Industry Voices: Altman's short goodbye, rapid re-entry and the saga of OpenAI

  • Appledore's AI analyst runs down the OpenAI saga

  • He examines the app store-based big tech move that preceded the crisis

  • He looks at what might come next for OpenAI

Back at the start of November, Sam Altman paced the stage calm and confident at the first OpenAI’s developer conference. In black jeans and Adidas sneakers, any angle seemed to be covered to portray himself as the next Steve Jobs, or at least, this year’s Mark Zuckerberg.

OpenAI’s most recognized exec had gathered about 1,000 software engineers and AI researchers into a former Honda dealership in downtown San Francisco for an event that projected his company directly into the tech giants league.

This kind of scene was popularized by Jobs, who used such events to announce the first iPod, and, years later, the first iPhone. Delivered in a signature black turtleneck, his presentations helped fuel the frenzy around Apple’s iconic devices — his recognized “reality distortion field” that made Apple into one of the most lucrative gatekeepers in business history. At that stage, Altman was using exactly the same playbook to reinforce OpenAI’s dominance in the AI era.

A little history

Launched as a nonprofit in 2015, OpenAI was originally created to keep advanced artificial intelligence (AI) out of the hands of monopolistic corporations. But since accepting a major investment from Microsoft in 2019, the company transitioned to a novel for-profit structure. 

OpenAI often says it is still pursuing its original goal of building AI that “benefits all of humanity.” But its path forward lately looks more like business as usual. In May, the company declared a open season, approaching executives from Meta, Apple and Amazon Web Services. Something very unusual for a company where jobs postings were traditionally dominated by academic and research requirements, without paying much attention to product and business management roles. 

Around the same time, Altman zipped around the globe, visiting world leaders and developers in dozens of cities, including Doha, Qatar and Tel Aviv, one-upping Zuckerberg’s 2017 tour of America. Last month, the company expanded its footprint in San Francisco, subleasing nearly 445,000 square feet of office space from Uber, purchased when former Uber CEO Travis Kalanick was still the most envied founder in the Valley. A peculiar behavior for a so-called non-profit organization like OpenAI.

A Big Tech power play

Half an hour into his keynote speech, Altman unveiled a Big Tech power play: OpenAI’s latest upgrades would make it easy to build customizable bots, called generative pre-trained transformers (GPTs) without knowing a lick of code. The company would promote the best options in a proprietary app store and share “a portion of our revenue” with people who build the most useful and popular tools, he said.

The GPT Store marked OpenAI’s effort to make itself the central hub for generative AI, and, just like Apple, an unavoidable platform for anyone building a business.

What Altman avoided mentioning was that, if all went to the plan, OpenAI’s store would demolish the AI strategy of many enterprises and the start-ups built on top of ChatGPT.

This was a threatening move for many investors that always imagined OpenAI being the generative AI back-end for their highly lucrative business, and a betray for some key employees that see OpenAI as the non-profit organization created to achieve superhuman level intelligence and be a counterpoint to Google’s dominance in big tech. More than enough reasons to steer the company trajectory in its way to conquer the world, and block anyone else from doing that.

By opening the store, which is only accessible to those with a ChatGPT Plus subscription, OpenAI captures the value that was going to outside builders on its own platform and makes paying for a subscription more enticing.

After the huge impact of ChatGPT launch (destroying the TikTok record of most visits within the first three months) they saw the opportunity, and they wanted to own it. It would be the equivalent of enabling apps to sell more iPhones, which for many tech observers and analysts, might be seen as an inevitable outcome of ChatGPT’s viral success.

OpenAI’s use of recognizable tactics in the hopes of becoming the dominant platform in the AI world has elicited the kind of push-back that giants like Facebook and Google faced much later in their rise. It has invited scrutiny about the inherent contradictions in the story OpenAI tells about its altruistic roots, including its transition to a for-profit. Many investors perceive that Altman skipped past the growth stage that a start-up naturally goes through and is moving straight into Big Tech with a closed platform.

Despite parallels with the rise of Facebook and Google, the speed of AI advancements makes the future much more unpredictable. It’s also hard to compare aggressive AI upstarts to the tech giants that ruled the last internet era, because the two groups are connected. 

Companies like OpenAI and Anthropic are reliant on Microsoft, Google and Amazon to access the computer chips necessary to train today’s large AI systems. Nevertheless, the first red flag was raised in October when it was announced that the company was exploring the possibility of developing chips in-house, according to a report from Reuters which cites sources familiar with the plans. Though no final decision was communicated, the company was apparently evaluating a potential acquisition target to accelerate this process. This was perceived by many as yet another declaration of independence of OpenAI on its way to conquer the AI market and leave them wondering what the reaction of Microsoft to such possibility would be.

Altman’s repeated calls for regulation in D.C. and willingness to link his products to existential risk have rankled start-ups and leading venture capitalists back home, including those who helped Big Tech craft their world-devouring strategies. Meta board member Marc Andreessen, founder of the firm Andreessen Horowitz, has condemned Altman as someone who has tried to use regulatory barriers to block competition. 

There is nuance in saying “regulate us, but not smaller competitors’ that somehow ends up getting lost and we just get bashed,” Altman posted on X after tech leaders criticized the White House’s AI-focused executive order, which Altman supports.

Nevertheless, what we are already seeing in the market is that could be challenging for OpenAI in solidifying their dominance and of locking out some of the new competitors. But, the question is not if OpenAI plays the dominant role, the question is… who will? We think Microsoft's CEO already has an answer in mind for this question.

Then came Black Friday

On Friday, Nov. 17, OpenAI announced that co-founder and CEO Sam Altman had been fired by the board of directors, and that Mira Murati, the CTO, would serve as interim CEO. OpenAI's board claimed that Altman was “was not consistently candid” with them, without providing details.

It also said that another co-founder and Chairman Greg Brockman, would be removed from that role while staying at the company. But Brockman then indicated that he would quit in solidarity with Altman. That meant that there was only remaining member of the core founding group behind OpenAI: Ilya Sutskever. A Russian-born Israeli-Canadian, Sutskever is a global leading expert in deep learning and for many the main obstacle for Altman’s commercialization plans. In fact, Sutskever voted in favor of firing Altman as CEO. 

In one of the many dramatic turns that occurred over that weekend, 500 OpenAI senior employees signed an open letter presented on the following Monday requesting a collective resignation of the entire board, including Sutskever, and Altman’s and Brockman’s reinstatement.

Some members familiar with the process said that the explanation for this shocking development was that Altman and Brockman were playing behind the scenes to transition the company from non-profit to, potentially, one of the most profitable and powerful companies on the planet, even putting profits before safe and common good.

Although as I’ve previously stated, given the success of ChatGPT during this year, plans for product commercialization cannot be qualified as secret and as the drama continues, this reason seems to lose credibility by seconds.

A final twist?

In a dramatic turn of events, Altman was reinstated as CEO of OpenAI after a collective resignation of the entire board. What might have been considered as a mastermind operation to take control of the company, seems to be the result of naive miscalculated actions from some members of the board, and making evident to the entire world the danger of putting mediocrity in the driving seat of disruptive companies.

Roman Ferrando, principal analyst at Appledore Research, is an accomplished AI professional with outstanding depth of knowledge and an extensive track record in developing and implementing AI strategies across diverse organizations, including Openet, Amdocs, MYCOM-OSI, B-YOND and Ericsson. Ferrando holds a Ph.D. in Generalized Anomaly Detection for Telecommunications Networks and has used a deep understanding of cutting-edge AI technologies and their applications in telecom to successfully spearhead AI initiatives at director and executive levels. Ferrando holds numerous patents and has published extensively in the domain of AI and telecommunications.

Industry Voices are opinion columns written by outside contributors — often industry experts or analysts — who are invited to the conversation by Silverlinings' editors. They do not represent the opinions of Silverlinings. Read all of our Industry Voices here.