The Year That Was For Generative AI— And The Shape Of Things To Come
The last 12 months have been a wild rollercoaster ride for generative AI. Almost a year after OpenAI’s ChatGPT sent waves of excitement across the world, its chief executive, Sam Altman, was fired from the company. The last time we saw something close to this was in 1985 when a power struggle at Apple pushed out Steve Jobs. However, the events at OpenAI will have a more profound impact on business thinking and technological advancement, which is of interest to all of us.
The boardroom drama unfolding at OpenAI exposed a cultural schism dividing the world around generative AI. The board disagreed over Altman's ambitions of sprinting ahead to commercialize generative AI. OpenAI’s chief scientist, Ilya Sutskever, wanted to slow down, concerned that the perils of adopting advanced AI were not being adequately understood or addressed. Altman’s sudden ouster sent shockwaves through the corridors of technology; Microsoft, a major investor in OpenAI, then hired Altman to drive its AI initiatives; by the third day of the boardroom drama, OpenAI had seen three CEOs; 747 of the 770 OpenAI employees threatened to resign; hectic discussions with Altman saw him return to the company; and the company board was reconstituted as a pre-condition to his return. Looked at one way, Altman sacked the people who sacked him. Phew! The events provide first-rate grist for the Netflix documentary mill.
In the last 12 months, ChatGPT has shown us what a real, 100 per cent certified technological revolution looks like. The product acquired a phenomenal 1 million users in the first week of its launch on November 30, 2022 (we wrote about it here). By November 2023, it had 100 million active users with 1.6 billion monthly visitors to its website.
Businesses have been finding use cases for generative AI faster than anyone can count. Alternatives to ChatGPT have sprung up at a furious pace, at an average of over two every month. Today, you can pick from Bard, Bing AI, LLaMA, Claude, Replika, Chatsonic, Jasper Chat, NeevaAI, Character.ai, CoPilot, Poe, Perplexity, and another two dozen such competitors. ChatGPT—and its look-alikes—have affected every industry, from entertainment to manufacturing, retail to life sciences, and financial education services. My favourite is ChatPDF, which lets me converse with my PDF documents (Confession: I am a “technology-for-everyday-things” junkie).
Today, 12 months after its launch, OpenAI is on track to generate revenue of USD 1 billion by the end of 2024. More importantly, over 60 per cent of its users are aged between 18 and 24, and over 50 per cent of people cannot distinguish between content generated by AI and that created by humans. Without a doubt, OpenAI has become a serious contender for a `Technology Company of the Decade’ award. And its story has barely begun.
On the business front, these developments have shown us how deeply organizational structure impacts a company's mission. Open AI was set up to benefit humanity. Co-founders Greg Brockman and Ilya Sutskever said in a blog in 2015, “Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Since our research is free from financial obligations, we can better focus on a positive human impact.”
The stated mission of OpenAI in 2015 is not sustainable from the point of view of investors who have pumped millions into the company. In 2018, when Elon Musk quit the board, it cut off OpenAI from badly needed funding. When ChatGPT 2.0 was released in 2019, OpenAI corrected. It announced the creation of OpenAI LP, a “capped-profit” company, letting investors receive “capped returns.” In the OpenAI context, returns would be limited to 100X the investment in the first round (the balance of profits goes to the nonprofit company). This ensured capital became available for the heavy-duty infrastructure required to develop and maintain the AI models.
With the ever-increasing need for capital and the duty to investors and employees, the company's structure will see change. Today, for example, Microsoft has an investment of $13 billion in OpenAI (and owns 49 per cent of OpenAI) but no seat on the board and no say in how the company is run. This seems untenable.
Many of us will rightly bat for the lofty ideals OpenAI began with—that of positively impacting humanity. But I am placing my bets on OpenAI becoming like other for-profit companies, without which it cannot keep finding investments. Its need for capital will grow as the company seeks more computing power, invests in strengthening its legal and ethical posture, and hires more of the best AI minds in the world.
AI was always going to be an extraordinary stream of technology. But could it become a threat to, for example, political systems? Or global financial stability? Or to open-source software? The answers tell us why everyone should closely watch developments at Open AI.
As it happens, Gary Gensler, the US Securities and Exchange Commission chairman, thought it would affect financial systems. He was concerned that AI companies may put the interest of advisers ahead of investors, “put the broker ahead of the investor.” Regulators could use the same technology to regulate the behaviour of financial companies. But as we know, regulators primarily respond to developing situations, rarely getting ahead of the technology curve. By the time they respond, it may be too late.
These apprehensions and other developments around generative AI have triggered urgent action worldwide. The US government, for example, is making it mandatory for developers to share internal testing information that the industry has traditionally kept private.
OpenAI could also disturb the open source movement with its ability to write code and develop applications/functionalities based on simple prompts.
In terms of technological developments, the return of Altman at OpenAI means the highly controversial “Q*” model – pronounced “Q-star” – will continue. The model can reportedly solve mathematical problems it has not encountered before. If this is true, we might see the birth of superintelligence (a.k.a. artificial general intelligence or AGI). But will this superintelligence eventually challenge human control? That is the multi-billion dollar question we might see answered in 2024. Strap in. The giga-coaster ride is about to begin.
_20240501181242original_image_36.webp)