LogicStar AI, co-founded by DeepCode creator Boris Paskalev, recently launched a platform designed to autonomously investigate, reproduce, fix and validate software bugs — all before humans intervene. The company claims this autonomous workflow can reduce the bug life cycle by 95% and frees developers from “the most hated part” of their job. Technology journalist Meredith Shubel explains how the platform works.
In other tech stories we’re covering today:
Large language models (LLMs) can be a bit persnickety about implementing frameworks. Some are better at particular frameworks than others, but all are flawed in some way when it comes to producing frontend code that follows best practices. To solve the problem, the Google Angular team developed and open-sourced Web Codegen Scorer, which provides quantifiable quality metrics for AI-generated code. Before you vibe code your frontend, find out how the tool assesses code accessibility and security, while incorporating framework best practices.
TNS Senior Editor Joab Jackson is at the Current event this week in New Orleans, and today he brings news of Confluent’s new offering, Confluent Intelligence. It’s designed to address a major bottleneck in enterprise AI: continuous real-time data. Joab reports that the new platform features a real-time context engine that streamlines and packages data from Apache Kafka and Flink. He explores how the platform transorms the data to ensure AI agents can efficiently consume it via the Model Context Protocol.
Simply "shifting left" is no longer sufficient to protect your software supply chain. Learn essential patterns for securing regulated environments.
Organizations running their business on open source software are faced with a more aggressive and complicated security and compliance landscape than ever before. Malicious actors are bypassing transitional security tools by directly targeting developers, according to Sonatype’s10th annual “State of the Software...
Missed our recent live session on solving the Kubernetes resource management challenge — or just want to dive back into the insights? Great news! The full webinar is now available on-demand, so you can watch (or rewatch) all the expert discussions at your convenience.
Think the newest LLM is the best for coding? Sonar put leading models to the test and unveiled insights on GPT-5, Claude Sonnet 4, and Llama 3 in their “The Coding Personalities of Leading LLMs” report. The results may surprise you.
TNS Agents: How Agentgateway Solves Agentic AI’s Connectivity Challenges
Agentgateway is an AI native proxy designed to help users control security and governance of their agents and tools, said Lin Sun of Solo.io in this episode of The New Stack Agents.
