Key takeaways of zdnet
- Google is introducing powerful techniques for agents and data.
- They are also offering a series of data-centered agents.
- A new command-line AI coding tool is now available.
I am not a stranger for hyperbolic claims from tech companies. Anyone who receives AI -related press announcements at the end of firehos understands. Everything is game-changing, world-changing, most, best, more, more.
And then Google is. There is no stranger for Google Hyperbole. But when a company is immersed in data management as part of its core DNA, talks about “fundamental changes”, and says that the world is changing because, “It is being engineered again in real -time by data and AI,” we can consider those claims quite reliable.
Also: Got 6 hours? This free AI training from Google can promote your resume today
Google Cloud is making a series of Google announcements for the next Tokyo 2025, which is a major change in managing data to enterprises.
Yasmin Ahmed, Managing Director of Data Cloud of Google, says in one blog post“The way we interact with data, it is undergoing a fundamental change that is moving forward to a collaborative partnership with intelligent agents beyond human analysis.”
She calls this agent shift, which she explains, “A new era where special AI agents work autonomously and cooperatively, to unlock the insight on a scale and the speed that was first unimaginable.”
Too: 5 ways to successfully integrate AI agents in your workplace
From almost any other company, such claims will look like just such a hot air. But Google is leaving a series of announcements about new offerings that provide the real world ability to data scientists and engineers in very tangible support of claims.
Agent shift
There is a fairly good line between AI chatbots and AI agents. The chatbott is connivable, while agents are devices that do autonomous functions. Some users appoint chatbots to perform tasks, as I did when I used Chatgpt to analyze some commercial data. The agents, like the chatgate agent, use a conjunctive interface to obtain instructions.
A good way to think about agents is as members of the surrogate team. Perhaps one agent performs data generalization (cleaning of data), while the other migration. Each agent performs one or more defined function using AI capabilities.
Too: Do AI agents want to work together? Linux Foundation has a plan
In this context, Google is looking at agents that can automate and simplify tasks for data workers, communicate with each other, and free professionals from tedious work so that they can focus on “high-value tasks”. Google is also trying to get agents to work together in virtual teams.
Of course, there are questions about whether agents are not really free from the time of senior professionals, but instead are overcome by more junior employees. On the other hand, when I am completely immersed in a project, I have no one to do the work of grunt. So I can hand over anything to an agent which is more time for projects and writing.
Cognitive basis
With all these agents walk around, traditional databases are not just dependent on the task of feeding them. Agents perform their logic or automation processes in Silos. They require access to both historical and live data.
Classic data management methods such as Real-Time OLTP (online transaction processing) and Deep-Dive OLAP (online Analytical Processing) make the data very differently different to get insight from trends and current activities.
A way to help unite all these abilities is to increase their database offerings. A few years ago, Google added a column engine for alloys. Alloydb Google Cloud Platform has a fully managed database service of the company that focuses on postgresql users, ideal for those who require postgresql-specific solutions.
Too: How AI agents can generate $ 450 billion by 2028 – and what stands on the way
A column engine is the one where queries of specific columns of workload data, only read the field required for analysis. This increases rapidly and allows the vector for execution, where the operation is applied to a whole column of data at a time.
Now, Google Spanner, its globally distributed, firmly added a column engine to the consistent database service provides high availability and scalability, designed for enterprises required by global access and high transaction integrity.
It is designed for business agility in BigQuery, Google’s server-free, highly scalable and cost-effective multi-cloud data warehouse. As the name means, Bigkwear is ideal for those who need to run queries like SQL on a large dataset.
The company says that this new column capacity in spanner speeds up analytical questions on live transactional data by something like 200X. With such a performance, we are talking about immediate accountability for real -time situations.
When creating an enterprise-based AI system, you need agents to make decisions based on real data. To take real -time action based on hallucination data, it can be ugly very quickly. This is the place where RAG (reconstructed generation) comes. Essentially, RAG combines large language models with real -time data access.
Too: According to business leaders, 5 ways to be a great AI agent manager
You can start to see how vectoring search in spanner and Bigkweari becomes necessary when you feed in real-time figures along with historical information. But it has been traditionally painful to search for a vector to work efficiently. Google is automatically adding adaptive filtering to alloydbi to maintain vector index and optimize rapid questions on live operational data.
Google is also presenting the autonomous vector embeding and generation in BigQuery, which automatically prepares and actually prepares multimodal data for vector discovery. This is an important step in creating a type of semantic memory for agents.
The company is also introducing the ability to run AI Query inside Bigkwear. This is important, big. Now, bigQuery users may have AI that he can do his magic in the giant Gobs of structured and unnecessary data, ask complex questions (such as “such as” who are disappointed? “Such as subjectives), and get a direct answer within the existing analytics tool.
New agent capacity
In addition to creating a foundation for agentic cooperation and data access, Google is announcing a series of new capabilities that embed the agents in their largest data tools. Let’s look at each in turn.
Data Engineering Agent: Especially manufactured for data engineers, this agent can simplify and automate complex data pipelines within BigQuery. The entire workflow can be powered by the indications of the natural language, from data ingestion to generalization in data-quality evaluation.
Spanner migration agent: Relating to data engineering agent, Spanish migration agent can simplify data migration from Legacy System to Bigkwear. Such migration is generally exceptionally tedious and potentially dangerous, but now agents can do most heavy lifting.
Data Science Agent: Data scientists focus on analyzing and interpreting complex data, while data engineers focus on data infrastructure. According to Google, the new data science agent triggers the entire autonomous analytical workflows “, including searching data analysis, data cleaning, features, machine-learning predictions, and more, and more, and more, and more, and more. It makes a plan, executes the code, presents the reasons about the results, and presents your conclusions, while allowing you to support you, while supporting you, and presents your conclusions. Is.”
Code interpreter: Made as an increase of the intellectual analytics agent introduced last year, the code takes into the interrelation business-analysis questions and converts them into a python code to prepare custom analysis for users. It runs within all Google data clouds and uses Google Data Cloud Security Infrastructure. It also includes APIs available for developers, which are available to include convergent analytics agents and code interpreters in the custom code.
New command-line coding equipment
As part of this large range of announcements, Google is starting an extension called Mithun CLI Gemini Cly Githib Action,
CLI means command line interface, which is basically a terminal interface for your computer. Even though most of the users left behind the terminal when MS-DOS migrated to Windows, uses the Kodar Command Line to date. Working in terminal mode add coders tools and control the coding process very fast, when they have to find and choose items from the menu and icon.
Too: Bad vibes: How an AI agent makes his way for disaster
Last month, when Google introduced Gemini CLI, it originally created the characteristics of Mithun Chatbot available in the terminal. Now, Google has increased the capacity that provides some agentic features within the terminal environment.
Some of you must be wondering how it compares with Jules, I wrote about the Google coding agent in May. First, Jules works in a safe cloud VM, while Gemini CLI moves into the action terminal and is integrated with the Zeethb action (zethab-based workflow tool).
Google says that Gemini CLI is quite narrow scope for action compared to Jules. Jules can read your entire codebase, plan and present an approach to a coding challenge, and then execute it on it. Gemini CLI GITHUB actions are targeted specifically intelligent issues triaz, quick bridge-infinite reviews and on-demand cooperation.
Too: Most developers use AI in their daily workflows – but they do not rely on it, get studies.
Problem-Triies capacity helps coders manage specific bug reports and convenience requests. Bridge request is the way Github asks to confirm the coders to integrate coding changes in branches and master codebase. Whenever you want to talk about your code, on-demand cooperation is essentially establishing a chat session.
I could easily see the use of both a programmer. Jules will be great for large projects and large swings, and Gemini CLI Githib actions will work well for quick updates and fix.
Are the agents a game-changer?
What do you think about agent shift promoting Google? Have you started integrating intelligent agents in your own workflows? Which one of the new data tools or abilities of Google likes you the most-Detta Engineering Agent, In-Kuery AI argument, or anything else? Do you see agents as a place of junior roles or both to help senior professionals? And how do you feel about running AI workflow directly in Bigkwari? Let us know in the comments below.
Too: AWS VP says AI agents will change the work and society in the way of internet size
Want more stories about AI? Check out AI LeaderboardOur weekly newspapers.
You can follow my day-to-day project updates on social media. Be sure to subscribe to My weekly update newsletterAnd follow me on Twitter/X @DavidgewirtzOn Facebook Facebook.com/davidgewirtzOn Instagram Instagram.com/davidgewirtzOn blue @Davidgewirtz.comAnd on youtube Youtube.com/davidgewirtztv,
