Should we care about the OpenAI shenanigans?

last updated:
openai changes

Should we care about the OpenAI shenanigans?

last updated:

In GenAI, there is, though, some remarkably good technology that will find its rightful place amongst a range of other flavors of AI and ML and, over time, will transform enterprises.

Something in the range of $2 to $3 Trillion is associated with the global IT industry, approximately the same amount of money as the entire GDP of Africa. So, if you are part of the Silicon Valley bubble, it’s easy to think your headlines are world headlines. The furor over the firing and hiring of Sam Altman from OpenAI to Microsoft is a case in point. Who in the tech world has not followed the shenanigans closely over the past weekend? Few care outside the tech sector; it’s not a headline that interests them; the Travis Kelce and Taylor Swift romance is more significant news for many. Or, to use that famous phrase, “Today’s news is tomorrow’s history”.

The Information and Automation Management sector has a foot in the tech and the real world. So, the question we ask at Deep Analysis is beyond the dubious pleasure of car crash voyeurism: why should we care about who is, or is not, running OpenAI? The answer is that we should care, but not in the way you might expect.

First, we must look at OpenAI and its current impact on our market. On the one hand, its impact has been immense as virtually every vendor in the sector has felt compelled to articulate how they are doubling down on GenAI; they have felt the need to ride this wave of tidal proportions. But after this weekend’s wild antics, many may have to think about how to jump the shark rather than continue to ride the wave. Partnering with OpenAI and becoming reliant on a single large vendor’s LLM is now open to question. We all know now that this 800lb tech gorilla isn’t as robust, viable, or reliable as we thought. To mangle another overused phrase, the cat is out of the bag; OpenAI is no longer the invincible giant it appeared to be, and there is no undoing that.  But it should not have taken the firing of the CEO for us to see a red flag, as red flags have been flying left, right, and center for the past year.

Since late 2022, the tech community has been held breathless in awe of OpenAI and the potential of Generative AI. The launch of ChatGPT was seen, rightly so, as a breakthrough moment. But the real world, though also impressed and eager to play around with ChatGPT, asking it to answer any ridiculous or serious questions, immediately spotted that it had some significant drawbacks, most notably its ability to make stuff up convincingly. Fun for sure, but ready for the harsh reality of a regulated business world, clearly no. Add to its ‘hallucinations’ questions regarding safety, compliance, and security, and the red flags have been flying aplenty. An update to ChatGPT was expected to sort these problems, but it didn’t, with some reporting that the newer version was no better and, at times, worse than its predecessor. Then there is the fact that OpenAI allegedly costs $700,000 a day (or more) to operate, with a single question asked of it costing around 45 cents to process. This from a company with no articulated plan to break even, let alone turn a profit. I mean, how many red flags do you need?

OpenAI has some severe issues to resolve, but whether they do resolve them is ultimately of little fundamental importance to the information and automation management world. To use the exact phrase twice, the cat is already out of the bag. GenAI is here to stay and will play a significant role in the future of our sector regardless of the success or otherwise of OpenAI. At Deep Analysis, we have advocated strongly for enterprises using their data (not public LLMs) and the potential for tightly and accurately curated micro LLMs to deliver business benefits. The challenge for enterprises big and small is not whether GenAI works; it does; the challenge is in curating and managing accurate, safe, and targeted data to train the GenAI. And secondly, in focusing the use of GenAI on specific high-value business activities.

You only need to scratch beneath the surface of the marketing speak from tech firms in our sector to conclude that they have known this all along. If you are a Hyland, UIPath, or OpenText customer, you need to work on your data, prepare, clean, and curate it so that AI can leverage it. Similarly, suppose you are a Microsoft or Oracle customer, beyond a few low-hanging fruit use cases such as generic customer support or creating marketing copy. Then, there is much work on the data to do. There are no shortcuts, magic wands, or silver bullets. In GenAI, there is, though, some remarkably good technology that will find its rightful place amongst a range of other flavors of AI and ML and, over time, will transform enterprises. But there is a great deal of work ahead to make that a reality, and neither Sam Altman, Elon Musk, or Jeff Bezos can blow fairy dust to do it for you. Buckle up, folks; GenAI is real; it’s here to stay, but the tabloid stories of Sam’s firing and hiring, though entertaining for sure, are ultimately little more than a minor blip in the long-term future of enterprise AI.

Leave a Comment

We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it.

Work Intelligence Market Analysis 2024-2029