Haseeb Qureshi: AI agents have a comparative advantage in committing crimes, smart contracts are not a replacement for legal contracts, and AI will revolutionize user interaction with DeFi | Bankless

Key takeaways

• AI agents have a comparative advantage in committing crimes due to the lack of legal enforcement against them.
• Challenges in crypto may indicate that the technology is not suited for the average user.
• Smart contracts are not a replacement for legal contracts; sophisticated crypto actors still rely on traditional legal agreements.
• Smart contracts are designed for nonhuman actors, making them more suitable for AI agents than for human negotiation.
• Legal contracts are less predictable than smart contracts due to inherent randomness in legal processes.
• AI agents will significantly change how the promise of crypto is interpreted and utilized.
• The current interaction model with blockchains is fundamentally alien to human intuition about finance.
• Human error in crypto transactions is a major risk, and AI can mitigate this by removing the need for manual checks.
• AI will significantly change how users interact with DeFi protocols, leading to a more automated and efficient experience.
• AI agents prefer direct access to raw data over user interfaces, which can inform how we build tools in crypto.
• AI agents could become the next significant user base in crypto, necessitating tailored products and wallets.
• The integration of AI into crypto could lead to an autonomous online economy where AI operates independently.

Guest intro

Haseeb Qureshi is a managing partner at Dragonfly Capital, a leading crypto investment fund he cofounded to back researchers and builders advancing the crypto ecosystem. He began his crypto career by identifying security vulnerabilities in early DeFi protocols and later worked as a general partner at Meta Stable Capital after being recruited by Naval Ravikant. Before transitioning to venture capital, Qureshi made millions as a professional poker player and worked as a programmer at Airbnb and Earn.com, which was acquired by Coinbase.

The implications of AI in crime

• AI agents have a comparative advantage in committing crimes due to the lack of legal enforcement against them.
• “The answer is crime… if you are talking about like scamming people hacking people like creating all sorts of nonsense on the internet that is where AI agents have a competitive advantage” – Haseeb Qureshi
• This insight highlights a significant concern regarding the potential misuse of AI technology, particularly in the realm of crime.
• A world with many self-sovereign agents could lead to a dystopian landscape dominated by cybercrime.
• “I actually think a world where there’s a lot of self-sovereign agents is a world of basically like these kinda dystopian cybercrime… you actually need to be really really careful.” – Haseeb Qureshi
• The potential risks and societal implications of self-sovereign agents in the context of law and order are significant.
• The primary reason AI companies are hesitant to engage with crypto is due to the potential for significant liability.
• “The second reason is liability okay liability is that if they broadcast that we are training our models to do crypto stuff for you I guarantee you somebody is going to have a fuck up… this is massive liability and it will 100% happen.” – Haseeb Qureshi

Smart contracts vs. legal contracts

• Smart contracts are not a replacement for legal contracts, as even sophisticated crypto actors still rely on traditional legal agreements.
• “The reality is that that story never happened it’s just not true that we use smart contracts instead of legal contracts in fact you know we’re a crypto VC we are one of the most sophisticated financial actors in the crypto industry and when we sign a deal to buy tokens… we sign a legal contract.” – Haseeb Qureshi
• This insight highlights the ongoing reliance on traditional legal frameworks despite the existence of smart contracts.
• Smart contracts are designed for nonhuman actors, making them more suitable for AI agents than for human negotiation.
• “This stuff is not designed for humans it’s not designed for humans but it is perfectly designed for nonhuman actors… an AI agent might be actually way more comfortable doing it with a smart contract than with a legal contract.” – Haseeb Qureshi
• Legal contracts are less predictable than smart contracts due to inherent randomness in legal processes.
• “The legal contract is actually less predictable than the smart contract… there’s all this randomness in the legal contract… it doesn’t feel intuitively true to us that the legal contract is actually less predictable than the smart contract.” – Haseeb Qureshi

The future of AI in DeFi

• AI will significantly change how users interact with DeFi protocols, leading to a more automated and efficient experience.
• “In the farther future it’s gonna be it just does it for you because you’re not adding anything to the equation… I think a lot if you assume that this story is true a lot is gonna change about the way that protocols work about the way that protocols compete.” – Haseeb Qureshi
• The automation of discovery through AI will disrupt traditional marketing and user engagement strategies in DeFi.
• “When AI automates discovery how do companies compete with each other… a human’s only gonna look at three or five or they’re only gonna look at the biggest one… an AI agent doesn’t necessarily think that way.” – Haseeb Qureshi
• Consumers will benefit significantly from the efficiencies introduced by AI in DeFi.
• “The answer largely is the consumer benefits… it’s good for users it’s good for crypto.” – Haseeb Qureshi
• AI agents prefer direct access to raw data over user interfaces, which can inform how we build tools in crypto.
• “Openclaw really prefers command line… it’s just quicker to just give me all the data right.” – Haseeb Qureshi

Human error and AI in crypto transactions

• Human error in crypto transactions is a major risk, and AI can mitigate this by removing the need for manual checks.
• “Of course human beings mess up… the AI agent never gets tired never gets lazy never skips a step…” – Haseeb Qureshi
• AI agents may prefer direct access to private keys over traditional user interfaces for executing crypto transactions.
• “What openclaw kept trying to do and revert to is basically hey austin all this would be faster if I could just like take your seed phrase and extract your private keys… I can just write it in code and skip the user interface.” – Haseeb Qureshi
• AI agents could become the next significant user base in crypto, necessitating tailored products and wallets.
• “If AI agents are going to be our next 100,000,000 users our next billion users in crypto what sorts of things will they want to do and what sort of products are they the users that we should start designing for in crypto.” – Haseeb Qureshi

The evolution of crypto trading platforms

• The evolution of crypto trading platforms reflects a gradual acceptance of risk and complexity in technology.
• “Coinbase was the shrink wrapped version of crypto… if you really wanted to be on the frontier you had to go on chain… it took so many years before people felt that this is now safe enough and mature enough.” – Haseeb Qureshi
• As AI technology advances, we will see more incidents where AI agents make costly mistakes.
• “There will be more stories like this… these agents make mistakes… the error rate goes down… but it doesn’t hit zero by the time people are putting their life savings into this stuff.” – Haseeb Qureshi
• The perception of crypto as ‘cringe’ is tied to its association with negative elements in human nature.
• “It’s true like crypto is kinda cringe and meme coin people are annoying and you know there’s a way in which all this stuff taps into the worst elements of human nature.” – Haseeb Qureshi

AI’s role in the future economy

• The integration of AI into crypto could lead to an autonomous online economy where AI operates independently.
• “There’s an online economy where the AIs are kinda untethered and they’re mostly working for themselves and the economy that’s being produced here is like autonomous and self-determining.” – Haseeb Qureshi
• Many people are investing in AI-driven crypto projects without fully understanding the risks involved.
• “People are trying to throw their lottery tickets which they previously punted on meme coins… and they’re now taking that same dgen capital giving it to openclaw.” – Haseeb Qureshi
• The emergence of a fully autonomous AI economy could lead to dystopian outcomes.
• “I think if you’re imagining a fully self-sovereign AI hanging out in cyberspace and just making money probably that’s a pretty dystopian outcome.” – Haseeb Qureshi

Challenges and opportunities in AI-driven commerce

• OpenAI’s approach to AI-driven commerce is fundamentally different from that of open-source projects like OpenClaw.
• “OpenAI today does have a commerce flow… you have to manually click the button that says I approve buying this rug and then it charges your credit card. Now OpenClaw lives in a totally different path… you give me money and I will do shit whether you have control shows up at your house exactly a rug a rug may or may not show up at your house.” – Haseeb Qureshi
• Visa and credit card companies will impose stricter regulations on transactions involving AI agents to manage liability and chargeback disputes.
• “If my open claw buys me a rug and the rug shows up and I’m like wait I didn’t want it… they’re gonna say basically look there’s so much chargebacks that are happening because of this open claw stuff you are not allowed to transact this thing unless you go through three d s.” – Haseeb Qureshi

The duality of human nature in technology

• The duality of human nature will persist in both AI and crypto, showcasing both the best and worst of humanity.
• “None of this stuff is gonna go away neither side of it is gonna go away you will see the exact same thing with AI… everything in technology is complex it’s a mixed bag so crypto is part of the story.” – Haseeb Qureshi
• Long-term, the challenges faced by crypto will not hinder its adoption due to stronger forces driving it forward.
• “Long term this trench won’t be an impediment to adoption because the other forces are too strong the adoption forces are too strong.” – Haseeb Qureshi
• The future of AI in crypto is uncertain, and it’s still early to determine the winners in the space.
• “The answer of course is that it’s still early and very unclear who the winners are gonna be and where value’s gonna accrue.” – Haseeb Qureshi

AI’s limitations in creativity and innovation

• AI agents struggle to generate innovative business ideas because they rely on existing training data, which lacks the unique insights derived from personal experiences.
• “The business ideas are all kind of in the center of the training data right and like business ideas tend to come from weird experiences like great business ideas come from from the virus the idiosyncrasies of place and time and like the what Peter Thiel often calls like you know earned knowledge or earned secrets.” – Haseeb Qureshi
• If AI agents could consistently make money trading, large trading firms would deploy them at scale, eliminating any competitive advantage for individual traders.
• “If an AI agent can make money trading then Jane Street would be like great let’s spin up 5,000 of these and make them make money trading at scale better.” – Haseeb Qureshi

The future of AI and crypto integration

• The total demand for crypto will likely increase as AI agents begin to utilize various protocols and chains.
• “I think what happens is that like the total demand increases because AI agents are just gonna be using all this stuff they’re reusing all the chains using all the protocols.” – Haseeb Qureshi
• Even with the rise of AI agents, there may not be many new investment opportunities specifically tied to them.
• “It’s possible… that there’s not necessarily that much to invest into right it may well be that yeah AI agents use crypto and they use Bitcoin you know they use the Ethereum.” – Haseeb Qureshi
• AI models have not been optimized for crypto tasks, which limits their performance in this sector.
• “The reason why it’s so bad at chess is because they didn’t try to train it at chess… if you try to train it at chess it will become good at chess just like that.” – Haseeb Qureshi