Discover more from Association of Cryptocurrency Journalists and Researchers
🥁 And the beat goes on 🥁
The best crypto and blockchain journalism and research in March
Crypto marches to the beat of a different drummer (not pictured)
We’ve got a new face! The most important news of the past month is obviously that the ACJR now has a brand new spanking website that we highly encourage all of you to check out. You can apply to be a member, see our upcoming events and learn more about why we even exist in the first place. Besides that wonderful news, March has been a month categorized mainly by the sheer number of shutdowns, indictments and criminal cases happening in crypto — from multiple banks failing to Do Kwon’s Montenegrin arrest to the SEC’s charges against Binance, Justin Sun, Lindsey Lohan, and others…
On a brighter note again, the ACJR also had a fantastic Off the Record session on crypto policy (more timely than ever!), which you can watch here if you missed.
Before we dive into the top crypto research and journalism this month…..
If you’ll be at Consensus in Austin this year, and you are a member of the crypto media, research, or PR space, we’d like to welcome you to our happy hour drinks and panel, “How to make your pitch…perfect.”
Moderated by Joon Ian Wong, co-founder of ACJR & independent journalist/researcher
Sam Reynolds, senior markets reporter, CoinDesk
Joyce P Hanson, community manager, ACJR
Saad Quershi, director, YAP Global
Molly Jane Zuckerman, opinion editor, Blockworks
Sign ups will be available here, keep checking! Please note that the first part of our happy hour is journos only while the panel runs — then, we open it up for all of y’all!
Sponsored by CoinDesk, Blockworks, Eleven International, the Filecoin Foundation and Shift Communications.
On to the articles!
Top crypto journalism in March
All about Do Kwon — even though this isn’t an article, we’d like to shout out DL News, who sent not one but TWO reporters to Montenegro to get this footage of Do Kwon leaving jail.
Amidst all the confusion of Do Kwon’s arrest — considering that it was first announced via an unverified Twitter account of a Montenegrin official — this piece did some digging to reassure its readers that Kwon was, in fact, under arrest in the Balkan country. How did it prove that the tweet was real? Well, by noting that the sketchy-looking minister’s account was followed by the prime minister of Montenegro’s official account — and that’s what you call fact-checking!
When writing this newsletter, we almost forgot that the biggest news of the month had nothing to do with foreign arrests or SEC charges — a whole bunch of banks failed, and it kinda (a little bit but not really) had something to do with crypto!
Barney Frank: Regulators Shuttered Signature Bank to Show 'Crypto Is Toxic' by Stephanie Murray, The Block
The Block got the exclusive interview with Barney Frank, which all other news outlets were probably jealous of — especially considering that Frank’s beliefs about a possible crypto “operation choke point 2.0” became a big narrative in the later coverage.
The Banking Crisis Is Not Crypto’s Fault by George Kaloudis, Coindesk
This opinion piece breaks down nicely the issues with conflating banks that deal with crypto customers, and why the banks themselves failed (or were taken over).
Barney Frank Was Right About Signature Bank by the Wall Street Journal Editorial Board
Plus, it’s always nice when a large, mainstream newspaper thinks that there is a conspiracy out there to shut down the crypto industry in the U.S., isn’t it?
A New Crypto Mixer Promises to Be Tornado Cash Without the Crime by Joel Khalili, Wired
Moving on from banking back to crime…this piece explores how one of the original architects of Tornado Cash is still going to attempt to build another mixer.
Reviewing Code Is Mind-Numbing: Q&A With Bitcoin Maintainer Andrew Chow by Frederick Munawa, Coindesk
Sometimes it seems like those developing Bitcoin do so in a black box — they might not be as public or vocal as other crypto influencers, and what they do is quite technical. Having a conversation with a Bitcore Core developer definitely provides an insight that we don’t see that often.
Euler Hack Victim Who Got 100 ETH: ‘He Was Probably Moved by My Message’ by Tim Craig, Eric Johansson, DL News
In a very strange story, DL News got the exclusive with an interview with someone who actually had money returned by a hacker (this was before the Euler hacker ended up returning even more money to Euler itself) — bizarre interview, but very cool get.
The Original King of Crypto Is Back Jen Wieczner, Intelligencer
In this lengthy interview, we learn that Arthur Hayes collects stuffed animals. Enough said.
NBA Top Shot CEO’s Decadent Lifestyle, ‘Public Shaming’ Led to Toxic Culture as Dapper Flails by RT Watson, The Block
Current and former Dapper Labs employees call ‘foul!’ on CEO Roham Gharegozlou. With a meaty eleven sources, The Block went to the nines to paint this picture of the creator of NBA Top Shot’s meteoric rise and fall.
💥EMERGENCY POD: Roham Accused of Lavish Lifestyle, Bullying by LG, Phil D, The First Mint
Another take on Dapper Labs CEO Roham from within the web3 sports community. Hosts LG and Phil D, being part of the Top Shots community specifically, have unique insight into the situation and other fan reactions.
Amazon NFT Marketplace Could Feature Beeple, Pudgy Penguins by Michael Bodley, Blockworks
Furthering Blockworks’ coverage of Amazon’s NFT initiative, Michael outlines what we know about the retail giant’s planned marketplace and what they might do next.
CoinMarketCap Acquisition of CoinDesk ‘On Hold’ by Jon Rice, Blockworks
In addition to reporting the latest in Binance’s possible acquisition of CoinDesk via CoinMarketCap, Jon includes the history of the sale as well as information from a source exploring the impact of the Binance brand on any acquired media companies.
$10B Mystery: Is OKB Really the 7th Largest Crypto? by David Canellis, Blockworks
Truly one of the great mysteries of 2023, OKB’s sudden presence among the top cryptoassets by market cap get broken down in meticulous detail by super sleuth David Canellis. A great blend of investigative journalism and on-chain analysis, David presents his findings in a credibly neutral way and includes OKX’s response to share their point of view. Wild read for fans of hard-boiled on-chain detective work and crypto corporate shenanigans alike.
We Wish You Did Better 😢
Binance Hid Extensive Links to China for Several Years by Scott Chipolina, Financial Times
In somewhat of an attempt to cover the recent lawsuit brought against Binance, this piece in the Financial Times falls flat in digging up old criticisms against the company and its chief executive without providing any substantial, current, or relevant sources.
And now — on to some research!
Top crypto research in March
US Federal Income Tax Analysis of Liquid Staking by Proof of Stake Alliance
This white paper from the Proof of Stake Alliance gives an overview of liquid staking, the mechanism by which users of a blockchain protocol can deposit (“stake”) tokens with a given product and receive receipt tokens in exchange and whether those tokens should be considered as taxable income in the United States.
The Ineluctable Modality of Securities Law: Why Fungible Crypto Assets Are not Securities by Lewis Cohen, Greg Strong, Freeman Lewin, Sarah Chen, DLx Law
This paper presents a compelling legal argument against fungible crypto assets (like Bitcoin, ETH, or most ERC20 tokens) as securities. From the abstract: This Article addresses the federal securities law status of fungible crypto assets not intended to be a type of traditional security, with a focus on secondary transactions in these assets, such as those effected on a centralized crypto asset marketplace, like Coinbase, or through the use of a smart contract-based protocol, like Uniswap.
From the policy analysts at Paradigm, this series explores what happens to crypto companies seeking to comply with regulation in the United States. The first part introduces the series and gives relevant background on crypto companies in the US, Part II outlines the SEC registration process and includes examples of crypto companies who have attempted it.
Polynonce: A Tale of a Novel ECDSA Attack and Bitcoin Tears by Nils Amiet, Kudelski Security Research
This blog post and paper document applying a method for attacking ECDSA, the algorithm by which public/private key pairs are derived on blockchain networks like Bitcoin and Ethereum. In experimenting with the attack, the team also uncovered evidence that a similar attack was successful in draining vulnerable wallets in the past. As an added bonus, the post includes a paper with open source code on recreating the attack and defending against it in public-facing products.
GPTs Are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models by Tyna Eloundou, Sam Manning, Pamela Mishkin, Daniel Rock
One of the questions on everyone’s mind: will ChatGPT take my job? This paper seeks to answer that question. From the abstract: We investigate the potential implications of large language models (LLMs), such as Generative Pre-trained Transformers (GPTs), on the U.S. labor market, focusing on the increased capabilities arising from LLM-powered software compared to LLMs on their own. Using a new rubric, we assess occupations based on their alignment with LLM capabilities, integrating both human expertise and GPT-4 classifications.
Bitcoin Mining Stock Roundup: February Monthly Numbers by Anthony John Power, Compass Mining
Bitcoin mining has been having a rough go of it lately. As the industry shifts and more operations spin up within the United States, Compass’ monthly summaries are a valuable resource for staying informed and observing trends.
Bitcoin Inscriptions & Ordinals: A New $5bn Market by Alex Thorn, Brandon Bailey, Charles Yu, Gabe Parker, Guillaume Girard, Simrit Dhinsa, Galaxy
This whitepaper from Galaxy takes a deep dive into Inscriptions, digital artifacts attached to ordinals, a single distinct satoshi of Bitcoin. The authors explain how these “Bitcoin NFTs” work, how they’re being used, what the future market could be, how they can help address the Bitcoin security budget problem, and address the inevitable controversy around anything good that happens in crypto.
Top crypto podcast in March
Have Crypto Detectives Killed the Cypherpunk Dream? by Laura Shin, Unchained
Veteran crypto reporter Laura Shin interviews Andy Greenberg, author of “Tracers in the Dark.” Some discussion highlights from the show:
How blockchain analytics started being used to tackle crime
What the impact of zero-knowledge technology will be for blockchain analytics firms
Whether the cypherpunk ethos is dead
Top crypto unhinged tweet(s) in March
Really? Is this specific case really a shared experience for anyone else?
And so we beat on…
Yes, we are telling you that.
Subscribe now! It’s never too late — in crypto, there’s no such thing.
Join the ACJR
Reminder! Becoming a member is now only $25 — that’s just one avocado toast! (If you live in San Francisco…). Dues-paying ACJR members get more access to ACJR stuff, like a secret members’ only chat, discounts and conference things and invites.
If you want to become an official member of the ACJR, please reach out to any of the current board members on Telegram, or Joyce Pavia Hanson: @JPHanson.
We’ll leave you with some food for thought about AI, crypto and how they can work together. It’s a long one, so settle in and don’t forget to click that “read more” if your email app is trying to deny you the ending!
Can ChatGPT be the Transformer of Crypto?
By LK Shelley
And if so, can we prevent any recurrence of large-scale financial fraud like FTX (which is yet to be judged in court) and strive to bring the next billion users to the blockchain safe and sound?
On Nov. 30, 2022, ChatGPT (built on GPT-3) was released and the ChatGPT fever — and simultaneously, the AI fever — began. By December, more and more YouTubers were posting content related to making money with ChatGPT, which led to an inflation of the AI hype. This aha moment is similar to the "Beeple" aha moment for NFTs, when his “Everydays: The First 5000 Days” digital art was auctioned off by prestigious auction house Christie for an astonishing $69M
It turns out that the text chat format and its human-like interactive dialogue feature of ChatGPT was what was needed to wow people with the possibilities of AI, far more than the AI-generated graphics and images.
The crypto space can often be a frustrating and unpredictable universe.
While there are tools available to access hard data such as on-chain analysis, there is no universal and regulated reporting and interpretation of data. Certain data can be easily manipulated through wash trading, data manipulation, and market manipulation. The total value locked (TVL) metric and indicator, for example, was manipulated in the case of Solana by two developers.
In summary, crypto has some complicated problems that AI can’t solve easily:
Reflexivity (h/t Arthur Hayes)
Tribal, social influencing
Sentiment and narrative, game theory driven
Lack of (standard, valid, and audited) data
Data and information asymmetry
Investigative journalism, the Twitterverse, and Crypto Twitter are often ahead of the game when it comes to detecting abnormal activities and irregularities, such as Cobie's detection of insider trading and postings on Twitter and other NFT detectives uncovering the Beanmaxi scams. Coindesk's reporting and revelation of the FTX and SBF frauds, which will be adjudicated in court, is an epic find. The problem is that these issues are often not detected ahead of time and may not be easily recognized by the general public, or even by veteran and experienced Wall Street firms.
Accurately labeling or classifying sentiment and emotions is one of the most difficult tasks in the field of AI. It is even more challenging to determine whether a sentence or block of text is truthful or not. Open competitions on Twitter Sentiment extraction and analysis are quite popular, drawing many researchers and participants.
Now, let's take a closer look at what ChatGPT can do (or cannot). A good way to illustrate the concept is with a tweet from a crypto marketer working for a firm that promotes token projects. The tweet was short and straightforward, but it's easy for human eyes to see that the intention was to hype up the $APT token and create a fear of missing out (FOMO) among followers. This type of pump-and-dump speech pattern is common, but it can't be detected by most general large language AI models like GPT yet.
During the 2021 bull run, it was relatively easy for many crypto outsiders to be taken in by the hype. However, it was relatively simple to differentiate between genuine enthusiasm and manipulative rhetoric. On the other hand, if a ChatGPT-like chatbot was asked to evaluate a short tweet post, the result often missed the subtle context and intention behind the message. Because of certain positive wording, the sentence was often identified as being overall positive.
The screenshot shows two ways of asking for a judgment from a ChatGPT-like chatbot. Initially, it thought the tweet was positive. When asked to judge whether the statement was true or a lie, it could not provide a definite answer.
As many cryptocurrency influencers are much more subtle in their messaging, readers need to delve deeper and analyze the subtleties in the writing and expression in order to determine the true intention and story behind the scenes.
IBM Watson Research Lab's Text Analyzer Classifier (built in 2000) provides a comprehensive look at the inner workings of text analysis in NLP. To demonstrate the capabilities of this text analyzer, we will focus on two classifications: Sentiment and Emotion. What makes this text analyzer particularly interesting is that it can detect both the overall positive emotion ("Joy") of the first tweet about the $APT token above, as well as the negative sentiment in the text.
This is one example that illustrates the potential of AI and machine learning technology. However, the success of such models lies in the quality of the training data and model.
Where do we go from here?
For specialized fields such as crypto and investments, a tailored approach and specialized model are needed to help reduce the workload and serve as the “superheroes” for both crypto natives and "normies" (outsiders). While the current state of AI and machine learning does offer some capabilities in dealing with the issues, the key is to go deeper and find the right model, as well as learn how to gather the right training dataset and train the model to solve the first layer of the problem: detecting and finding red flags in crypto projects using a combination of large amounts of unstructured data and some quantitative analysis to validate and evaluate the projects in question.
Large language models and products from ChatGPT are still in their early stages when it comes to integrating AI into society. They are neither evil nor superheroes, but are capable of taking over and automating certain mundane tasks to improve productivity and efficiency in various industries, particularly those involving structured tasks or outputs that don't involve subjective elements such as emotions, sympathy, or other aspects of human life. It is important to remain calm and let AI innovation continue to develop in a healthy and sustainable way.
Reach out to Molly Jane or Anthony on Telegram if you have some ideas for our next op-ed of 2023.