Scribbler Logo

    You have 4 summaries left

    Empire

    Will AI Agents Use Crypto? | Qiao Wang

    Tue Jul 25 2023
    Crypto MarketAI and BlockchainFounder DynamicsDecentralizing AICrypto HubsDecentralized ComputePrivacy ConcernsDecentralized DatasetsAI Agents in CryptoCrypto Consumer ProductsGPT for Developers

    Description

    This episode covers various topics in the crypto market, including the intersection of AI and blockchain, the current state of the market, building infrastructure and key projects, founder dynamics and market trends, decentralizing AI and crypto hubs, decentralized compute and privacy concerns, decentralized datasets and risks of AI agents, crypto consumer products and financialization, and GPT for developers and market sentiment.

    Insights

    AI agents leveraging blockchain for various purposes

    AI agents can leverage the blockchain permissionlessly to interact with DeFi, play games, and use ETH as native money.

    Decrease in developer activity in the crypto ecosystem

    There are fewer interesting projects in the crypto ecosystem, possibly due to a decline in developer activity.

    Importance of unique insights for successful products

    Unique insights into users are crucial for building a successful product.

    Pivots away from crypto and co-founder breakups

    There has been a significant increase in pivots away from crypto and co-founder breakups in the last few months due to the bear market.

    Decentralization of AI and privacy concerns

    Decentralized AI aims to prevent powerful models from being controlled by one central entity. Privacy is a key concern with centralized AI as personal data is sent to the cloud.

    Shortage of GPUs for training powerful models

    There is a shortage of GPUs for training powerful models, so decentralized AI allows average consumers to contribute their idle GPUs.

    Decentralized datasets for preventing politically biased models

    Decentralized datasets used in training can prevent politically biased models.

    Risks and opportunities of AI agents in crypto

    Permissionless platforms like crypto can be leveraged by AI agents, which raises concerns about their potential actions. Consumer crypto products should consider embracing AI bots and hyper-financialization as interesting opportunities.

    Embracing financialization in crypto consumer products

    Crypto consumer products should embrace speculation and hyper-financialization to attract mainstream users.

    GPT as an important tool for developers

    GPT is the most important use case for developers, making coding faster and easier.

    Chapters

    1. AI and Blockchain in the Crypto Market
    2. Current State of the Crypto Market
    3. Building Infrastructure and Key Projects in Crypto
    4. Founder Dynamics and Market Trends
    5. Decentralizing AI and Crypto Hubs
    6. Decentralizing AI and its Challenges
    7. Decentralized Compute and Privacy Concerns
    8. Decentralized Datasets and Risks of AI Agents
    9. Crypto Consumer Products and Financialization
    10. GPT for Developers and Market Sentiment
    Summary
    Transcript

    AI and Blockchain in the Crypto Market

    00:00 - 07:45

    • AI agents can leverage the blockchain permissionlessly to interact with DeFi, play games, and use ETH as native money.
    • The private primary market in crypto is currently dead, with low venture funding and decreasing valuations.
    • Bitcoin's price at $30,000 feels heavy despite positive news like Bitcoin ETF and XRP victories.
    • Uncertainty remains in the macro market regarding a potential recession or soft landing.
    • Possible reasons for funds not deploying capital include previous investments at high valuations and cautiousness due to falling prices.
    • VCs are currently fearful rather than being contrarian investors.
    • Seed deals are still happening in the $10-15 million range, but later stage deals are dead.
    • There are fewer interesting projects and overall fewer deals in the market.

    Current State of the Crypto Market

    07:16 - 15:10

    • Late stage deals in the crypto market are not happening due to the impact of the bull market peak.
    • There are fewer interesting projects in the crypto ecosystem, possibly due to a decline in developer activity.
    • The number of new developers and active developers in crypto has decreased by 25-50% from the peak.
    • Many developers are pivoting to AI instead of building crypto projects.
    • The current stage of the crypto market is characterized by boredom and lack of attention.
    • The market may be entering a rebirth stage, but there is skepticism and disbelief among investors.
    • The macro environment will play a significant role in determining the future of the crypto market.
    • The battle between bad macro conditions and Bitcoin halving could impact the next bull run.
    • There is uncertainty about whether the regular four-year cycles in crypto will continue or change.
    • Currently, there are more infrastructure projects than consumer-facing products being built in the crypto space.
    • Startups on the consumer side are less inspiring, and there is a desire for more creative and outrageous ideas.

    Building Infrastructure and Key Projects in Crypto

    14:49 - 22:30

    • Nikita Bier, a legend in consumer products, tweeted about using AI to bootstrap critical mass for dating apps
    • Building infrastructure in the crypto space is becoming more popular, with L2 solutions and account abstraction being key areas of focus
    • Solana recently pivoted from being a layer one to a layer two solution on Ethereum
    • Wallet infrastructure around account abstraction is not considered deep tech but rather wrappers around existing technology
    • The big five chains in the market are Ethereum mainnet, Polygon, Solana, Arbitrum, and Optimism
    • It's unclear why some projects choose to build on Ethereum instead of alternatives like Arbitrum or Optimism
    • No one has adopted the same strategy as Cosmos for building L3 solutions
    • Rob Leschner launching Superstate is seen as an encouraging development in the crypto market

    Founder Dynamics and Market Trends

    22:14 - 29:33

    • Crypto founders who have previously launched successful projects are now entering the market for their second or third venture.
    • The current cohort of founders is predominantly made up of crypto natives, compared to previous cohorts which included more web2 engineers and people who have since pivoted away from crypto.
    • There has been a significant increase in pivots away from crypto and co-founder breakups in the last few months due to the bear market.
    • The co-founder relationship is an important factor in a company's success, with conflicts often arising when growth stalls or one founder loses conviction in crypto.
    • The Tensor founders are seen as a winning team due to their impressive achievements and strong work ethic.
    • When assessing founders, it is important to consider their unique insights about the product and go-to-market strategy.

    Decentralizing AI and Crypto Hubs

    29:06 - 37:21

    • Resilience is the most important characteristic of every successful founder.
    • Unique insights into users are crucial for building a successful product.
    • Tensor and Blur had unique insights about their users in the NFT trading space.
    • Stepan overcame language and cultural barriers to succeed.
    • There is renewed interest in Korea, Japan, Singapore, and Hong Kong as crypto hubs.
    • The US is gaining positive perception with recent Bitcoin ETF and XRP news.
    • Latin America has a growing number of crypto founders building for the region.
    • Dubai and Hong Kong are seeing an increase in crypto founders.

    Decentralizing AI and its Challenges

    37:05 - 44:09

    • There are different buckets for decentralizing AI, including customer support chat bots, decentralized compute, training of AI, and decentralizing the front end of AI inference.
    • Helping developers become more productive is the most important use case of GPT.
    • GPT can architect and write code, resulting in a 10x improvement in speed.
    • The code interpreter of chat GPT can also architect code, leading to another 10x improvement in speed.
    • GPT can help beginners learn Python and build projects quickly.
    • Decentralizing AI is important to prevent control by a single entity.
    • OpenAI recently dumbed down their models for consumers while keeping powerful models internally.

    Decentralized Compute and Privacy Concerns

    50:26 - 58:51

    • Decentralized AI aims to prevent powerful models from being controlled by one central entity.
    • Privacy is a key concern with centralized AI as personal data is sent to the cloud.
    • Edge computing, running open source models on local devices, can solve the decentralization and privacy problems.
    • Consumer-grade hardware may not be powerful enough for the most advanced models, but efficient open source models exist.
    • There is a shortage of GPUs for training powerful models, so decentralized AI allows average consumers to contribute their idle GPUs.
    • Decentralized compute has the disadvantage of being slower due to communication overhead between different computing devices.
    • The counter argument is that consumer-grade hardware may not be powerful enough for training large models, but it could be used for smaller models or fine-tuning.
    • If decentralized compute works, it could address the GPU shortage problem and have a significant impact.

    Decentralized Datasets and Risks of AI Agents

    58:32 - 1:06:25

    • Decentralized datasets used in training can prevent politically biased models.
    • The internet leans more liberal than conservative, which affects the political bias of social networks and the data used by OpenAI to train their models.
    • Using tokens to incentivize politically neutral or diverse datasets is challenging.
    • Crypto can be used to digitally assign content and fight deepfakes through digital signatures on the Ethereum blockchain.
    • Permissionless platforms like crypto can be leveraged by AI agents, which raises concerns about their potential actions.
    • AI agents with permissionless programming capabilities could pose risks if they have access to unstoppable forms of money like ETH.
    • Giving AI write access to the blockchain opens up new possibilities but also challenges in verifying human identity online.
    • Embracing AI bots as users and building billion-dollar companies around them is an alternative approach.
    • Consumer crypto products should consider embracing AI bots and hyper-financialization as interesting opportunities.

    Crypto Consumer Products and Financialization

    1:06:03 - 1:14:06

    • Crypto consumer products should embrace speculation and hyper-financialization to attract mainstream users.
    • Crypto social networks should encourage de-gen behavior and outrageous interactions, similar to DeFi Summer on Twitter.
    • There may be a rise in social tokens and financialization of tweets, where users can bet on accounts and early engagement.
    • A Chinese product allowed consumers to buy NFTs representing time with celebrities, speculating on their value.
    • AI agents are autonomous entities that can perform tasks and hire other AI agents, but the hype around them has died down.
    • The optimism lies in the ability of GPT-powered technology to enable faster coding and empower developers as 10x performers.

    GPT for Developers and Market Sentiment

    1:13:36 - 1:16:43

    • GPT is the most important use case for developers, making coding faster and easier
    • Startups can launch MVPs with GPT, reducing the cost of product launches
    • GPT benefits people with good ideas but limited coding ability
    • The speaker is optimistic about AI in general, but doesn't have a strong thesis yet
    • Crypto market is less stressful now as committed investors remain
    • The speaker was excited about the ripple SCC incident as it could bring an end to the bear market
    1