Generative AI Insights, an InfoWorld blog open to outside contributors, provides a venue for technology leaders to explore and discuss the challenges and opportunities presented by generative artificial intelligence. The selection is wide-ranging, from technology deep dives to case studies to expert opinion, but also subjective, based on our judgment of which topics and treatments will best serve InfoWorld’s technically sophisticated audience. InfoWorld does not accept marketing collateral for publication and reserves the right to edit all contributed content.
RamaLama makes it a snap to spin up AI models locally in containers and streamlines the path from experimentation to production.
The governance journeys of SaaS and Web2 tell us that today’s ad hoc AI governance will give way to a continuous and automated approach.
A glimpse at how DeepSeek achieved its V3 and R1 breakthroughs, and how organizations can take advantage of model innovations when they emerge so quickly.
Building generative user interfaces into your applications can be an effective way to deliver better user experiences. It means orchestrating fully interactive responses.
Deep neural networks have hit a wall. An entirely new, backpropagation-free AI stack promises to be orders of magnitude more performant.
To unlock the full potential of AI and machine learning, understand the keys to model selection, optimization, monitoring, scaling, and metrics for success.
Hype can be excessive and obnoxious and lead to waste and false promises. It is also a crucial catalyst of experimentation and innovation.
With AI observability, we can guard against hallucinations, catch irrelevant and incomplete responses, and identify security lapses in generative AI applications — ensuring they meet the needs of the business.
How event-driven design can overcome the challenges of coordinating multiple AI agents to create scalable and efficient reasoning systems.
Why relying on retrieval-augmented generation and prompt engineering is preferable to investing in model training and fine-tuning.
While retrieval-augmented generation is effective for simpler queries, advanced reasoning questions require deeper connections between information that exist across documents. They require a knowledge graph.
Before deploying agentic AI, enterprises should be prepared to address several issues that could impact the trustworthiness and security of the system.
AI red teaming offers an innovative, proactive method for strengthening AI while mitigating potential risks, helping organizations avoid costly AI incidents. Here’s how it works.
Developers are tired of hearing about AI as a panacea. The backlash may be just what organizations need to effectively implement the technology.
Small language models shine for domain-specific or specialized use cases, while making it easier for enterprises to balance performance, cost, and security concerns.
1.7 million AI chats in 30 days... and seven big lessons Campfire learned about building better AI chat products.
What’s the best way to store, search, and analyze content not based on their technical characteristics but on their meaning?
How Gencore AI enables the construction of production-ready generative AI pipelines using any data system, vector database, AI model, and prompt endpoint.
How can enterprises secure and manage the expanding ecosystem of AI applications that touch sensitive business data? Start with a governance framework.
Failed AI projects waste time and resources, damage reputations, and stifle innovation. To succeed with AI, put the necessary practices in place to ensure high-quality data.
The potential for generative AI to deliver a significant return on investment is being demonstrated by early adopters across various industries. My company provides one example.
By giving developers the freedom to explore AI, organizations can remodel the developer role and equip their teams for the future.
Python developers are uniquely positioned to succeed in the AI era, with a little help from upskilling.
SB 1047 missed the mark. A far better solution to managing AI risks would be a unified federal regulatory approach that is adaptable, practical, and focused on real-world threats.
How high-quality, synthetically designed data sets enable the development of specialized AI models.
Five of the most common and complex challenges organizations face in putting large language models into production and how to tackle them.
Combining knowledge graphs with retrieval-augmented generation can improve the accuracy of your generative AI application, and generally can be done using your existing database.
LLMs are powering breakthroughs and efficiencies across industries. When choosing a model, enterprises should consider its intended application, speed, security, cost, language, and ease of use.
Once you get your retrieval-augmented generation system working effectively, you may face new challenges in scalability, user experience, and operational overhead.
Both the US and the EU have mandated a risk-based approach to AI development. Whatever your risk level, ultimately it’s all about transparency and security.
Understanding the lumpy pattern of technological evolution is essential for organizations that want to make informed decisions about when to invest in and adopt new technologies.
Any leading large language model will do. To succeed with retrieval-augmented generation, focus on optimizing the retrieval model and ensuring high-quality data.
A modern AI-enabled iPaaS solution that supports collaborative workflow design and management can break down silos between IT and business teams and propel automation initiatives forward.
For IT admins, engineers, and architects, language models will save time and frustration and increase confidence in troubleshooting, configuration, and many other tasks. Here are six ways they’ll make operations easier.
With the rise of AI-generated code, development teams must become smarter about how they conduct code reviews, apply security tests, and automate their testing.
Just as AI-powered programming assistants make developers more productive, AI will streamline workflows for data analysts. It will also bring vast benefits to business users.
By integrating domain-specific data, RAG ensures that the answers of generative AI systems are richly informed and precisely tailored. More sophisticated techniques are on the horizon.
Key to the success of any large organization is effective governance of a vast, distributed landscape of data stores. AI can help.
It’s no longer how good your model is, it’s how good your data is. Why privacy-preserving synthetic data is key to scaling AI.
Successful integration of AI into daily operations hinges on front-line employees, yet the impact on their morale is often overlooked.
Balancing performance, energy efficiency, and cost-effectiveness, CPUs adeptly handle the less-intensive inference tasks that make up the lion’s share of AI workloads.
The average user of AI lacks an adequate understanding of the tools they increasingly depend on for decision-making and work. We need to change that.
Fine-tuning and retrieval augmentation are time-consuming and expensive. A better way of specializing LLMs is on the horizon.
Generative AI not only makes analytics tools easier to use, but also substantially improves the quality of automation that can be applied across the data analytics life cycle.
Responsible AI isn’t really about principles, or ethics, or explainability. It can be the key to unlocking AI value at scale, but we need to shatter some myths first.
Through natural language queries and graph-based RAG, TigerGraph CoPilot addresses the complex challenges of data analysis and the serious shortcomings of LLMs for business applications.
Retrieval-augmented generation brings to generative AI the one big thing that was holding it back in the enterprise.
The key to reaping the benefits of AI while minimizing the risks is through responsible development and use. Here’s how SAS Viya puts ethical AI practices to work.
Hardware requirements vary for machine learning and other compute-intensive workloads. Get to know these GPU specs and Nvidia GPU models.