtains voting rights. Quote "Of course, an app developer could choose to do something more complicated. The swirld could start with the hybrid model, then allow users to sell voting stake to each other. Or it might be permissioned, but with the permissioning done within the swirld itself, by having an actual election, with the humans talking with each other prior to voting. Or it might start with proof-of- burn, and automatically transition to proof-of-stake once the total value gets large enough. In all cases, the consensus is always decided by the platform, using the current record of voting stake, as managed by the app." So, for convenience, within the paper you rightly point out that Leemon refers to other DLT and voting rights methodologies, however it is not an exclusive list. There is more discussion in 'Section 2: An Example Scenario', which I would encourage everyone to read (the whole paper is only 2 pages). It's important to think in relative terms here. Hash graph is not perfect, but in my view it's vastly superior to blockchain. Time.... will tell. MR

Tuur is very bitcoin biased.

good point

You can have sound money and credit money (paper money). Remember the last gold standard? What you shall not have is shity money and credit money. This leads to Venezuela-type hyper-inflation cases. You may have thrilled in a shit-money society, but it does not mean it is right or fair for less favoured people. #$ #Largest scam ever.

Paul N. I think Leemon Baird addressed the voting topic - it´s virtual and depends on info already there. As to closed/open source didn´t Leemon said at the begining it´s up to the person as to the "world" be open or closed? And for the searching and reading that I´ve done (not being in the biz), this guy explains the concept it very well. Much better than hundreds of proponents in the fintech industry. And he has a lot of speeches, Q&A online so every one can judge for himself and can dispute it. Finally, as with all technologies, it may not pass the test of time. But I don´t think Leemon Baird can be guilty of cheating, colluding or showing any hubris.

Agree with your comment entirely. Blockchain is only good at one thing: censorship resistance. Proof of work is necessary to externalize a cost (electricity) on miners too keep them from attacking the network, and in exchange they are rewarded with something intrinsic to the network (new coins). If hashgraph nodes require tiny amounts of energy botnets (read Windows machines) can easily be used to sybil attack the hashgraph network. Hashgraph works great in interbank transaction networks where everyone is friends, but not so well in adversarial environments. In his own assumptions, which aren't mentioned in the video, an attacker only needs to control >1/3 of the hasgraph network nodes. I'm not sure what affect that would have on the hashgraph but I'm guessing it's not just a partial transaction history reorganization. Bitcoin requires 51% of hash power to attempt a partial 1 block reorg, or double spend. Which network is more expensive to attack? Bitcoin definitely has issues, but these two networks aim to solve different problems. He also seems to be a bit confused about what a bitcoin full node is. Here is a link that shows you how to make your own raspberry pi full bitcoin node. http://www.raspberrypifullnode.com/

Their response to the Sybil attack problem is addressed here: http://www.swirlds.com/downloads/Swirlds-and-Sybil-Attacks.pdf

Their response is also nonsensical. Okay lets just use all the known consensus proposals and you can pick one because we dont know which one works... Proof of stake and proof of burn are completely untested, the other are straight up permissioned, and a network cannot be secure if only a small % are doing proof of work. Im more convinced than before that this project is a straight up scam.

Thanks Ricardo, you would think more people would be asking this question, why else do we watch RV?

Its nice to know whats going on .. but its even nicer to make money.

An interactive interview for the #realaudio fans: https://www.hiddenforcespod.com/leemon-baird-hashgraph-distributed-ledger-technology-blockchain/ Does anyone else find Leemon reminiscent of @elonmusk? Still, it's a great concept.

I think I agree with Amir. This felt like a radio DJ condescendingly explaining bitcoin and blockchain and then trying to sell me hashgraph. Like Amir I ask myself - was this useful / can I use it? - and in this case I don't see a well spent half hour. Just my feedback but I'm not loving this and the recent AI talk. I come here for some link to finance / investment and these have veered off the path for me.

To Eric F - if it was 2011 and somebody presented a talk on blockchain technology and how bitcoin is a new type of money, would you be complaining that it has nothing to do with investment and finance?

Henry, I am a trader on a commodity cash market. I am confident I can replace myself with a computer using a hashgraph (back office) and RPA (front office). I need to find experts in each field. If you or someone you know might want to talk, please call 978 500 9301 or email Andrew.gibson@fctg.com

The 1/3 figure comes from theoretical mathematics. Look up the Byzantine Generals Problem if you want to understand why that result is so important. What he's describing here is a truly brilliant solution to a problem that has confounded theoretical computer scientists for decades: a byzantine fault tolerant consensus algorithm that operates over asynchronous networks (i.e., the internet) and doesn't involve an explosive number of messages to be sent relative to the number of transactions being performed. Regarding your question about fake news, dealing with any such type of traitorous behavior is precisely the point of a Byzantine tolerant algorithm. As long as the contagion afflicts less than 1/3 of the nodes, the problem will be detected and the correct decision will be made by the majority.

Danger..... Patented equals ultimate centralized control equals same danger all centralized systems represent. Might be a credible alternative to blockchain if made public opensource but don't think this guy will get there..... at least not soon.