If you’re at all involved in technology, unless you've been living under a rock or in a cave, you’ll no doubt have heard of generative artificial intelligence, or gen AI, for short.
The technology is a type of machine learning (ML) that can produce content as diverse as images, text, audio and code, following user prompts. AI applications like ChatGPT, Google’s Bard, and DALL-E can be used to create this type of content.
Full disclosure: Silverlinings uses AI app Midjourney to create our own fantastic artwork.
[Ed note: No more robot pics, please!]
Generative AI does this by using various learning methods “to easily and quickly leverage a large amount of unlabeled data to create foundation models,” according to Nvidia. These foundation models can be used as a basis for AI systems, such as ChatGPT, which in turn, can be used to execute myriad functions.
E-commerce (and cloud!) giants like Amazon and Google are already using AI to make their supply chains more efficient. While social media titan Meta uses ML algorithms to help decide what delights it can toss onto your feed.
Companies as diverse as Chase, Ford and Twitter are also using AI for various tasks, and many other enterprises are starting to breathe in the AI atmosphere.
Earnings don't lie when it comes to AI - or do they?
Some of the firms that are supplying the infrastructure for AI, such as Nvidia, are already reaping the benefits. The chip designer blew past its second quarter earnings expectations with revenue up 88%. Similarly, software companies like IBM are seeing a boost from AI.
“In the context of the telecom sector, the feasibility of applying generative AI, beyond the well-known chatbot use cases, is cast into serious doubt, as most of AI/ML use cases were already grappling with substantial challenges in delivering business value,” Appledore AI director Roman Ferrando commented in a recent LinkedIn post on Gen AI. “The resources and data required to construct Gen-AI solutions serve to underscore this skepticism.”
AI apps like Chat GPT and Bard are subjected to training using mountains of data from the wide realm of the open Internet. “This reliance on data brings along substantial energy consumption and intensive computational demands,” Ferrando said. “When enterprises, and telecom operators, contemplate the construction of Gen AI models internally, they must weigh the significance of the data volume integral to achieving relevant outcomes.”
In other words, companies shouldn’t just take puffery from the AI crowd at face value. Ferrando said that firms need to figure out if they have the data, resources and infrastructure necessary to establish systems that are actually useful to the firm.
“AI...comes with a higher price tag, presenting a steeper challenge for companies to thrive,” Ferrando said.
So we could think of this as a classic tech industry bubble. A few companies will benefit greatly from generative AI. Many, many fresh and fancy startups will start strong but end up buying bought up for peanuts or going bust. While plenty of companies will spend big on AI and derive little from it.
Are you looking to understand key cloud business cases, identify and solve deployment challenges, innovate through interactive agile workshops, learn from hyperscalers and service providers successfully deploying cloud services and network with key players in our industry? Check out our in-person Cloud Executive Summit in Sonoma, Calif., from Dec. 6-7, 2023.