Author| David Linthicum
Planning| Yan Zheng
Nowadays, no one doubts the power of AI, but enterprises must be aware that it will also lead to the deployment of too many Applications, scaling issues and cost overruns.
Because of my experience in AI development and integration with enterprise and cloud architectures, I know the benefits of generative AI. However, I also know that while there are many benefits, there are also drawbacks that must be considered simultaneously. Because generative AI is developing so quickly, it is important to decide how to manage it effectively and reduce any negative impacts.
I propose three major shortcomings of generative AI that cloud computing professionals need to understand and manage.
This is the biggest problem I see. Now, with no-code or low-code mechanisms, we can build applications more quickly using generated AI-driven development tools. The number of applications deployed (all needing to be managed) can easily get out of control.
Of course, it’s good to speed up application deployment to meet business needs. The application backlog limited business growth in the 90s and early 2000s, so any way to improve it is good for business, right?
However, I see an almost reckless approach to application development. The work required to build and deploy these systems takes only days and sometimes hours. Companies don't put much forethought into the overall role of applications, and many applications are purpose-built for tactical needs and are often redundant. CloudOps teams are trying to manage three to five times the number of applications and connected databases they should. The whole mess is not going to scale and the cost is too high.
Generative artificial intelligence systems require more computing and storage resources than currently provided. Driving greater scale requires leveraging these resources and is not as simple as turning on more storage and compute services.
To rapidly expand the use of generative AI systems, consideration and planning must be made to obtain and deploy additional resource support. This generally depends on the operations team correctly deploying the right amount of resources without compromising the value of the system or limiting its capabilities. The trade-offs here are almost endless.
We may notice that while we are busy building financial operating systems that monitor and manage cloud costs, the cost of generating AI systems has increased dramatically. what should you do?
Actually, this is a business issue, not a technical issue. Companies need to understand why cloud spending is happening, how it is happening, and the return on business benefits. The costs can then be included in the predefined budget.
This is a hot spot for businesses that have limits on cloud spending. Often, line-of-business developers want to leverage generative AI systems for legitimate business reasons. However, as stated earlier, they come at a high cost, so companies must find funding, commercial motivation, or both.
Many times, although generative artificial intelligence is the technology used by today’s so-called “cool kids”, it is usually not cost-effective. Generative AI is sometimes used for simple tactical tasks that are comparable to more traditional development methods. The overuse of AI has been an ongoing problem since its inception; the reality is that the technology is only applicable to certain business problems. But it's popular, hyped, and therefore overused.
These issues illustrate the need for more experience as the technology matures. However, this could have a negative impact on cloud operations, as it did when the cloud first took off.
Original link://m.sbmmt.com/link/a26475af783877529bf81eed81743d71
The above is the detailed content of Three major challenges of generative AI to cloud operation and maintenance. For more information, please follow other related articles on the PHP Chinese website!
![[Web front-end] Node.js quick start](https://img.php.cn/upload/course/000/000/067/662b5d34ba7c0227.png)
