Honest Harmony ICO review

Published in

Golden Borodutch

15 min readMay 17, 2019

Note: this is an English translation of the Russian article we wrote recently with an in-depth analysis of the Harmony IEO.

The review was created by “Avocado Approves” community with the support of Golden Borodutch Telegram channel. For the convenience of readers, the review is divided into six parts: product, advisors, team, partners, legal part and conclusion. We’d appreciate if the project team answered the questions in bold publicly.

Worth noting that everything written below only represents the personal opinion of the author, which fully complies with the Fair Use laws and the First Amendment. It’s is in no way trading advice, and all the information was gathered from public sources.

Product

Harmony is a fast general-purpose blockchain based on sharding. Yes, there are plenty of similar projects. We’ve even reviewed a couple — MultiVAC and The Power. But none of them works under real conditions yet. Even Zilliqa and QuarkChain, which have “launched” their main networks, are now in a debug mode with all transactions disabled [1|2].

Sharding is a type of database partitioning that separates transactions into parts and distributes these parts between nodes for multi-threaded processing, which speeds up the overall verification and data transfer process.

Well, then what makes Harmony so appealing? We list a number of statements from their marketing materials:

The team that used to work at Google, Apple, Microsoft, and Amazon.
$18,400,000 that were raised on the seed round.
The negotiations with major exchanges for IEO opportunities. Harmony dismissed a usual ICO to avoid legal risks connected with SEC. Harmony IEO will be held at Binance Launchpad.
The ten billion people that said to be able to use the product when it’s ready
More than 20 major partners, which are ready to integrate the Harmony blockchain
The test network that shows the speed of about 114 000 transactions per second.

In the following sections of the review, we will check whether those are true or mere marketing tricks. Now, let’s take a closer look at the product itself.

The Harmony blockchain consists of many existing and experimental network technologies: OmniLedger, RapidChain, Chainspace, Kadmelia, QUIC, Bloom Tables, 5G, Unikernels, and Mosaic. The old White Paper had only but a brief description of those technologies, no technical details provided. The new White Paper says nothing about them at all.

People call Harmony a Frankenstein monster for it’s being an insane mixture of laboratory technologies that have never worked in practice. The Harmony team, though, promises to figure out the technologies, connect them, and put all of it in order. They justify the promise by drawing a story of how iPhone was made; that if you look under the hood of it, you’d also see a monster build together from different technologies.

As if that wasn’t enough, the project develops four more technologies:

OmniLedger based Deep sharding, that not only separates transactions, but also divides all levels of consensus, connections, and network states. The separation is followed by resharding meant to protect the network against attacks. Theoretically, this thing is used in MultiVAC.
Zero-knowledge proof, that is an application protocol that allows users to share their data without revealing sensitive information. It is said to a necessary technology for the data monetization, but the team keeps their secret of how the monetization will be realized. Zero-knowledge proof is also used in Zcash.
Min Language, that is a secure programming language for smart contracts. Judging by the old White Paper, the language would be based on formal verification just like the language of the recently reviewed CertiK project.
fBFT, that is a hybrid technology made of BFT, DAG, and PoS consensus protocols and BLS cryptographic multi-signature. In theory, this should increase the consensus productivity a hundred times, add staking and make all nodes equal.

Now let’s move on to the questions. We did not divide them into parts, as they are all general purpose.

If a node has to store the data of many shards and constantly synchronize with other shards, can we call it a complete sharding? What will you do if there is a problem with communicating the master node? And isn’t that a vulnerability to a “1% Shard Attack”?
When will you start using VDFs? Who will provide altruistic nodes running an ASIC to compute VDFs? And what will motivate the owners of these nodes?
From the White Paper “<…>an attacker with a faster computing device could calculate the result before other honest nodes.<…> this problem can be mitigated on the smart contract layer with a proper delay”. What are the technical aspects of this solution?
You assume malicious groups to be slow; therefore, they don’t have time to sync. Why don’t you assume the existence of a group of fast malicious nodes?
What if a malicious node creates a smart contract to encourage the shard participants to test bad blocks that would create money out of thin air?
How do you verify whether a transaction in a shard is incorrect?
How did you solve the problem of trust and inter-shard transaction verification?
In the case of blocking, how does the system conduct parallel processing of transactions between shards from the same account?
What is the structure of shard accounts? Is it a single account, a number of separate accounts, or several combined accounts (sidechains)?
How are you going to use QUIC and update its codebase? After all, the server side of QUIC uses epoll meaning you can only synchronize blocks on Linux. Could you provide any technical details?

What has already been developed:

Why do you dissolve your efforts into side projects, instead of focusing on the development of the main product?
In testnet monitoring, the results do not exceed four transactions per second. The team claims there are 114,000 transactions per second in the testnet. Where is the truth?

There are sixteen repositories on Harmony’s Github. Most of the code is copied from other projects. There is little activity in the main repository that has an average of a dozen commits a week.

According to the project roadmap, the following should have already been implemented: resharding, VDFs, cross-shard routing, BLS sharding with a multi-signature, zkproof, and ECDSA algorithm.

Where can we look at how you’ve implemented these technologies?

In June 2019, you are planning to launch a public main network, and at the end of the year, you intend to begin the transfer of your partners’ dApps from Ethereum to your blockchain.

What are the technical aspects of the dApps transfer from Ethereum to Harmony?

Advisors

There no “Advisors” section on the project website, but there is a section titled “Collaborators.” Four people mentioned there turned out to be advisors after all, as we guessed from what the team says. After letting the team know, we wrote to the advisors via LinkedIn and e-mail asking to confirm their participation in the project.

Hakwan Lau, a full professor of cognitive and behavioral neuroscience at the University of California, Los Angeles. Lau founded CONSCIOUSNESS & METACOGNITION LAB, a laboratory studying the human brain and analysis of perception and behavior. From the Harmony website: “With Harmony, he is exploring the connection between probabilistic consensus protocols and brain communication. Hakwan is also studying privacy-preserving modeling of health data on blockchain”. However, we couldn’t find any research works in these areas written by Lau. In his CV, Lau indicates working on simple rules of the Harmony protocol in June 2018.

Could you share links to Lau’s studies on blockchain and machine learning?

Ka-yuet Liu , an associate Professor specializing in social networks and social epidemiology at the University of California, Los Angeles. From Liu’s profile page at the UCLA website: “My research asks questions such as: what are the network effects of individual suicides? What are the roles of social influence in the rising prevalence of autism?”. Liu doesn’t mention Harmony in her CV. The Harmony blog on Medium features an article about storing scientific data in the blockchain authored by Liu.

What kind of issues does a sociology researcher help you with?

Zi Wang , a former product leader at Google.
Bruce Huang , a former Director of Alibaba Cloud and Credit Ease. He also used to be the CEO of Madailicai and work as an engineer at Microsoft.

We wrote to the advisors on March 20, 2019. It’s been 2 months since and still there is no answer.

Could you ask your advisors to respond to us on LinkedIn?

Previously the Collaborators section featured six more advisors that are now removed from the website. We wrote to them too via LinkedIn. Only one of them answered:

Aaron Li, the founder of the Qokka feedback aggregator. He used to work as a developer at Google (summer internship) and Scaled Inference (for a couple of years). Li also took part in research on artificial intelligence. He’s active on his Github account.

He immediately responded that he had been advising Harmony on technical issues last year: artificial intelligence, machine learning, distributed systems, and natural language processing. Also, Li asked us to mention that he is not associated with the project financially and did not take on any formal role.

Team

The team page lists twelve people. All of them are located in Silicon Valley according to the website.

Stephen Tse, the founder and CEO

PhD in Cryptographic protocol, type theory, functional compiler at the University of Pennsylvania;
Co-author of eight scientific articles on cryptographic protocols and typed programming languages;
He participated in the development of architecture and tools for verification of security protocols at Microsoft for four months (probably during summer internship). He also developed a prototype of a functional Java compiler in F# to test .NET programs;
Tse points out that in 2006 he was a Research Intern on Google’s fast computing and charting algorithms. And from 2007 to 2010 he worked as a front-end engineer in Google Maps.
Since 2011, was a co-founder and the CTO of the Spotsetter app. Spotsetter recommended users places to go based on their social media news feeds. According to the media, the application was popular and had five million users in 2013. In 2014, Apple acquired the app and shut it down.
After the startup was bought-out, Tse worked as a search engine ranking engineer at Apple for one year.
He developed Voice AI, a voice companion project. According to Stephen, the project was not successful because of the unsuitable business model.
Tse wrote a platform for automated stock trading for the brokers of Interactive Brokers and IQFeed using OCalm language. We couldn’t find any information about the platform on the web.
Tse’s Github account is empty and inactive.

Where can we find the automated stock trading platform that Tse wrote?

Do people now use the products that Tse used to work on? Are they useful?

Where can we find evidence of Tse’s work at Google and Apple?

Nicolas Burtey, a co-founder and the COO. From 2012 to 2018 Burtey was working on a startup called Orah, a live streaming software for panoramic VR videos. The project website won’t work, but judging by the web archive, the website used to a considerable audience back in 2017. The project’s YouTube channel has about 10 thousand views for each of its dozen of videos. Crunchbase reports that the company conducted a private IPO, raising $8 000 000 as a result. After a couple of years, Orah was shut down without any public announcements.

Why was Orah shut down? What happened to the investors’ money?

Does Burtey have any other experience in managerial positions?

Alok Kothari , a co-founder and engineer. Kothari co-authored the collection of motivative stories for entrepreneurs titled “Game Changer.” For four years he was engaged in research on information search. For almost three years, he worked at Apple on Siri machine learning. What we assume to be Kothari’s Github account was created at the time Harmony was founded. Kothari participates in the public development of the Harmony blockchain and smart contracts. Kothari’s Github account doesn’t have any working projects of his own.

Where can we find evidence of Kothari’s work at Apple?

Rongjian Lan , a co-founder and engineer. For four years Lan worked as a search infrastructure engineer for Play Store at Google, for two years he was a co-founder and the CTO at Joobali, a payment service for childcarers and parents. For over a year now he’s been a co-chair of the ABCer crypto community. Lan is active on Github and intensely involved in the public development of the Harmony blockchain. Lan’s Github account doesn’t have any working projects of his own.

Where can we find evidence of Lan’s work at Google?

Minh Doan, a co-founder and engineer. Judging by Doan’s CV, he used work as a developer at Yandex, COA Solutions, Thefind, Ooyola, and Google. Doan also claims to have worked on TTS Google Assistant, Google Home, Screen Devices, Play Store, and Google Plus. He patented “Publisher Click-Ring Fraud Detector” during his work at Google. We found no public evidence of any of the above. Doan is active on Github and intensely involved in the public development of the Harmony blockchain. Doan’s Github account doesn’t have any working projects of his own.

Where can we find proof of Doan’s work at Google? Could you give us a link to the patent?

Nick White, a co-founder. He’s also a yoga instructor and surfer. Until 2017 he worked in the field of electrical engineering. Then he advised Asian AI startups for a year. His Github account is empty, and his informational footprint is practically non-existent. In one of AMA threads, Nick mentioned that his role at Harmony is research, development, hiring, dealing with partners and investors, organizing events and interviews.

What experience does qualify White for his role in the project? What startups did White consult? How did he manage consulting AI startup while having no practical experience in the area (based on his LinkedIn account)?

Sahil Dewan, a co-founder. Dewan holds an MBA degree from Harvard Business School. For a couple of years, he was President of AIESEC India, a nonprofit students association. Dewan created the Alumni app, a database of University graduates based on their LinkedIn profiles. The application wasn’t a success. Later he used to advise startups at Draper Dragon fund.
Leo Chen , an engineer. Chen worked for nearly ten years as a developer in a number of large (Ericsson and Broadcom) and small IT companies in Vancouver. For seven and a half years, Chen worked as a software development Manager at Amazon. Chen is active on Github and intensely involved in the public development of the Harmony blockchain. He also has several small projects of his own there.
Eugene Kim, an engineer. Kim worked for almost seventeen years as a leading network Protocol developer at NTT, for a year and a half at Blizzard, and three years at AWS. While working at Amazon, he registered two patents [1|2] for calculating the network speed. Kin is active on Github and intensely involved in the public development of the Harmony blockchain. He hasn’t published any working projects of his own.
Chao Ma , an engineer. Ma worked for four years as a research and development engineer at Ricoh and for two years as an applied scientist at Amazon. Ma’s Github account only features his works on Harmony.
Li Jiang, a business development manager. He worked for seven years as Vice President of GSV, an investment advisory company. Jiang is a co-founder of Ympact, a company that shoots interviews with startups.
Helen Li, a PR manager and marketer. We didn’t find any information about her online. According to Harmony’s website, she worked a lot as a journalist.

In addition to the leading team, there are five community managers, seven content managers, and two bounty managers — all in all 26 people.

Harmony now is recruiting new team members, promising cash rewards for recommending specialists in development in Rust and Go.

The team willingly answers questions in their official chat and often conducts AMAs.

Still, Harmony’s marketing policy is dishonest. If you go to any of the project’s publication on Medium, you see about 50 flattering and obviously bought comments. The commenters’ profiles don’t look real either. You can easily see that those are specially created accounts to imitate user activity.

Harmony’s Telegram chat used to undergo intense mass inviting at nights until the real users asked to add Shieldy to stop cheat bots activity.

Partners

The project website lists ten partners and twelve investors.

Most partnership projects are raw, and Harmony’s advisors actually own three of them. Of all the partners, we could only find one publicly confirmed. It is Hyperion, an Isaac Zhang project, who was previously listed as one of the advisors.

We wrote to all publicly unconfirmed partners via e-mail and Facebook. We’ll update this review as soon as they answer.

Who else are your partners and are they publicly confirmed? Do they plan on moving onto your blockchain from the ones they’re using now?
What would attract and motivate end users to use your blockchain?

Of the twelve listed investors, only a half is publicly confirmed. Investors like hayek.capital, uva.fund, bca.fund, qtum.org, and skunk.capital have not added Harmony to their portfolios.

The team claims to have received $18,400,000 in the seed round for 22.4% of the total tokens that were estimated at $82,000,000.

Who conducted the evaluation, and what exactly did they evaluate? Can we see the results of the evaluation?

In the token purchase agreement, Harmony indicated an Ethereum wallet as a destination for investors’ money. According to our calculations, the amount of about $5,600,000 was collected and withdrawn from this wallet in 2018. The current balance is 51 ETH (~$9 000).

How did you get the rest of the money and is there any evidence of that happening?
How much of the collected money have you already spent? Can we see your financial plan?

Legal part

Harmony refuses to publish any legal information, but the team is ready to communicate on the topic in their Discord.

According to media reports, the project is registered in the United States. A Harmony’s PR managers, who answer our questions on Discord, said that the company’s Fund is registered in Panama. The token purchase agreement, in its turn, refers to a certain company titled Ten Times Technology Ltd, registered in the Cayman Islands. We couldn’t find any public information confirming any of it.

According to the team, when it comes to legal issues, Harmony is advised by Fenwick & West LLP.

The project’s utility tokens are designed for node functioning, staking, and payment transactions. The team forecasts that users would receive up to 15% of income for the staking per year.

Forecasting income, you risk falling under SEC sanctions. Wouldn’t you agree?

The project doesn’t own any patents, although the team claims that they have developed many proprietary technologies such as Verifiable Random Function, Verifiable Delay Function, Fast Byzantine Fault Tolerance, and Adaptive IDA.

Where can we find these technologies, and can they be used in other projects?

Conclusion

The project has a beautiful wrapper, like any other project conducting IEO. Harmony creates an image of a trustworthy project thanks to their ability to promptly answer any questions, develop the community, and enter into partnerships. The team also has an impressive academic and technical background. In fact, for many reviewers, the mere existence of a testnet is enough evidence of an MVP.

But still, let’s put the pink glasses aside. Harmony is a thick bundle of forks. Even if the team gets all those technologies together, it won’t work. And this is not an iPhone, in which all the components are tested in practice and specially sharpened by engineers to become parts of a mobile phone. It’s just a set of high-level lab concepts that require a great amount of work to become anything practical at all.

Take Zilliqa, that has been struggling with sharding for years and still struggles. Most likely, Harmony is in for the same path. They either launch the mainnet in June, as was announced, and dive deep into debugging or, as other sharding projects did, just postpone the launch date. All in all, the market is already overflowing with “fastest” blockchains that have nothing close to the practical use that Ethereum brings. Does the market really need another one?

And as for the team. Try to name at least one now active product made by any of them, a product that they launched and scaled up to at least ten thousand active users. Anyone can write “Apple,” or “Google,” or “Amazon” on their LinkedIn account, but do they know how to bring real benefits to real people?

$18,400,000 on the seed round is just way too much. Usually, no one ever gives that much money on the seed round, let alone for tokens. The seed round is the riskiest round, where the minimum amount of money is allocated to run the minimum viable product: from $25,000 to $1,000,000 per company share. A bunch of Github forks just can’t possibly be estimated at $82 000 000!

Legally, the project is murky — the corporate structure, the country of registration, reports on early investments submitted to SEC, none of it is public. Everything is a mystery for the time being. But perhaps we’ll get this information when the project comes to IEO.

That’s all. Feel free to clap or rant in the comments if you see errors in the text. Thank you!