In my last article, I talked about OpenAI’s weird and quirky corporate structure which led to its CEO, Sam Altman, being ousted by the board. This ended up being short-lived as he was reinstated with a new board merely three days later. The growing pains of working at a start-up, one can say.
I wanted to focus more on the state of OpenAI’s products and take a step back to see if it can fulfil its mission of creating AI for the betterment of humanity whilst also making this open to all, in 2024. Currently, OpenAI has two main products; ChatGPT (the consumer-grade chatbot that we use for anything and everything these days), and the GPT-4 interface (a stripped-down version of ChatGPT’s underlying technology that businesses can customise). It is apparent OpenAI have been hugely successful in marketing and selling licenses for these products, with them being valued at over $86 billion, and recently announcing that they have accrued $2 billion in revenue on an annualised basis. These values would put OpenAI up there with the dominant players of Silicon Valley in a much shorter timeframe. They show no sign of slowing down either, as they are forecasted to double this revenue figure in 2025.
While this may look like a promising future moving forward for them, its real game-changing AI use cases from its products amongst the major companies seem kinda scarce? Sam Altman claims that 92% of Fortune 500 companies are using their technology but there has not necessarily been a sensationalist product created with AI seeing exponential productivity growth, or it rapidly accelerating a business’s functions. Instead, there still seems to be the feeling of OpenAI still setting its feet in what direction they want to truly go towards – do they go about creating specialised commercial-grade products for their clients, or do they still follow their original mission of creating this overarching super intelligence for all? I can’t say for sure.
Regardless,Sam Altman and his team have created partnerships with the likes of Microsoft(they kind of had to here - see my last article) and Rakuten, which bring strategic and financial firepower that OpenAI can use. Zeroing in on Microsoft, they license OpenAI’s product directly given that they are also a key investor, and have recently the AI-powered Microsoft Copilot, that we can experience on their Office suite of products. I’ve tried the product first hand and while it isn’t as neat to use as ChatGPT, its integration into existing software that millions of end users are familiar with is a great advantage to have. This is one of a handful of examples of OpenAI’s products that promise of a game-changing, killer factor for productivity for its clients, but it is just that – one of a handful.
The other successful use cases that we see from OpenAI’s services seem to only work in narrow workflows, such as creating a chatbot for customer responses and interactions, but this is probably a textbook use of ‘Narrow AI’ – which is the level of AI that we are at currently. Narrow AI can work best in isolated use cases for businesses, being really good at one thing, but asking it for a holistic business solution seems out of reach.
Many businesses are still in the experimental phase of implementing this type of AI and seeing how, or if it’ll even create significant cost savings or productivity increases. This feeds into the idea of the initial scepticism that we all have when dealing with new technologies that claim to greatly speed up our processes.
On the other hand, the vast amounts of capital and financing needed to train these models are also seeing an exponential increase. So, whilst OpenAI are seeing record revenues in the billions, a genuine question may be asked of AI models costing more(!) to train. This is only exacerbated by Sam Altman’s dream of creating this super-intelligent AI which may cost upwards of $7 trillion. Wow.There may well be a possibility that OpenAI do go bankrupt before they even reach these goals. With typical sources of capital such as VC funds now not being able to compete on the pure values and returns of what OpenAI is currently outputting, other higher-end private sources will have to be sought after, think infrastructure-level PE funds and investors.
Again, this comes back to the underlying question of the true mission or motivation of OpenAI as a company; do they commit to becoming a profit-making start-up that follows their client’s needs to a tee, creating super specialised products that may well see productivity gains and cost-savings, or do they fully lean into the ‘betterment of human society’ route of their original mission, accumulating and spending an obscene amount of capital for something which could truly be a game-changer in the way we perceive machines? Right now I see Sam juggling these two opposed goals, creating some sort of identity crisis for the company as a whole, as they win new business and clients but all roads leading to this fantastical idea.
Investors and academics will give you different answers on what is the better route to take, but as of now, OpenAI needs to at least ensure that their short to medium goals of capital requirements are met, and truly win businesses with tangible and not novel gains in business efficiency of their products. Less bigger picture world-changing stuff in place of more viable solutions.
.jpg)
