Open AI – One direction.

Open AI’s future is to be a subsidiary of Microsoft.

• Open AI has no intention of going public not because the CEO wants to maintain control over technology but because I think that Microsoft intends to acquire OpenAI and will not let anything else happen to it.
• Sam Altman is currently on a global charm offensive to bring regulators and governments into the loop in terms of the benefits and threats of AI and also to influence regulation of AI when it finally comes.
• When questioned on the subject of long-term ownership of the company, he stated that he had no intention of listing the company saying “When we develop superintelligence, we’re likely to make some decisions that public market investors would view very strangely”.
• What he is referring to here is the possibility that while a super-intelligent AI could make the company a lot of money, the company could be forced to shut it down for safety reasons.
• This brings us right back to the Armageddon debate (see here) where my view is that while the machines are very good at creating the illusion of being able to reason, they remain nothing more than very sophisticated statistical pattern recognition systems.
• Hence, I think that Sam Altman is talking his own book in terms of creating the impression that if superintelligence is going to be created, it will be OpenAI that gets there first.
• I think that the reality is that we are as far now from superintelligence as we were 10 years ago largely because all of these generative AIs are based on the same technological foundation, and it is not until that foundation is changed that the weaknesses of hallucination and errors will be properly dealt with.
• Despite these weaknesses, generative AI based on large language models (LLMs) are extremely good at understanding human language and ingesting all sorts of data that can be recalled with a single request.
• The responses need to be fact-checked, but the time saving by not having to search for this data is very substantial and offers the possibility of a substantial boost to white-collar productivity.
• Not missing a trick, Microsoft is already embedding OpenAI-derived LLMs in all of its products which should greatly improve the execrable search and help functions of Microsoft Office as well as all of its other products.
• However, this also means that Microsoft has created a dependency of its core products on OpenAI technology, and I think that there is no way that Microsoft can allow this dependency to slip from its control.
• With its most recent investment in OpenAI, Microsoft now owns 49% which with no single investor controlling the other 51%, hands Microsoft effective control of OpenAI.
• Hence, I doubt that any corporate action that alters the control of this company is going to pass without Microsoft’s agreement.
• Any change that lessens Microsoft’s influence is going to increase its technology risk and therefore, I think is very unlikely to ever see the light of day.
• What is far more likely is that Microsoft buys the remaining 51% and OpenAI becomes the AI department inside Microsoft, but I suspect that the current status quo is probably the best option.
• This will allow OpenAI employees to enjoy the illusion of independence, making them less likely to leave and go elsewhere while at the same time, Microsoft has control of the company and its roadmap.
• This does have the downside of potential competitors having access to OpenAI technology, but should this become an issue, I suspect that Microsoft would move fairly quickly to formally acquire the company.
• This is why RFM Research considers OpenAI and Microsoft to be one company when it comes to positioning within the global AI ecosystem and this is unlikely to change.
• This catapults Microsoft from being a laggard in AI to being a leader but with the current valuation of the company, I still do not want to own it.
• I would continue to look elsewhere.