Traditional wisdom believes that you need to fix your data management shortages before success with AI. But according to some tech executives, it is no longer true, who look at the ability to understand the language and implement the ability of generative AI to fix the language and fix data management issues.
Vice President of Statistics Rahul Pathak goes to the market in AWS, thinking of himself as an old school data boy, which is of a kind that will never recommend a shortcut to show success on paper. So when he suggests that Genni can allow you to move forward in your data management capabilities and get results faster, you would like to take notice.
Pathak says, “We were in a world where you have to serials through it, where you have to meet the data house in order, then you have to make the app that is sitting on the data.” “I think you can really change this process a bit, where you can start unlocking your data for AI, almost Immediately immediately, using the state of well -known MCP locations and art models. [They] Can really help you unlock data that can then help you illuminate AI applications.
Obviously, not all AI use matters are the same. In some cases of use may need to collect, clear and prepare data before touching the AI algorithm. But when it comes to operating workloads on pre -trained models, stage of the data is unlikely, in which case the federated approach will be in order. The good news is that the model context (MCP) covers many sins of data that may require plenty of atonement (not mentioning data management pain and dollars).
“I am an old school person at the moment, but you can think of an MCP server as is like a fed up,” said Pathak. “The model allows you to get data. It is somewhat flexible. And then the basis of knowledge and index is almost like a substance view. And likewise, you can get data very fast. And in the models, intelligence data engineers and data scientists enhance their ability to move faster than before.”
In the form of a manufacturing company, Pathak was an example of a real world that wanted to use Generative AI to accelerate production. The company had already collected telemetry data, but it was difficult and time to get knowledge from the telemetry data to apply to the factory line.
The solution was to use Genai’s natural language processing (NLP) capabilities to remove appropriate pieces of data out of telemetry data. Subsequently, these insights were fed in the traditional machine learning optimization model. At the previous end, Genai was once again used to prepare instructions in which operators were told how to edit the production of their process to accelerate production.
(IM Imagree/Shutter Stock)
“This is the kind of integration that allows us to move faster than ever,” said Pathak. “Because otherwise you will have a big data and a type of ETL and data -like project that you have to do to make this telemetory quick and useful. And we can now work very fast, very fast. So it’s a huge unlock.”
Another supporter is to leave the Big Data Management Project and jump straight into the Geni projects. The company developed a genius -based inquiry device that allows users to immediately inquire of their data, without going through the demand process at the time of construction of a meaningful layer.
People with the prompt QL say that a spiritual layer is still important, as it works to translate a business specific terms and matrix into technical table names, which the device needs to offer the right questions. But the biggest difference is that the prompt QL supports the construction of the cementic layer when you leave, and over time it supports it. He says spending months or years on the big bang data management project is the path to endless POC and eventually failure.
The high failure rate of early AI projects is elephant in the room. Recent MIT studies, which have found that 95 % of live projects never get out of the test phase. Investing of trillions of dollars is being invested in the acquisition of high -speed GPUs, large -scale storage arrows, and AI’s major data centers, with some very wealthy companies being larger on AI.
Small resource companies have to be more careful about how they attack on the AI. The good news is that the abilities to understand the language of Ginai can be used in many ways, including how to understand how data is made, which can potentially allow you to leave the data management phase, at least to deal with it.
“These are no longer settling measures,” Pathak says. “I think this is a great example for many companies that are dealing with the challenges of heritage data, which clearly, we have been dealt with since we have had more than one table in the database.”
He says, “I think AI has done and what is different now.”
This article first appeared on our sister’s publication, HPCWIRE.
Relevant
