Aion is yet another blockchain project that claims it will turn the world upside down. And save the Internet!
(It’s unrelated to the popular MMO game of the same name.)
Aion has just done a marketing push making out that it’s the “first” blockchain for the Java programming language — in some sense that’s not entirely clear from the coverage, e.g., the Coindesk headline “Aion Network Launches First Blockchain Virtual Machine on Java”.
It’s nothing of the sort, of course. But then, what is it?
Aion and Java: first to be second or third
Java is a programming language that’s popular in enterprise environments — it’s a bit bureaucratic, but it makes it somewhat easier to code things in a correct and maintainable manner. (My day job is as a sysadmin to Java programmers.)
Aion’s smart contract engine, the Aion Virtual Machine, has been rewritten from C++ to Java. Aion posted about why they wanted a Java-based virtual machine in November 2018, and they put the new Java-based AVM live a few days ago.
Aion’s precise claim (archive), from their blog post announcing the move, is:
the world’s first JVM-compliant Blockchain virtual machine.
It’s possible you could twist shades of meaning of the words in that phrase to be true — but in any reasonable and obvious sense, it’s just false.
Implementing a simple blockchain in Java is a standard developer introduction to blockchains. So that bit’s not new. Nor is implementing your smart contract virtual machine in Java.
But nobody cares about your platform’s back-end — they care what they can do with it. Aion’s big promise here is that you can finally write blockchain smart contracts in Java!
This is not new in any way — R3 Corda and Hyperledger both do smart contracts in Java just fine. They’re enterprise — of course they’re going to do Java. R3’s implementation is built in Kotlin, a close cousin of Java that also runs on the JVM.
Matthew Spoke, founder of Aion, tells TechCrunch (headline: “Aion Network introduces first blockchain virtual machine for Java developers”):
Blockchains have a set of unique criteria. They need to be deterministic; the computing needs to happen across the distributed network of nodes; and the JVM was never designed with this in mind.
Aion is publicising their Java thing as if nobody ever did this before — let alone the two biggest enterprise blockchains.
As far as I can work out, Aion’s magic to make Java into a smart contract language is:
- whitelist classes that you’re allowing into the sandbox;
- mark a subset of the JVM platform classes as available;
- shadow those classes into a new set of package names;
- rewrite Java bytecode as needed.
I asked a Java expert about Aion’s claims — Ben Evans, who is principal engineer and JVM architect at New Relic, co-wrote Optimizing Java (UK, US) and Java in a Nutshell (UK, US) for O’Reilly, and is an editor for Java-heavy programmer’s magazine InfoQ.
Evans says: “For a decent enough low-level JVM engineer, it’s a clean and somewhat obvious architectural approach.”
In fact, it’s sufficiently obvious that Evans pointed me to a video in which he talks about how and why you would run smart contracts on the JVM — and sets out, step by step, how to make a Java program provably deterministic, as Spoke tries to take credit for — in 2017, when he was working on this exact problem for R3 Corda. (Here’s the InfoQ article.) This covers a lot of the specific technical ideas that Aion are proudly claiming credit for in 2019. It’s a good and very clear video, well worth an hour of your time.
The baffling bit is that Aion thinks anyone in the Java space would see any of this as novel. It wasn’t new when Evans and R3 did it in Corda — and they didn’t claim it as new — and it’s not new now Aion’s doing it.
I contacted Aion and asked just what they were claiming — given the long history of the ideas. Calvin Sribniak-Jones, head of marketing, got back to me:
Although there are other platforms that enable Java code to interact with the Blockchain this is the first time that native-JVM code can run directly on a public Blockchain safely and reliably.
So, the first public blockchain to do this trick — but that’s not in the publicity, and they’re still not acknowledging the previous work.
Evans says that Aion’s code quality is okay — which immediately puts it ahead of 95% of the blockchain space. So that’s nice.
The origins of Aion
Aion’s founder Matthew Spoke started as a chartered accountant at Deloitte in Toronto, Canada. He got Deloitte into blockchains in 2014 — he sold them on the idea that Bitcoin could “render auditors irrelevant” — and led the team that produced Deloitte’s Rubix blockchain software.
Spoke took two others from the Rubix team, and founded a company called Nuco in 2016 — which became the Aion team.
Nuco’s original ambitions were focused on enterprise blockchain:
Nuco’s business model is to build, license and deploy customized private blockchain networks for clients across industries. Specifically, Nuco will begin building tools for tracking compliance duties for firms accountable to the Securities and Exchange Commission (SEC) and federal environmental regulators.
Spoke worked with Vitalik Buterin on an enterprise-suitable private version of Ethereum, and is still on the board of the Enterprise Ethereum Alliance. Nuco also got Buterin and Alex Tapscott on board as advisors.
The Aion ICO
Aion ran an ICO in October 2017. There was a presale to accredited and institutional investors, such as Michael Novogratz’s Galaxy Digital. This raised $23 million. Here’s the technical white paper (archive) and the executive summary.
Aion cancelled the planned public token sale … possibly because the AION token was already trading below the $1 asking price.
Here’s the promotional video. It speaks of “first-generation” blockchains like Bitcoin, that just run a token; “second-generation” ones like Ethereum, that you can run programs on; and “third-generation” ones like Aion, which will permit “disparate blockchain networks to communicate with each other, for the very first time” — providing “federated inter-chain communication of both data and value,” i.e., money.
What Aion actually offered was an Ethereum ERC-20 token, AION — though those are now being swapped to native Aion-based coins.
You can now write dapps (smart contract programs) on the Aion blockchain.
We’ve yet to see evidence of inter-chain connectivity. The video promises Aion tokens moving “seamlessly” between the Aion and Ethereum blockchains — though in practice, they’re moving the AION tokens by hand.
The video also promises that Aion will be just the thing for supply chains and health care, because of course it will.
The AION price chart is a completely standard altcoin decline. Their Coindesk article from December 2018 trumpets that the token is listed on Binance and Bitfinex — but the volume is overwhelmingly on bottom-of-the-barrel altcoin exchanges.
Aion in practice
At present, Aion is a stand-alone blockchain network, with a token, that runs smart contracts.
Aion doesn’t seem to be marketing itself to businesses, but to people who want to do stuff on a public blockchain — an Ethereum-but-better blockchain.
Aion uses Proof of Work — literally wasting electricity to give coins away in a lottery. Specifically, it uses Equihash, most notably used by Zcash.
Zcash does 4.2 billion hashes per second, Aion does 800,000 — you could trivially 51% attack this thing with spare hash power you found behind the sofa cushions. A single Antminer Z11 does 135,000 hashes per second.
Aion plans to move to a new consensus algorithm, Unity Hybrid Consensus — a combination of Proof of Work and Proof of Stake.
The Unity white paper notes the natural centralising tendencies of both PoW and PoS — and proposes that requiring both hashing power and a stake will make the system not centralise. Yao Sun from Aion blogged about what they were thinking here.
The obvious question is — what if someone has both? Say, if they’re mining a thinly-traded token they can’t cash out easily. This already happened with Bitmain and their “inventory” of a million Bitcoin Cash — and BCH is fabulously liquid compared to Aion.
Fundamentally, fiddling with the algorithm doesn’t solve the problem of lack of participation. Every mined crypto, even Bitcoin, started as a mining monopoly. Aion mining is 55% a single pool.
Looking to the future is fine — but Unity is a hypothetical solution that can’t be tested properly, to a problem Aion is nowhere near having.
The Aion Foundation
The Aion Foundation is incorporated as a Cayman Islands non-profit foundation company. Nuco is now a subsidiary of the Foundation.
The Foundation publishes regular reports on its reserves and spending. The Aion Foundation Report of 30 November 2018 was prepared with “strategic guidance” from Deloitte, who they remain on good terms with, but is not an “audit” as such.
As of 31 October 2018, the Foundation had $14 million in dollars, bitcoins and ether. Spoke told CoinDesk that Aion had 18 months’ funding runway — though by 31 March 2019, they had $8.3 million left. About half the spending is developers’ wages.
The Aion Foundation predicts at least five years to their goals being practical — “services comparable to centralized alternatives like Amazon Web Services,” Spoke told CoinDesk.
Aion doesn’t have any other income, or claim prospects of any other income — “The Foundation intends to earmark a portion of its AION holdings for long-term funding of operating costs.” Spoke told CoinDesk that they “might seek additional funding from crypto funds and accredited investors in return for Aion tokens.”
“In service of the cause”
Aion looks like someone found an idea that was fascinating technically — and they just haven’t really answered “why would you even do that?”
This post from Mike Mason, VP of Product at Aion, is one Aion answer to “why” — and it reads like extruded blockchain hype gibberish. But I expect they’re completely sincere.
His thesis is that:
- the Internet is now run by a few giant companies — yep, and that’s a problem
- that this is a propaganda outlet for nation-states — maybe … but you’re going to try to fix this with blockchains?
- that a legislative response cannot work — what?
- and therefore the technology of “Web3” will work around these social problems — and evade government oversight on a massive scale … using blockchains.
This is a rerun of the Bitcoin conspiracy theorist dream I outlined in chapter 2 of Attack of the 50 Foot Blockchain — where techies who think that understanding technology transfers to understanding society or politics, try to fix social and political problems with technical hacks.
There’s one thing Mason doesn’t argue at any point — that Aion has a path from point A to point B, let alone any plausible prospect of getting us there.
The elephant in the room is that Aion is a me-too coin, with nearly no users, that started eight years after Bitcoin — when Bitcoin had already failed in all these ambitions and become just a speculative commodity. There are any number of coins that claim this as their mission.
When you say this to blockchain dreamers, they come back with claims like “time will tell” — but it pretty much already has, and the burden of proof is now entirely on those claiming all of this is even possible, let alone feasible.
(Maybe Aion can make a real impact, and achieve the power waste of Bitcoin.)
Like a lot of Bitcoin and blockchain weirdness, Mason starts by asking some good questions — and then gives bizarre and disconnected answers, with handwaving for the hard bits. But being able to ask good questions doesn’t mean your answer will work, or even makes sense as an approach.
Sribniak-Jones from Aion clarified that this post is an unofficial personal opinion piece from Mason. Though Spoke signs off The Aion Foundation’s November 2018 report with “In service of the cause” — so I suspect Aion’s well into it.
Mason’s earlier post “A Request for Decentralized Infrastructure” is more concrete —though mostly it’s a list of problems with blockchains that run smart contracts.
The key issue is that all of the problems Mason lists are self-inflicted. They all stem from thinking that bad ideas — immutability, smart contracts — are good ideas, if only you can iteratively patch the problems as bits burst, explode or fall off.
We have ten years’ experience of the ways this stuff fails to work in practice — it’s not at all clear how Aion can do better, from first principles.
Aion appears to be a perpetual motion machine powered by unicorn feathers and phlogiston — but all the fittings are precisely machined and beautifully polished.
Your subscriptions keep this site going. Sign up today!