OpenAI – Implosion Incoming

The chances of collapse have increased materially.

• The abandonment of the plan to become a public benefit corporation means that the current unstable and conflicted corporate structure will persist, greatly increasing the risk of repeating the almost lethal implosion of 2023.
• OpenAI’s rivals, and particularly Google, will benefit as OpenAI will be less inclined to compete against them and will, instead, focus on pursuing a dream (AGI) that I think it will not achieve with its current technology.
• OpenAI has abandoned its plan to become a public benefit corporation (PBC) and will remain as a non-profit with a for-profit division, which I think greatly increases the risk of both a complete collapse of the company that ends its life in its current form as well as a bitter legal fight with Microsoft.
• This has come about as the result of a backlash from “civil leaders”, the attorney generals of Delaware and California, the lawsuit from Mr Musk and general negative commentary from AI academia.
• A PBC is pretty rare structure, but a good example is Black Rifle Coffee which is a coffee roasting company that also supports US veterans, law enforcement and first responders as its public benefit.
• OpenAI’s, of course, was to “ensure that artificial general intelligence benefits all of humanity”, but unlike Black Rifle Coffee, many of OpenAI’s actions have been in direct contradiction to this statement.
• This is because these are the actions of a for-profit entity that needs to maintain as much of a lead over its rivals as it possibly can as well as create barriers to entry into its market.
• Furthermore, with SoftBank leading a monster $40bn investment round at a post-money valuation of $300bn, Mr Son will be expecting to see OpenAI become a multi-trillion-dollar company meaning increasing pressure to make money.
• The nature of the conflict of interest is simple but impossible to reconcile without OpenAI becoming a for-profit company.
• On the one hand there is the governing body that controls OpenAI that has a mission to deliver artificial general intelligence (AGI) for all of mankind which means giving the IP away for free.
• On the other, there is the very large business opportunity that exists in selling AI services and technology to companies and potentially replacing Google Search for consumers.
• There is every sign that the bigger OpenAI becomes, and the more revenue it earns, the greater the conflict of interest between the two will become.
• These two sides are irreconcilable when it comes to how the company operates and, as a result of the growing conflict, this is a bomb with an escalating probability of going off.
• Furthermore, the scene is set for a bitter legal dispute with Microsoft, the risk of which would have been largely dispelled by the move to a PBC.
• This is because when AGI is reached, Microsoft’s participation in the value upside in the company is capped.
• The problem here is that it is the governing non-profit board of OpenAI that decides when AGI has been reached, which is yet another large conflict of interest.
• Consequently, I think that the risk of a terminal event which causes the company to collapse and its employees and IP to be absorbed or split up between one or more of its rivals, has increased materially.
• The main beneficiary here is Google because OpenAI has parked its tanks on Google’s lawn and for some time, has been signalling its intention to come after Google’s business.
• There has been no sign whatsoever that consumers have any intention of switching from Search to generative AI when looking for information, but for more complex enquiries, there is no doubt that it is a better service.
• Should the complexity of search enquiries increase as consumers realise that AI can now deal with them effectively, then we have arrived at a dislocation point in the market.
• It is always at these junctures that there is potential for market leadership to change.
• This is why at I/O in a couple of weeks, we are likely to see Google focus almost entirely on Gemini and making it so good that it can migrate all its search traffic to Gemini without its users thinking of trying ChatGPT.
• Thanks to greater internal conflicts of interest and decreased incentives for profit, OpenAI will now be less incentivised to aggressively come after Google’s business.
• Furthermore, any impact that it has on Google will increase the risk of a fatal implosion, which would benefit Google greatly.
• I have been increasingly concerned with the long-term (we haven’t seen it yet) impact of OpenAI on Google’s search business, and becoming a PBC would have increased my concerns still further.
• This is almost a get out of jail free card for Google.
• My view on OpenAI’s valuation has, until now been relatively simple in that if it invents AGI and monetises it, then it is worth many trillions but if it fails, then it is worth the present value of the net cash flow that it can generate from selling its services to companies and consumers.
• Given that it already has something like 400m users but is still burning billions of dollars every quarter and will now remain a non-profit permanently, I think that net cash flow to shareholders could easily be zero or less than zero.
• If I were Mr Son, I would be thinking twice about leading or even participating in this round.