Your browser is not supported. Please update it.

25 July 2019

Microsoft disguises exclusive Azure deal as $1bn OpenAI investment

Microsoft has announced a $1bn investment into OpenAI, the non-profit AI research project that counts Elon Musk among its founders, and signed a completely totally separate honest-to-goodness “exclusive computing partnership” that will see Microsoft become the sole provider of computing resources for OpenAI.

But because OpenAI is a non-profit, Microsoft can’t really just give the firm $1 billion, and so, the investment looks like a somewhat roundabout way to pay for the Azure instances that OpenAI is going to be using. There’s still an element of philanthropy here, but Microsoft wants to get in at the forefront of this AI research, which should have some beneficial effect on the combined Azure platform – which now represents around a third of Microsoft’s quarterly revenue.

To this end, OpenAI is going to be licensing some of its technologies to Microsoft, which are then going to be commercialized in Azure. Timelines and deliverables haven’t been discussed with regard to this facet, but it seems quite clear that Microsoft views this as a business opportunity and not some charitable endeavor.

One does wonder what happened to the $1 billion announced at formation, and whether an extra $1 billion from Microsoft is going to make all that much difference if OpenAI ran out of cash in about 3.5 years. Sure, it might keep things ticking along, but the jump from where we are now to an Artificial General Intelligence (AGI) is going to require rather a lot more cash to fuel the run-rate. At launch, it said it didn’t plan on spending the $1 billion within a few years, but now, speaking to Tech Crunch, CTO Greg Brockman said it plans to spend the new $1 billion within five years – a cash investment that it can call as needed.

OpenAI is perhaps best known for building a system to beat the best human players at video games, as well as a Fake News generator, the code for which was kept hidden due to fears for its misuse. Founded in December 2015, the firm was launched with around $1 billion in funding that came from Musk, Peter Thiel, AWS, Infosys, and YC Research. Ostensibly, its goal is to create a safe AGI, such that the AGI can’t act in a manner that will harm humans – a concern Musk has been pretty vocal about.

But it seems that the money ran out, or came close to at least, as OpenAI announced that it needed more money to continue operations – and then set up a for-profit firm to help alleviate these cash-flow problems. This is notable, because much of the founding discussions focused on how OpenAI would be removed from commercial concerns, and so could develop an AGI in a proper and safe fashion – with the subtext being that the business that created the first AGI would be doing so to make money, and therefore there would be a higher chance that things could get all Skynet very quickly.

In February 2018, Musk left the board of OpenAI, using Twitter to cite disagreements over the company’s direction as the cause, as well as his commitments to Tesla and SpaceX. “Also, Tesla was competing for some of same people as OpenAI and I didn’t agree with some of what OpenAI team wanted to do. Add that all up and it was just better to part ways on good terms.”

OpenAI’s AGI is a controversial goal. This would be a system that could master any task given to it, outperforming any human in its capabilities, and effectively being superhuman in its capabilities. However, the industry consensus on its achievability is that such an AGI, if it is at all possible, is a few decades away at least, with a survey citing only half of industry respondents thinking it was doable by 2099.

As the firm puts it in the Microsoft announcement, the project requires a lot of capital for computational power, and that the most obvious way to secure that would be via productization. It wants to avoid this, as it would mean changing its focus.

“We believe that the creation of beneficial AGI will be the most important technological development in human history, with the potential to shape the trajectory of humanity. We have a hard technical path in front of us, requiring a unified software engineering and AI research effort of massive computational scale, but technical success alone is not enough. To accomplish our mission of ensuring that AGI (whether built by us or not) benefits all of humanity, we’ll need to ensure that AGI is deployed safely and securely; that society is well-prepared for its implications; and that its economic upside is widely shared. If we achieve this mission, we will have actualized Microsoft and OpenAI’s shared value of empowering everyone,” reads a statement.