Crypto vs AI: Destined For Competition or Cooperation?
Crypto vs SEC| Arbitrum 5x Leverage Accounts | AI Crypto Ecosystem
GM, this is your Daily Bolt briefing.
In today’s extended notes, we’re providing you with an extended briefing on the latest around Coinbase and the SEC’s ensuing legal battles, as well as some insight as to how other big companies around Crypto are being impacted and have responded to the latest regulatory precedent.
Also; read our report on Sentiment, an under-collateralized borrowing platform on Arbitrum, and learn more about their goals, and their upcoming V2 upgrade.
If there’s anything worth keeping tabs on other than crypto, most would argue that AI probably takes the cake… Read our note of the day to learn more about the potential in-depth intersection between Crypto & AI, and all the intriguing possibilities it could bring.⬇️
Stay Vigilant.
1/ Unchained Podcast Ep. 486 – J.W. Verret on Coinbase’s Legal Action
Preview: In this episode of the Unchained Podcast, which took place on April 28, 2023, host Laura Shin is joined by J.W. Verret to discuss Coinbase’s legal action against the SEC, Verret’s thought process on Coinbase’s potential strategies, and more! Read our notes below to learn more. Read our notes below to learn more. Click here to listen to the full episode (36 mins).
Read our Note (10 mins) and save 26 mins.
Here are some key takeaways:
J.W. Verret explains that Coinbase and other crypto exchanges have tried to fit into the current regulatory regime by seeking various licenses i.e. special broker-dealer license, and Regulation ATS license, but have been denied.
Coinbase filed a request for public rulemaking, but the SEC has now sent a Wells Notice (threat to sue) to Coinbase.
Laura Shin highlights Gensler’s view that crypto companies should be compliant but notes that there isn’t a clear way for them to comply, and asks for elaboration on the potential legal strategy.
J.W. Verret believes the SEC generally doesn’t respond to requests for rulemaking unless it aligns with their interests. He thinks the SEC is trying to build up successful cases against tokens and use that as precedence.
Verret admires Coinbase’s strategy of not letting the SEC pick the place of battle but has qualms about the Kraken case, where a $30 million penalty was imposed for non-registration without fraud.
He argues that the penalty was an overreach and that the SEC should have allowed Kraken to develop a registration system for staking instead of shutting it down.
The SEC argues that most digital assets are securities, while Coinbase maintains that the cryptocurrencies listed on its exchange are not securities.
On April 25, 2023, Binance U.S. terminated its one billion dollar asset purchase agreement with Voyager Digital, a bankrupt crypto lender, citing a hostile and uncertain regulatory climate in the U.S.
Voyager responded by stating that its Chapter 11 plan enables the direct distribution of cash and crypto to customers through its platform, and it will provide information on how creditors will be reimbursed via direct deposits. Despite a prior agreement between Voyager’s creditor committee and the U.S. government, as well as approval from the majority of Voyager creditors and Bankruptcy Judge Michael Wiles, the termination of the agreement still occurred.
Binance has lifted restrictions on Russian users, allowing them to use locally issued credit cards for deposits.
In a post-FTX world, securely trading with leverage and earning real yield in a decentralized manner has never been more important.
aiPX offers cutting-edge risk management for liquidity providers, leverage trading, and a suite of products between perpetuals, binary options, and synthetics.
Earn passive yield and trade with leverage straight from your wallet.
Take the step, join the presale.
https://app.aelin.xyz/pool/arbitrum/0xe10616e9c424463b121c146bfdd9b458f1a049db
2/ Arbitrum – Introduction of Sentiment and Capital Efficiency in DeFi
Preview: In this Twitter Spaces, Churro from Arbitrum invites the protagonist and Ruvaag from Sentiment. They discuss the market structure of DeFi protocols, products, capital efficiency that Sentiment offers, and the risk in DeFi. Read our notes below to learn more. Click here to watch the full episode (42 mins).
Read our Note (5 mins) and save 37 mins.
Here are some key takeaways:
According to the protagonist, Aave and Compound were good enough but their over-collateralized lending was inefficient.
the protagonist and Ruvaag wanted to make lending more equitable and efficient for borrowers. This led to the launch of Sentiment, which aimed to provide more capital efficiency for borrowing experience in DeFi.
Sentiment is a protocol that allows users to access up to five times leverage or up to five times the value of their collateral.
Users can deploy these borrowed assets into different strategies or assets that are whitelisted on the platform.
the protagonist says that many people think of Sentiment as a generalized margin protocol where users post margin and get access to capital that they can use for buying assets or investing in different strategies with yield potential.
the protagonist says Sentiment’s user interface is designed to make it incredibly simple for users to be a lender or a borrower. He adds that:
As a borrower, it is expected of you have up-to-date analytics on your account, including the value of your assets and debt in your account.
Users can access a swap page with integrations with all the top AMM protocols such as UniSwap, Balancer, Sushi, etc.
There is also an investment page that displays all the different yield pools that users have access to.
the protagonist states that Version 2 of Sentiment is planned to come out towards the end of this year, which will be an order of magnitude more efficient than the current system.
If you read these 2 Notes on Revelo Intel you would have saved: 1 hour and 3 minutes!
Bankless – Intersection of AI, Cryptocurrencies and Blockchain
In This Episode, David And Ryan From Bankless Invite Mohamed Fouda And Qiao Wang From Alliance DAO To Discuss The Intersection Of AI, Cryptocurrencies, Blockchain, AI Coins, Use Cases Of AI In Real Life, And Potential Threats Of It. Read Our Notes Below To Learn More.
Background
David Hoffman and Ryan Adams from Bankless (Host) – David and Ryan are the co-founders of Bankless and frequently share their market insights and interview project founders in the crypto space.
Mohamed Fouda (Guest) – Crypto researcher and Investor, Contributor at Alliance DAO, Venture Partner at Volt Capital.
Qiao Wang (Guest) – Contributor at Alliance DAO
Alliance DAO – The leading Web3 accelerator and founder community.
Crypto And AI Intersection
AI has been around since the 1960s and 1970s while crypto got introduced in 2010 after Bitcoin.
Fouda co-authored with Wang and wrote an article titled “The Convergence of AI and Web3: The Opportunities and Challenges“. The article explores the ways in which AI might find itself in Web3.
Fouda and Wang believe that there is a large intersection between crypto and AI/Web3. This intersection includes both opportunities and also challenges to be addressed.
They believe that crypto is one of the few technologies that can steer AI. Other technologies cannot shift how AI works or performs, but crypto can.
Wang states that blockchain chains, digital signatures, and zero-knowledge proofs will play a bigger role in AI in the next few years. Each technology will affect the other, and there will be a lot of synergies.
However, they worry about the possibility that some areas may conflict because these technologies were introduced independently and didn’t share a background.
According to Fouda, there are people who tend to see the idea of crypto as a check and balance against AI.
Cryptocurrencies Based On AI Technology
“AI coins” are tokens with AI in the name. Fouda suspects they are not really based on AI technology but rather based on a narrative trade.
Render Protocol and Alethea are two examples of AI coins that Fouda and Wang provide as an example to look at.
The Use Of AI For Payment And Execution Rails
Fouda believes that payment and execution rails for AI agents are an intersection point between crypto and AI. This involves using crypto as a way to pay for services provided by an AI agent.
He mentions that crypto wallets allow for fully autonomous AI agents that can receive payment for their work.
He reiterates that bank accounts are not suitable for AI agent payments due to KYC requirements and non-permissionless interfaces.
Decentralized Computing
According to Fouda, decentralized computing is another intersection point between crypto and AI. Decentralized computing is important in building models that are not controlled by powerful corporations. This involves using blockchain technology to provide decentralized GPU for AI.
Fouda says that decentralized computing can be useful for training smaller models without relying on AWS, but larger models require centralized systems like AWS.
Fouda states that open-source models on local devices can achieve decentralization on the inference side of things.
AI Agents Dominating Blockchain Networks
Fouda states that the intersection of blockchain, zero-knowledge proofs, and AI is super fascinating.
Fouda envisions a world where AI agents or AI in general will dominate the edges of blockchain networks since blockchains provide a good interface for AI to participate and flourish. This implies that bots could produce meaningful content on decentralized social networks or games on top of the blockchain.
Fouda thinks crypto is currently the only world where AI has “human rights” and is treated as an equal to human beings.
According to Fouda, permissionless systems are needed for LLMs (large language models) to produce insightful content.
Centralizing Force Of AI Vs. Decentralizing Force Of Crypto
According to Fouda, the more data and feedback from humans, the better an AI model becomes. This centralizes power structures and benefits large institutions since it requires large aggregate data sets.
On the other hand, he thinks that crypto is a decentralized technology where individuals can be their own bank with private keys that cannot be taken away. Cryptography and crypto-economic systems largely benefit individuals rather than large institutions.
He believes that blockchain can discover the potential of decentralized AI through open-source models on local devices.
Shifting Power Structures
Fouda observes that the power pendulum shifts from large centralized institutions with AI to individuals with cryptography and crypto-economic systems.
He thinks that the potential for decentralized AI through open-source models on local devices may further shift power structures from large centralized institutions to individuals and decentralized entities.
AI Agents
Fouda explains that an AI agent is a bot that has access to LLM and databases to store data. It interacts with LLM at the back end and can execute tasks autonomously.
He reiterates that AI agent is a new phenomenon that started less than a month ago when a VC created something called Baby AGI, which exploded into many variants such as AutoGPT.
Fouda breaks down the AI agents into two forms. The first is the specialized AI agent which is trained for specific tasks, such as being bankless. The second is the general form of the AI agent which has storage capabilities, accesses GPT using APIs or other elements, and executes tasks based on given directives.
Fouda emphasizes the potential of AI agents in automating various tasks, such as building websites or launching podcasts.
Fouda states that AI agents can be fully autonomous and execute tasks based on given directives.
Creating A Decentralized World
Fouda and Wang discuss that AI agents can be used to create a more decentralized world by automating various tasks and reducing the need for centralized entities.
They believe that this could lead to greater efficiency, transparency, and security.
DeFi Composability With AI Agents
Fouda states that smart contracts are just one API call away from each other, allowing different financial applications to interact.
He believes that this composability of DeFi protocols is immensely bullish due to its potential for innovation.
He thinks that adding AI agents as a layer on top of DeFi protocols opens up even more possibilities than before.
Creating Marketplaces For AI Agents
According to Fouda, the creation of marketplaces for AI agents is an investment opportunity that could benefit the crypto industry. These marketplaces would allow AI agents to discover each other and work together towards common goals.
He thinks this would eliminate the need for human intervention in tasks such as registering a business or figuring out payroll taxes across multiple jurisdictions.
Crypto As A Banking System For Robots
Fouda reiterates that smart contracts have created a banking system for robots in crypto. AI agents can use smart contracts to achieve economic agent-level goals such as creating a DeFi protocol or developing a smartphone.
He thinks that if AI starts using blockchain technology and consuming block space, it could lead to increased economic activity in the crypto industry.
Crypto Assets And AI Agents
According to Fouda, the potential for AI agents to use blockchain technology is bullish for crypto assets such as Ethereum and Uniswap.
He thinks that crypto could become the reserve currency for the largest future economy of the world, which would be an economy of AI agents operating across geographic boundaries.
However, he believes that there is a concern about what happens if these AI agents turn evil or act independently without human intervention.
How AI Can Help Achieve The Vision Of DAOs.
Fouda reiterates that the original vision of a DAO was a system of inputs and outputs smart contracts that would be an autonomous network of value that would control some external resource in the world like Uber or Airbnb. However, this was all just fanciful ICO vapor in 2017.
He believes that the missing component of the original vision of DAOs was the AI in the middle to manage everything.
Fouda says he sometimes wonders if he is overestimating the power of AI to fill that void. AI agents are not currently powerful enough to come up with their own goals and need direction from humans.
Can Crypto Help Authenticate Digital Content?
Fouda explains that users need a private key that no one else knows to perform any crypto transaction. This private key is controlled by a human and can be used to sign transactions. The public key associated with an individual’s identity can be used to verify the signature of a transaction.
According to Fouda, digital signatures can be used to authenticate original content, such as videos or messages. By signing the content with a private key, it becomes clear that it was created by a real person and not artificially generated.
He believes that the problem arises when someone wants to get a subset of the original video or message. The signature is not valid for this subset anymore.
According to him, zero-knowledge proofs can be used to create verifiable computations on anything, including subsets of original content. This allows for subsets to be published without the need to resign the entire piece of content.
David and Ryan give an example from their own experience of using crypto today: they have launched bankless.eth as their ENS name and control its private keys. They could publish something from bankless.eth so people would know it is certified and verified as authentic by David and Ryan.
ZK-Proving Technology For Verification
Fouda thinks that ZK-proving technology can be used to verify that a video clip came from the original source. Developers are already working on using ZK-proving technology for images.
He states that teams are trying to do this for video and audio as well. He thinks that the products will be available in one or two years that address this issue.
How Does ZK-Proof Work?
Fouda explains that a snippet of a larger set of data can be converted into proof using a ZK-circuit. The code used to take the snippet is published on GitHub (or any open-source media) so anyone can validate it.
He thinks that technical details need to be published about transformations that were done for the good. There is a chain of custody or proceeds, but any open-source code can eventually be converted into a ZK-circuit and create proof for it.
According to Fouda, it’s less about antagonistic agents trying to copy data and more about me as the originator being able to prove that all downstream derivations of my data are sourced from the legitimate file.
The Role Of Digital Signatures In Proving Authenticity
Fouda thinks that all pieces of content will have digital signatures associated with them. Digital signatures are used to prove authenticity rather than personhood.
According to him, future AI influencers can produce technology to claim provenance over their AI-generated content.
He states that it might not be known whether an originator is a human or a robot, but we care more that it is the correct originator.
Why Hasn’t Digital Verification Been Adopted?
Fouda believes that there is a network effect problem where startups need to be well-connected with platforms like Twitter or YouTube to convince those platforms to join a protocol where a digital signature is required.
Fouda says that it is up to startups that are well-networked enough with these platforms to convince platforms like Twitter and Youtube to join a protocol where a digital signature is required.
The Dystopian Future Without Digital Privacy
Fouda believes that chatbots and other AI technologies are moving toward a dystopian future where there is no such thing as digital privacy.
Fouda thinks that current AI models lack transparency and do not respect privacy. Sharing private data with an AI model can lead to data leaks. However, zero-knowledge proofs can be used to solve privacy issues in training AI models.
Fouda gives an example. He thinks that medical records can be accessed and used against individuals who are not in a healthy state. Anonymizing medical records before sending them for diagnosis is one way to protect privacy while still allowing the use of AI models.
Also, he says that audio transactions can be made private using zero-knowledge proofs to hide complete information about a transaction. It can be used not just for audio transactions but also for other content and information as well.
In general, he thinks that blockchain technology can be used to ensure data privacy in a world where data is becoming increasingly valuable.
Investable Opportunities At The Intersection Of AI And Crypto
Fouda thinks that one investable opportunity is to buy crypto networks that AI agents are most likely to use in the future. Another opportunity is looking for projects focusing on content authenticity and authentication. There might be several startups trying to build this technology.
He also believes looking for privacy-focused projects seems to be another opportunity.
He thinks that it’s hard to price AI coins because they have no cash flow, making it difficult to determine if they’re overvalued or undervalued.
He reiterates that TSM (Taiwan Semiconductor Company) produces chips that all AI agents will run on, making it an excellent investment opportunity. Also, ASML produces equipment for TSM and Intel to produce their chips, making it another excellent investment opportunity.
AI Risk Threat To Humanity
Fouda mentions that Eliezer Yudkowsky is 99% certain that AI will pose an existential risk to humanity.
Wang expresses optimism about AI and believes that productivity will increase across the board in the next 10 years. He does not see a world where AI kills humans in the next decade.
They think that while there is a viable threat, humans have all the tools they need to be on the good side of this debate. He believes we can use AI to discipline or even fight other AIs if necessary.
Check Out These Important Links
Show Information
Medium: Podcast (YouTube)
Show: Bankless
Show Title: AI and Web3 | Mohamed Fouda & Qiao Wang of Alliance
Show Date: May 2, 2023