17 Jul How Can Bitcoin Improve Access to AI?

On July 6th, Lightning Labs announced a set of developer tools to enable  Lightning and AI developers to integrate the Lightning Network for micropayments directly within Large Language Models (LLM). This is intended to make payments for Artificial Intelligence a lot more easy. Integrating Lightning payments mean that fast, quick payments to AI services is possible, fostering much greater inclusion into this exciting technology, and allow for AI agents and LLMs to make micropayments and device to device transactions.

The Benefits of Integrating Lightning with Artificial Intelligence

AI will need to utilise some kind of native payment method, in order to autonomously access data and meet the demands of paying for the cost of electricity and computer power that sustains it. It is only with this capability that can truly operate as an independent entity online. 

Many people think that Bitcoin fits the bill because payment methods for AI needs to be accessible at all times, must be digital, and must have a transparent set of rules that cannot easily be changed. It would also need to be automated without the need for reliance on human interaction. This could then be leveraged by AI models and platforms offering AI-focused services.

Another major benefit for AI-based systems to use Bitcoin for payments is its easy onboarding. As it currently stands, LLM users pay for access to models like ChatGPT by using a credit card to purchase credits/tokens which allow the user to access the service. 

For an AI agent which only exists online, this is untenable, as they exist solely in a digital environment, and are not considered to be entities which can be registered with any nation’s economic system. This means AI agents do not have the ability to access payment systems based on current fiat infrastructure. 

Bitcoin fixes this by making it easy for any non-human digital entity to simply onboard and begin using the system for autonomous censorship-resistant payments, at anytime. Another benefit of democratising the way payments are made to access AI platforms and services, is that AI becomes more open and accessible.

Currently, AI models have high costs, and if they become popular, developers must either pay for these costs out of pocket, or find their service either rate limited, or unsustainable to operate. In order to scale these kinds of applications globally, costs need to be offloaded via a globally accessible low cost payment system, which anyone can use to pay for AI services, as they need.

Paying for LLM Access with Satoshis via Lightning Network

This breakthrough took place just last week, when Lightning Labs developers Olaoluwa Osuntokun and Michael Levin announced the Lightning Labs release of a new set of developer tools which will make it easy for Lightning network developers and LLMs like ChatGPT to easily and seamlessly integrate Lightning Network.

Users who want to access LLMs can now pay for that access using Lightning BTC, using the L402 Protocol, as well as the popular AI development library Langchain, using the newly developed LangchainBitcoin suite of tools created by Lightning Labs. These tools make it simple for developers to create agents that can hold a Bitcoin balance (on-chain and Lightning), send/receive Bitcoin on Lightning, and also interact with a LND node.

They also include a Python wrapper for LangChain LLM chain (APIChain). This wrapper makes it possible for developers to create AI agents that can transparently access L402 payment-metered APIs. This means users of LLMs now can pay for their LLM services and access with Bitcoin over Lightning, making it easier for those without access to financial infrastructure to access LLMs.

Lightning Labs also announced the latest release of Aperture, which is a reverse-proxy server for LLM APIs, which can turn any LLM API into a pay per use resource behind a L402 “paywall” which can be accessed via micropayments in Satoshis over the Lightning Network.

Lightning Labs sees potential in the L402 standard, as being a new addition to the current 402 payment required response code, which exists in HTTP. L402 would make it possible to charge for services and authenticate users in distributed networks.

By adding micropayments via Lightning Network, access to AI services becomes dramatically improved, for both human beings looking to use the platforms, as well as AI agents, bots, or machine to machine payments can also be made. This solves many of the issues surrounding access to the best AI models.

This seems a great way to ensure that AI is made more accessible with Bitcoin technology at its core, facilitating payments.