A new ChatGPT-powered bot named Satoshi will soon help crypto traders

Artificial intelligence may soon make waves in the cryptocurrency industry, but maybe not in the way you think. Instead of merging the two technologies, Chicago-based prime broker FalconX plans to put a chatbot in the co-pilot’s seat for investors.

Using technology created by OpenAI, whose ChatGPT program is helping companies like Microsoft rewire online search, FalconX clients will be able to ask questions like “What are the three biggest differences between two blockchain platforms?” or “What is the delta between the Sharpe ratio for a Bitcoin base strategy or a Bitcoin Hold strategy over a two-week period?” to a bot called Satoshi.

Satoshi – named after Bitcoin’s alleged founder Satoshi Namakmoto – will also be able to generate investment ideas for users based on their historical trading activity, portfolios and interests, says FalconX CEO Raghu Yarlagadda. Although the technology is very much in its early stages — the current prototype primarily allows users to get customized news summaries similar to traditional ChatGPT responses to user requests, and trading backtesting has only been available for a few weeks — progress is likely to come quickly.

FalconX is a natural bridge to bring OpenAI’s technology into crypto. Prathab Murugesan, the company’s chief engineering officer spent 2.5 years at Google working to bring machine learning technology, a process by which computers are trained to recognize patterns and predict actions, into products like Gmail and Google docs.

Yarlagadda, who himself was recruited to Google in 2014 by current CEO Sundar Pichai, spent three years finding ways to transmit high-quality video over the Internet. “Sundar said Google wanted to be a machine learning company,” says Yarlagadda. “This was a complete and radical departure from the norm, because machine learning had never been operationalized at a scale where you could freely provide access to all these huge products.”

This machine learning approach was built into FalconX from the start in 2018 because it was the only way to get a clear picture of the market. Therefore, initial use was focused on mundane tasks such as cleaning up market data to filter out fake volume and wash trades, notorious problems in crypto.

However, machine learning algorithms cannot tell traders what to do next. Yarlagadda says that you can train a computer model to recognize images of cats by sharing an image library with the program. It can become very adept at distinguishing cats from dogs and even identifying different types of cats, but no matter how many pictures it sees, it can’t draw one. Taking this analogy one step further, even if this model were trained to recognize dozens of types of animals, it would not be able to perform a task such as predicting how a platypus might evolve in 1000 years in a scenario where the sea temperature rises by 2 degrees.

In trading, this analogy is the same as asking a traditional algorithmic trading model, which probably took a team of developers to code, to build a strategy for circumstances that have yet to happen, and perhaps adapt it to a specific portfolio.

Large language models (LLMs), such as those used by OpenAI and Google, can take this machine learning foundation and build what is known as generative artificial intelligence on top of it by enabling these platforms to take in volumes of unfiltered and imprecise data and answer any question. Yarlagadda says the company had been working with Satoshi for more than nine months, before the ChatGPT hype. However, it ran into roadblocks until OpenAI cleared the way.

“We didn’t have a breakthrough in the first five to six months because we were mostly relying on machine learning, and while we were aware of what OpenAI was doing at the time, it wasn’t until we used ChatGPT that we had the tools to solve this problem on a large scale, says Yarlagadda.

Today, FalconX uses OpenAI’s API stack and infrastructure to test and build Satoshi and frequently interacts with the firm’s account management and engineering teams. It’s all part of the company’s effort to make Chat GPT’s LLM a base layer for a wide variety of applications. FalconX, for its part, says it will integrate other LLMs beyond those offered by OpenAI, such as Google’s Bard. OpenAI declined to comment on its partnership with FalconX when reached for comment.

That’s not to deny the possibility, but the real question is whether it will work. Yarlagadda says that in crypto, 90% of all legitimate trading is done by 10% of traders, most using algorithmic models. These are usually large firms that have the resources to hire teams capable of building the tools. These businesses can implement a variety of approaches from market neutral strategies such as market making and arbitrage (exploiting asset price differences across trading platforms), to quantitative long/short approaches. According to PriceWaterhouseCoopers 2022 report on the crypto hedge fund industry, these two approaches make up about 55% of the industry.

That still leaves 45% of hedge funds using a discretionary approach to at least some of their trading, and the percentage of such trades goes up when you consider other categories such as venture capital funds, family offices, brokers and retail traders. Satoshi is designed to help these groups compete on a level playing field with the large quantitative operations.

It will do this in three ways. First Satoshi can map all relevant news and information across traditional and social media to provide briefings targeted to clients’ interests or holdings that can answer questions like “How has my portfolio done in the last 24 hours?” or “Who are the top three social media influencers posting about a particular asset and what are they saying?”

Then the user can test trading strategies by asking service questions like, “How much will it cost me to put on a $1000 short position on bitcoin?” or “What is the best strategy to buy $5M worth of Ether without paying more than 25bps?” Finally, Satoshi will eventually have buy/sell buttons built right into the platform so that the user manifests these strategies right away.

Much of this is ambitious. Satoshi remains in testing and is not yet integrated with necessary platforms such as exchange order books and cannot produce trading charts and other necessary tools for professional users.

A potentially disastrous blind spot is likely to be the inability to assess levels of leverage or financial soundness for trading with crypto counterparties. An important lesson from the collapse of major crypto players like BlockFi, Three Arrows Capital, Genesis Global Trading and FTX is that so many of them were indebted to each other and took on huge amounts of leverage to maximize gains in what was believed to be. an upcoming crypto superbike.

There are also unresolved issues with generative AI, including privacy concerns. Also, the technology causes hallucinations, or at least their virtual equivalent. The phenomenon that occurs when AI platforms give wrong answers in a matter-of-fact and yet convincing way can be extremely dangerous if they give incorrect trading strategies.

ChatGPT is notorious for some very damaging hallucinations, such as the time it falsely claimed that a George Washington University law professor was accused of sexual harassment, and even made up a Washington Post story to support the claim.. The consequences could be even more critical for users should trade large amounts of money based on hallucinations provided by Satoshi.

“Because we specialize in this particular use case, our goal over the next year is to reduce the variance by 10 times,” says Yarlagadda. However, he has yet to even have a failure rate given the newness of the product.