The combination of generative AI and cloud brings both opportunities and challenges

In the era of rapid development of generative artificial intelligence, no one doubts that artificial intelligence has become a mainstream trend, and there is no need to doubt the changes that AI will bring to the world. However, when enterprises think about the sparks that will result from the collision of AI and cloud computing, they must first think of a practical problem, that is, deploying too many applications will cause expansion problems and lead to cost overruns.

Although the application of artificial intelligence technology with generative AI as the core can bring benefits to enterprises, there are also some problems. We must consider it comprehensively and consider the pros and cons. Compared with the rapid deployment of generative AI, it is crucial to comprehensively think about how to effectively manage the application of these new technologies so that technological innovation will not have a negative impact on the enterprise.

Specifically, generative AI will encounter three problems when running in the cloud:

1. Accelerate cloud application deployment

This is the first misunderstanding. In the current situation, we can quickly create applications using no-code or low-code mechanisms with the help of generative AI development tools. But as the number of deployed applications increases, it's easy for enterprises to lose control.

Of course, in the general direction, we very much agree with this technology trend. There is no doubt that generative AI plays an important role in accelerating application deployment, meeting business needs and improving efficiency. Because many applications developed in the 1990s and early 2000s were not satisfactory and limited business development to some extent. Any improvement methods are good for business!

Only sometimes, we see an almost reckless approach to application development, where the work required to build and deploy these systems takes only days, and sometimes even hours. Companies don't put much thought into the overall role of applications, and many are purpose-built for tactical needs and are often redundant. They need to manage three to five times as many applications and connected databases as they should. Not only will the whole mess not scale, it will also keep costs high.

2. Reasonable use of resources

Generative AI requires a lot of computing and storage resources, certainly much more than currently used by enterprises. Turning on more storage and computing services does not simply drive larger scale expansion, but also requires full utilization of these resources.

Thinking and planning must be done into resource sourcing and deployment to support the rapidly expanding use of generative AI. This often falls on the shoulders of the operations team to deploy the right amount of resources in the right way without destroying the value of these systems or limiting their functionality. The entire process is a trade-off that does not happen overnight.

3. Cost overrun

As enterprises focus on deploying specialized systems to monitor and manage cloud costs, we can observe a significant increase in funding to support generative AI. At this time, what should the company do?

This is a business issue, not a technical issue. Businesses need to understand why cloud spending is happening, why it's happening, and the commercial benefits to the business. The costs can then be included in the predefined budget.

For enterprises limiting cloud spending, this is a starting point. Line-of-business developers want to leverage generative AI, often for business reasons. Although the high computing and storage costs of generative AI have been explained above, companies still need to ensure business value and raise funds.

Although generative AI performs well in many situations, it is often still in its basic stages and lacks reasonable cost assessment. Generative AI can be applied to simple tactical tasks in some situations where traditional development methods are equally feasible. This overuse has been an ongoing topic since the inception of artificial intelligence. The reality is that this technique only works for certain business problems. The current situation is that generative AI has become very popular due to widespread publicity and overuse.

It is necessary for enterprises to think more deeply about implementation plans when the AI ​​generation technology is mature. During this period, if cloud support cannot keep up, it may bring negative effects.

The above is the detailed content of The combination of generative AI and cloud brings both opportunities and challenges. For more information, please follow other related articles on the PHP Chinese website!