IBM was into AI early on. Its first Watson implementations showcased that it wasn’t the cost of the hardware or software that made AI difficult; it was the cost of creating the massive datasets needed to train the platform. These training sets were exorbitantly expensive and very time-consuming, so that initial Watson implementation failures were tied more to the cost and unavailability of the training sets. However, over the decades since Watson was created, IBM has acquired or created some of the largest data sets on the planet and now can largely handle your AI training needs if they reside in one of IBM’s core markets, like cloud, finance, banking, and government. IBM’s more complete offering is called watsonx, and it is impressive.
Already 100 IBM clients have begun to leverage watsonx and its massive practical, operational, and integration knowledge that IBM typically supplies as part of any Watson engagement.
Let’s talk about watsonx this week and how it is about to change much of our world.
The power of watsonx
There are a ton of generative AI startups in the market now. Most of the solutions I’ve seen are impressive, at least on paper. But the issue for most of them is their lack of training data, which is usually acquired and sorted by the client firm’s staff and can take months to build up and integrate with the new AI system.
But watsonx already has much of the data needed to train it for initial tasks that aren’t specific to the company, and IBM’s focus on assuring that training data is accurate and that the solution is highly secure makes for a very attractive offering at a time when it seems everyone and their brother is bringing new generative AI solutions to market.
IBM worked with I50 companies across a variety of industries to complete this uniquely complete AI offering, and the cost savings from not having to build a training set from scratch likely makes watsonx one of the biggest enterprise bargains currently in the market.
Banks like Citibank and world-class manufacturing companies like Samsung have embraced this product aggressively, seeing it as providing a significant competitive advantage. Given Samsung’s position against Apple and Apple doesn’t appear to be anywhere with generative AI yet, there is a good chance Samsung will be able to use this recently acquired watsonx technology to more aggressively compete with Apple in what should be an early case of what can happen if one competitor embraces and effectively uses generative AI and the other does not.
The increased interest in watsonx is amazing, given the platform isn’t even complete. The last module that builds on watsonx.ai, watsonx.data, and watsonx.governance isn’t quite ready yet and won’t ship until the end of the year.
watsonx.governance is particularly fascinating because it focuses on transparency, explainability, and responsibility for the solution. This is the part of the solution that assures that it complies with regulatory and policy requirements, something often left out of other AI solutions, and better assures that the outcome of deploying watsonx is positive for the company acquiring it.
Wrapping up: Generative AI is exciting but…
It is great that everyone seems to be extremely excited about generative AI, and is hell-bent-for-leather to deploy it. However, many solutions are very immature, have no intrinsic training data to speed implementation, and don’t yet take into account the need for AI transparency and governance.
watsonx may be the only truly complete generative AI solution in the market. While its training sets don’t encompass every industry and function yet, they go farther than anyone else I’m aware of. There is an old saying that applies to new products that the “prospectors get the arrows; the settlers get the land.” Most of the AI solutions in market aren’t solutions and will clearly require a tremendous amount of oversight and participative cost by the purchasing company. IBM watsonx is different in that it is far more complete and thus should be faster and cheaper to implement, allowing IBM to showcase many customer testimonials on the product this month.
watsonx is finally available to the broad market, and I expect this market will prefer to be more of a settler than a prospector.
- HP’s Threat Report – New Threats, Bigger Problems - February 23, 2024
- Lenovo Partners with Anaconda for AI to Stretch its Workstation Leadership - February 19, 2024
- AI and the Supply Chain: Making More Resilient Companies - February 9, 2024