Sam Altman's Reckoning
I can’t help the feeling that Sam Altman is running a con job.
I have had the feeling with a number of companies and people in the past. Here are a few examples (I have more):
I would count Cintas (CTAS) and its former CEO Scott Farmer as one such example which was more than a hunch. I reported that one to the FBI, but zero came of it. No worries. Harry Markopolos had it far worse trying to get the FBI to pursue Madoff, who it turns out was protected by a family connection within the FBI. Read Harry’s book HERE. By the way, Harry was right about GE.
I like Elon Musk for what he did for free speech by purchasing Twitter and preventing the FBI from censoring content as it did under previous ownership. However, I did not like how Musk front-ran his crypto buying and selling back in the 2020-2021 period. Further, I was not a fan of how he tweeted about TSLA without SEC penalty when 99.9% of CEOs would have faced the music (at least when the SEC had teeth in the 1990s and early 2000s). The SEC was a joke under former Chairman Jay Clayton and is almost as bad under Gary Gensler.
I am not a fan of these clowns that ran SPACs in the 2021 period to cash in on the retail investor while valuations and investor animal spirits were at their peak. Count Chamath Palihapitiya among this group.
I detested WeWork’s Adam Neumann, FTX’s SBF, and Nikola’s Trevor Milton, all of whom found it difficult to give a substantive explanation for how their respective businesses worked.
Sam Altman is not a fraud in the vein of Madoff. AI is real. Advanced Automation has been around since the 1950s. We saw a revolution with machine learning and deep learning that began around 2010. The benefits are real.
However, the picture that Altman paints of Artificial General Intelligence (AGI) in several years time is suspect and seems self-serving. First, define exactly what you mean by AGI. What exactly will the first AGI systems be capable of? Provide a couple of concrete examples of the practical output. Gen AI is not an exact science, I get it. Who knows if the scaling laws will continue to apply? I get it. Who knows how users will deploy AGI? I get that too.
However, the timing of OpenAI news over the past few weeks is a bit suspect. For example:
Sam Altman has been in the midst of a capital raise for months now. It was leaked (obviously by OpenAI) weeks ago that OpenAI’s negotiated valuation could be north of $100 billion.
OpenAI drops its o1 preview model on September 12th. While everyone I came across wass impressed by o1, I also know of a number of developers who prefer Anthropic’s Claude LLM to o1 (Anthropic is also raising money now). The o1 release followed cryptic tweets by Altman designed to create hype in my view.
OpenAI may be valued as high as $150 billion in its latest round. This is also an Altman leak. Who else would leak it? Not the investor cohort. The valuation is dependent upon a change in corporate structure so said the September 14th Reuters story. OpenAI is now restructuring to be a for profit company.
Sam Altman posts a rare blog post about the coming benefits of AI which is full of generalities designed to create hype, much as Altman did with his cryptic tweets leading up to the o1 preview release.
Meanwhile, OpenAI is bleeding senior executives (see our previous notes HERE and HERE).
Sure, OpenAI’s $4 billion revenue run rate is impressive, but what would that number be if we stripped out MSFT? What percentage of revenue is recurring? What is the customer retention rate? How many end-users come back to run queries either on ChatGPT or via APIs each day, week, month? Who are the operators given the senior level departures? How has the culture changed since the November 2023 drama and the employee departures?
Sorry, Sam. You want me to pay approximately 40x revenue, I want to know why you feel so insecure as to sport fake hair on your head (I would have asked Elon the same question in the early Tesla days).