Cognitive solutions have developed beyond traditional AI capabilities into machine thinking, and startups can now leverage cognitive technologies like IBM Watson through the cloud to offer better solutions to their customers.
IBM’s Watson can digest and reason through information, and come up with solutions. This allows startups to offer their customers APIs like natural language processing without the obstacles that prevent easy, robust and responsive user experiences.
Why AI and the Cloud Go Together
Systems like Watson need two things most startups don’t have: a lot of computing power, and a lot of data to work with. Watson’s original question-answering system is a good illustration. When it played Jeopardy! in 2011, it worked from 200 million pages of content, and IBM master inventor Tony Pearson estimated its hardware cost at about three million dollars, according to Wikipedia.
No early-stage company could buy that machine to add one feature to an app. The cloud splits the cost. The provider owns and runs the hardware, keeps the models up to date and spreads the cost across thousands of customers. Each startup pays only for the requests it actually makes, and can use more when demand grows without buying anything new.
Get the TechPortfolio briefing
The month’s fintech and funding news, plus new tools for founders, in one short email.
Natural language processing, the example IBM highlighted, is a good fit for this model. It covers tasks like understanding what a customer is asking, pulling key facts out of documents, or judging whether a review is positive or negative. These are things a small product team might want but would never have the time or specialists to build themselves.
Placing this advanced technology in the cloud provides the resources that cognitive needs to function smoothly, and for it to be accessible through platforms including IBM Bluemix.
What Happened to Bluemix
Bluemix, IBM’s platform for building and running apps in the cloud, launched in public beta in February 2014 and became generally available that July, according to Wikipedia’s IBM Cloud entry. In October 2017, IBM announced it would bring everything together under the IBM Cloud brand, retiring the Bluemix and SoftLayer names.
IBM’s cloud strategy then shifted toward “hybrid” cloud, meaning software that runs across several public clouds and a company’s own data centres. Its biggest step was buying Red Hat for about $34 billion, the largest acquisition in IBM’s history, completed in July 2019, according to the same entry.
Watson’s Mixed Results
Watson’s commercial journey turned out to be harder than the early pitch suggested. Healthcare was one of its earliest and highest-profile targets: in 2013 IBM announced Watson’s first commercial application would help with decisions about lung cancer treatment at Memorial Sloan Kettering Cancer Center in New York, Wikipedia notes.
By 2022, IBM had decided to step back from that business. In January 2022, it agreed to sell healthcare data and analytics assets from Watson Health to the investment firm Francisco Partners, IBM announced. The business became a separate company called Merative later that year.
The lesson for anyone building on AI is that impressive technology isn’t the same as a product customers will rely on. Medicine, in particular, needs accurate data, careful testing and trust from the people using the tool, and those take far longer to build than a demonstration.
Watson Today
IBM reused the name for its next generation of AI. On May 9, 2023, it announced watsonx, a platform for building and running AI with foundation models and generative AI, along with tools for managing data and governing how AI is used, according to IBM’s announcement. IBM said watsonx.ai would let developers train, test, tune and deploy models, and would offer thousands of open models and data sets through a collaboration with Hugging Face. The company also said the platform was built to run across any cloud environment.
The core idea from 2016 is still there: AI delivered as a cloud service that a business can tap into rather than build. What has changed is the kind of AI on offer, and how many companies are now offering it.
What It Means for Startups
For startups, the case for using AI through the cloud is stronger than ever, because the most capable models are far too expensive for a young company to build or run alone. But the Watson story is a reminder to judge an AI service on what it reliably does for your customers, not on how impressive it looks in a demo.
It’s also worth choosing tools in a way that lets you change your mind. Product names, platforms and pricing in this market change quickly, as Bluemix’s short life shows. Keeping your own data and product logic separate from any one provider makes it easier to switch when a better option comes along.
Frequently Asked Questions
Why does cognitive technology need the cloud?
Systems like Watson need large amounts of computing power and data that most startups can't afford. Placing the technology in the cloud gives it the resources it needs and makes it available to startups through platforms like IBM Bluemix.
What can startups offer customers through IBM Watson?
Startups can offer APIs like natural language processing. That covers tasks such as understanding customer questions, pulling facts out of documents or judging whether a review is positive or negative.
Why did IBM sell Watson Health?
In January 2022, IBM agreed to sell healthcare data and analytics assets from Watson Health to Francisco Partners. The business became a separate company called Merative later that year.
What did IBM's acquisition of Red Hat mean for its cloud strategy?
IBM bought Red Hat for about $34 billion, completed in July 2019, the largest acquisition in its history. IBM's strategy shifted toward hybrid cloud, running software across several clouds and companies' own data centres.
What is IBM watsonx?
IBM announced watsonx on May 9, 2023, as a platform for building and running AI with foundation models and generative AI. IBM said it would offer thousands of open models and data sets through a collaboration with Hugging Face.