Finding vulnerabilities in modern web apps using Claude Code and OpenAI Codex | Semgrep," focuses on a security research experiment conducted by Semgrep to assess the effectiveness of AI Coding Agents, specifically Anthropic's Claude Code and OpenAI Codex, in identifying vulnerabilities within real-world web applications. The research highlights that while these AI tools can find genuine security flaws, they suffer from high false positive rates and significant non-determinism, meaning they produce inconsistent results with repeated scans. Semgrep also details its comprehensive security platform, which offers various tools like static application security testing (SAST), software supply chain analysis (SCA), and secrets detection, aiming to provide more reliable and consistent code security solutions.

Send us a text

Support the show

Podcast:
https://kabir.buzzsprout.com


YouTube:
https://www.youtube.com/@kabirtechdives

Please subscribe and share.

This episode analyzes a report titled, "State of AI in Business 2025", from MIT NANDA, explores the surprising "GenAI Divide", revealing that despite significant investment, 95% of organizations see no return on their GenAI initiatives. The research, based on a multi-method design including a review of over 300 AI initiatives, interviews, and surveys, highlights that while general-purpose tools like ChatGPT are widely adopted for individual productivity, enterprise-grade solutions often fail due to poor integration, lack of contextual learning, and misalignment with workflows. The report suggests that success in GenAI hinges on adopting learning-capable, deeply customized systems, focusing on strategic partnerships over internal builds, and prioritizing back-office automation for higher ROI, all while acknowledging the emerging "shadow AI economy" of employee-driven tool usage. Ultimately, the document concludes that organizations must move towards "agentic systems" that learn and adapt to successfully bridge this divide and transition into an "Agentic Web" where autonomous systems coordinate actions across the internet.

Send us a text

Support the show

Podcast:
https://kabir.buzzsprout.com


YouTube:
https://www.youtube.com/@kabirtechdives

Please subscribe and share.

Weekly Review of AI

Send us a text

Support the show

Podcast:
https://kabir.buzzsprout.com


YouTube:
https://www.youtube.com/@kabirtechdives

Please subscribe and share.

This episode reviews AI hot news till August 16th, 2025.

Send us a text

Support the show

Podcast:
https://kabir.buzzsprout.com


YouTube:
https://www.youtube.com/@kabirtechdives

Please subscribe and share.

This episode primarily discusses the recent release and capabilities of OpenAI's GPT-5 model, contrasting it with Anthropic's Claude Opus 4.1 and earlier AI versions. They offer a comparative analysis focusing on coding performance, multimodal understanding, agentic functionality, and pricing structures for developers and general users. While GPT-5 is highlighted for its unified architecture, reduced hallucinations, and cost-effectiveness for versatile tasks, Claude Opus 4.1 is often praised for its superior precision in complex coding, especially with niche or multi-file projects, despite its higher cost. The texts reveal a split community sentiment, with users often choosing between the models based on their specific needs for speed, accuracy, or specialized development environments, underscoring an evolving, highly competitive AI landscape.

Send us a text

Support the show

Podcast:
https://kabir.buzzsprout.com


YouTube:
https://www.youtube.com/@kabirtechdives

Please subscribe and share.

This epislode primarily critique the recently launched GPT-5, highlighting widespread user dissatisfaction. Many users, especially on Reddit, report the new model as "horrible," citing issues like inability to perform basic math, poor image analysis, slow and unhelpful responses, and a lack of the "personality" or creative flexibility found in older versions like 4o and 4.1. Some speculate these perceived downgrades are due to cost-cutting measures or problematic internal "routing" of queries to less capable models on the ChatGPT website, rather than the API. Furthermore, concerns are raised regarding GPT-5's significantly increased energy consumption and the potential for a discrepancy between OpenAI's ambitious claims about AGI and the actual performance and utility of their latest model, suggesting it may not be the "paradigm shift" many anticipated for the AI industry.

Send us a text

Support the show

Podcast:
https://kabir.buzzsprout.com


YouTube:
https://www.youtube.com/@kabirtechdives

Please subscribe and share.

This episode centers around the launch of OpenAI's GPT-5, a new large language model. The Decrypt article announces its public release, highlighting its availability to all users and new features like video options and business integrations, while also providing a list of cryptocurrency prices. Wikipedia offers a concise overview of GPT-5's capabilities, launch date, and technical specifications, noting its "PhD-level" abilities. Finally, Vellum AI provides detailed benchmark comparisons, demonstrating GPT-5's superior performance in areas like math, reasoning, coding, and reliability against predecessor models and competitors, solidifying its position as a leading AI model.

Send us a text

Support the show

Podcast:
https://kabir.buzzsprout.com


YouTube:
https://www.youtube.com/@kabirtechdives

Please subscribe and share.

This episode discusses hierarchical decision-making across various fields, from computational models to biological systems and robotics. "Hierarchical Decision Making" explores how context-dependent decisions can be structured in a hierarchy, utilizing machine learning and reinforcement learning to enhance human operator effectiveness. Complementing this, "Hierarchical reasoning by neural circuits in the frontal cortex" investigates the neurological underpinnings of such processes, identifying specific brain regions involved in multi-timescale decision-making in primates. Finally, "Multi-Level Reasoning for Delicate Assembly using Dual Arms" demonstrates the practical application of hierarchical reasoning in robotics, showcasing how complex multi-robot assembly tasks, like building with LEGOs, benefit from physics-aware planning and asynchronous execution within a hierarchical framework.

Send us a text

Support the show

Podcast:
https://kabir.buzzsprout.com


YouTube:
https://www.youtube.com/@kabirtechdives

Please subscribe and share.