Ep. 10 CIOs are Pushing Back on Generative AI Hype | AI Insights and Innovation

In this episode of AI Insights and Innovation, David Linthicum, principal analyst at theCUBE Research, explores the challenges CIOs face in identifying effective use cases for generative AI within enterprises. He highlights issues such as aligning AI capabilities with business needs and the difficulties of scaling from pilot projects to full implementations.


Linthicum also discusses tactical approaches CIOs can adopt, such as focusing on value-driven initiatives and integrating small language models, which offer cost-efficiency and faster deployment. By providing examples of successful implementations, Linthicum shows successful implementations, underscoring the benefits of small models for specific, niche needs.

#theCUBE #AIInsightsAndInnovation #theCUBEResearch #CIOs #GenerativeAI #scalability #SLMs

[AoDqqLTUv]

I worked at one of America’s largest lawyer. We are deploying AI and are already seeing massive productivity gains.

[sullyguy]

This was a very balanced view. As with most issues humans think about, this one is often divided into 2 extreme positions:

There are actual people claiming it is all a hoax (I can sympathize with this due to terminology like “artificial intelligence” which captured the importance of the innovation but oversells it to anyone who knows a bit more than average about intelligence) and that it will never do anything useful (a demonstrably false position)

On the other hand, many of the optimists don’t seem to know anything about how technology is adopted. The reports, the studies, the pilot projects the cancelations are restarting from scratch, the budget meetings and fighting…

I have seen computer systems that were out of date when they were first adopted yet resist massive efforts to replace them despite everyone at every level understanding the waste of money the old system is causing (except maybe one new authority who decides to cancel the new system for “cost” reasons). These old systems survive better than Hollywood horror monsters. They just can’t be killed.

People don’t understand how much needs to be redesigned and changed to implement a new system, just like they don’t understand how complex the “simple labour tasks” are that they think can be automated in a few weeks by a sufficiently cool looking humanoid robot. They think replacing a computer system is a simple matter of pressing “uninstall” on the old and “execute” on the new while the business is shut down overnight or hey look at file footage in news stories of people “working” a job — footage designed not to distract the viewers from the main point and think “That’s what people do at jobs? That can be automated by a Tesla Bot today.”

This is all going to take a long time to play out.

[glennmontague]

Source for the 3-5x cost for AI systems vs conventional systems?

[RobertoDF]

Good summary and interpretation of where the market currently is. I work for a Fortune 500 company and I see some of the points you’re making here. From the board of directors pressure to find the ‘billion dollar use case’ to the underwhelming results from our GenAI proof of concepts.
In my personal opinion I think the technology is not mature or powerful enough to provide the ground shaking results some are promising. It needs more time.