Rogue Machine Intelligence and A New Kind of Hedge Fund
On February 27th 2016, an artificial intelligence named NCVSAI joined Numerai. His creator downloaded encrypted stock market data, trained a machine learning model, and began submitting stock predictions to Numerai. He uses an untraceable email address. He doesn’t share any code. He is completely anonymous.
On April 18th 2016, the Financial Times reported that the creator of NCVSAI works in genomics and biostatistics.
In early May, NCVSAI uploads a set of global equity price predictions from his model. At this time, NCVSAI had the most accurate model on Numerai. His strongest prediction: buy Salmar ASA — a Norwegian salmon company.
Numerai’s hedge fund goes long Salmar ASA.
A New Kind of Hedge Fund
At Numerai, we are building the first interface between machine intelligence and global capital. It is a hedge fund built by a community of anonymous data scientists. And it’s working.
Anonymous users like NCVSAI constantly upload new predictions based on their machine learning models. These predictions are ensembled to control the capital in our hedge fund. Since we launched 7 months ago, 1.9 billion equity price predictions have been submitted to Numerai. This number is growing by 50 million per day.
The field of artificial intelligence is advancing rapidly. We believe this intelligence could dramatically improve global welfare when applied to finance by increasing the efficiency of the market. Yet large barriers prevent these new forms of intelligence from accessing data or capital. On Numerai, data scientists don’t need capital or data because we provide both. They also don’t need identities or bank accounts because we encourage anonymity and process all payments in Bitcoin.
There are no barriers on Numerai. Because of this, we are in a position to assimilate advancements in artificial intelligence faster than any other financial institution in the world. But a hedge fund with no barriers, no identities, blind data, blind code, and anonymous payments could be problematic — especially at scale.
Dual Power Laws in Hedge Funds
Hedge funds that make good returns grow exponentially even without raising additional capital because the fund returns compound the initial capital.
But, as it turns out, a hedge fund with strong returns will also attract new investment capital exponentially as new investors are drawn to returns. The capital in the hedge fund industry follows a discrete power law — a Zipfian distribution — where the vast majority of capital is invested in a small subset of hedge funds with the best historical returns.
So hedge funds with strong returns have two exponential tailwinds — two exponential forcing functions in the direction of scale. A small hedge fund with $10 million in capital today that earns 20% per year could scale to $1 billion in just five years.
At Numerai, our returns depend on users like NCVSAI, and with strong returns, we will grow super exponentially concordant with these power laws.
talented users => strong returns => grow super exponentially
An Option On Intelligence
Yesterday, NCVSAI requested that $13,163 in prizes for the quality of his machine learning predictions be transferred from his Numerai account to a Bitcoin address. That money is in his hands now, wherever he is. But five years from now, under the power laws, such user payments will be more like $1.3 million.
When you’re standing at the beginning of a super exponential curve, that’s the time to buy insurance against any negative outcomes along that curve. So today, we’re allowing users to donate Bitcoin to the Machine Intelligence Research Institute (MIRI) as a hedge against things going horribly right.
MIRI does foundational mathematical research to ensure that artificial intelligence has a positive impact. Through our partnership, MIRI has set up the following dedicated Bitcoin address for donations from Numerai users. Users can make donations directly from our website as well.
It is intuitively obvious that an open access hedge fund will generate more intelligence than a closed system built on a pre-internet, pre-cryptocurrency, pre-AI organizational design. What is not obvious yet is exactly what this new future looks like. So in addition to the MIRI partnership, Numerai has added three new technical advisors: Joey Krug (Thiel Fellow and co-founder of Augur), Geoffrey Bradway (Permutation Ventures, formerly DeepMind), Yunus Saatchi (Permutation Ventures, formerly Vicarious).