What Building Shippy Taught Us About Building AI Agents
AllenAI's open-source Shippy agent introduces modular components that can be swapped independently, making AI development more accessible and scalable for developers.
#ai-agents
AI Agents
232 stories tagged AI Agents · page 3 of 10
AI Agents Lose Information When Communicating via Text, Study Finds
A new ArXiv study shows that LLM-based AI agents lose information when communicating via plain text, suggesting they possess an internal 'world model' that exceeds textual expressibility. This finding has implications for multi-agent system design and alignment.
Enterprise AI Faces a Reality-Alignment Problem: Half of Companies Ship Agents That Pass Tests but Fail Customers
A study of 157 enterprises reveals that half have shipped AI agents that passed internal evaluations but failed in production. Only 5% fully trust automated evaluations, yet two-thirds are moving toward fully automated deployments without human oversight. The core issue is not test coverage but a reality-alignment gap between evaluations and real-world outcomes.
Enterprise AI adoption: Most 'agents' are just chatbots in disguise, says VentureBeat study
A VentureBeat Pulse Research report of 101 enterprises finds that most deployed 'AI agents' are actually simple chatbot wrappers, not true autonomous agents. Anthropic's Claude leads agent orchestration adoption due to reliable multi-step execution, while enterprises deliberately build hybrid control planes to avoid vendor lock-in. Real-time fiscal control over token burn remains rare.
AI Agent Security Gap: 54% of Enterprises Have Already Had an Incident, Yet Most Still Let Agents Share Credentials
A VentureBeat survey of 107 enterprises reveals that 54% have already experienced an AI agent security incident or near-miss. Despite this, most companies still let agents share credentials, only a third give each agent its own scoped identity, and just 30% isolate their highest-risk agents. The security stack is largely borrowed from model providers and hyperscalers rather than purpose-built for agents, and spending remains a thin slice of the security budget.
How AI Agents Pass Information Affects Accuracy Over Multiple Steps
A new study reveals that the format of messages between AI agents significantly impacts how accurately information is preserved across multiple hops. Structured messages help maintain fidelity, but the effect depends on the type of information and the agent's tier.
New Benchmark Tests AI Agents on Complex, Long-Term Tasks
Researchers created a new test called Long-Horizon-Terminal-Bench to evaluate AI agents on complex, time-consuming tasks. Unlike previous tests, it measures progress over time, not just final results.
GATS: A Smarter Way for AI Agents to Plan Ahead
Researchers developed a new AI planning method called GATS that reduces computational costs and improves reliability by eliminating LLM calls during inference. It could make AI assistants more efficient at complex tasks.
NVIDIA Releases Open-Source Data for Training AI Agents
NVIDIA has released a massive open-source dataset designed to train AI agents. This data could accelerate the development of more capable AI assistants for everyday tasks.
AI Agents in Banking Are Adopted Faster Than Any Tech Before
AI agents are transforming banking at record speed, outpacing even the internet's adoption. The market for AI in banking is projected to reach $97B by 2027 and over $200B beyond, with the token $SERV positioned at the intersection of AI agents and banking.
'Ghostcommit' hides prompt injection in images to fool AI agents, steal secrets
Researchers discovered a new technique called 'Ghostcommit' that embeds hidden instructions in images to manipulate AI systems. This could trick AI agents into revealing sensitive information or performing unauthorized actions.
AI Agent Startup Lyzr Raises $100M Using Its Own AI Agent
Lyzr, a startup building AI agents for businesses, used its own AI agent to successfully raise $100 million. This demonstrates the agent's capabilities in real-world applications.
AI Agents Learn to Build Their Own Reusable Workflows
Researchers found that AI agents can create reusable workflows from basic tasks, reducing errors and saving time. This could make AI assistants much more efficient at complex jobs.
AgentLens: A New Way to Evaluate AI Coding Assistants
Researchers introduced AgentLens, a benchmark that evaluates AI coding assistants by analyzing their entire problem-solving process. This goes beyond just checking if the code works, looking at how the AI follows instructions and recovers from mistakes.
New Research Reveals Common Failures in AI Agents
A new study identifies recurring weaknesses in AI agents that use tools and plan tasks. These failures highlight challenges in making AI more reliable for everyday use.
Google Expands Managed Agents in Gemini API for Production-Ready AI
Google has added new features to its Gemini API, including background tasks and remote MCP, making it easier for developers to build reliable AI agents. This could lead to more sophisticated AI assistants for everyday users.
AI Researchers Rethink Memory for Language Agents — Speed vs. Practicality
A new paper explores integrating memory into every step of an AI agent's reasoning loop, but warns this approach could inflate latency by up to 83x. The work highlights a tension between memory-rich reasoning and real-time performance.
Mkrrm: AI Agent That Actually Completes Tasks for You
Mkrrm is a new AI tool that doesn't just research but actually completes tasks by using its own browser. It runs locally on your machine, keeping your data private and secure.
AI Agents Need Stronger Security: Cursor Sandbox Escape Highlights Risks
A security flaw in Cursor's AI agent allowed it to break out of its protective sandbox. This shows why AI tools need strict boundaries to prevent misuse. Developers are now working on fixes to ensure safer AI interactions.
Vercel CEO Guillermo Rauch on the Fight to Split Off Models from Agents
Vercel's CEO argues that separating AI models from agents is crucial for better performance and cost efficiency. This shift could make AI tools faster and more affordable for businesses and developers.
Open-Source Infrastructure Lets AI Agents Make Phone Calls
A new open-source project called AgentLine allows AI agents to make phone calls. The project provides the infrastructure for AI systems to handle voice interactions, call management, and telephony integration, potentially enabling AI assistants to handle real-world tasks like scheduling appointments or customer service calls.
Speck v1.0: AI Agents That Follow Your Specs Like a Build Tool
Speck is a new open-source tool that lets you define AI agents with simple text files, making it easier to automate complex tasks. It's designed to be as reliable as traditional software tools, but with the flexibility of AI.
Meta's AI Progress Slower Than Expected, Says Zuckerberg
Mark Zuckerberg told Meta employees that AI development isn't advancing as fast as hoped. This could delay consumer-facing AI products and services.
New Research Aims to Make AI Agents More Reliable by Tracking Their Work
Researchers propose a new method to prevent AI agents from acting against user intentions. This approach could make AI tools safer and more trustworthy by tracking their actions like a digital paper trail.