October 1, 2025

news  
Zhang Peng Skeptical of 2030 Superintelligence, Zhipu AI CEO

The rapid integration of Artificial Intelligence across various sectors is prompting a significant evolution in security measures and identity management. Companies like Wing Security are pivoting to an AI-first approach, expanding their platforms to secure not only SaaS applications but also AI tools and third-party integrations. This involves offering comprehensive visibility, AI-specific security controls, and real-time threat detection to manage risks associated with uncontrolled AI tool use and data exposure. Similarly, Permiso is enhancing its identity security platform to encompass AI users, builders, and agents, utilizing runtime intelligence to track usage patterns and identify vulnerabilities. Entro Security and SailPoint are also introducing solutions focused on securing AI agents, mapping them to non-human identities, and governing their access to data and credentials. Meanwhile, in the business and product development sphere, companies are leveraging AI to streamline operations. BullFrog AI is expanding its sales team to accelerate the adoption of its AI-powered clinical trial optimization tools, including the bfPREP module, which standardizes clinical trial data. Kolmar Korea has launched an AI platform, Kolmarpkg.com, to expedite the sourcing of beauty product packaging, reducing timelines from six months to one month. Shortcut has introduced Korey, an AI project manager designed to automate software development tasks from planning to progress summarization. Hance is set to demo its compact AI audio processing software at TechCrunch Disrupt. In a different application, Donald Trump's team has utilized AI to generate viral portraits and, controversially, a debunked 'MedBed' video. On the predictive analytics front, an AI model forecasts Bengals quarterback Jake Browning to exceed 219.5 passing yards. In the realm of AI development and its future, Zhang Peng, CEO of China's Zhipu AI, expressed skepticism about achieving full artificial superintelligence by 2030, suggesting it will likely remain limited in scope. Finally, the Base network's NFTs have seen a surge in sales, largely driven by the AI trading game DX Terminal, where NFTs function as AI trader agents.

Key Takeaways

  • Wing Security is expanding its platform to become an AI-first security company, focusing on securing SaaS, third-party integrations, and AI tools with enhanced visibility and AI-specific controls.
  • Permiso has updated its identity security platform to include AI users, builders, and agents, using runtime intelligence to monitor AI usage patterns and identify vulnerabilities.
  • Entro Security and SailPoint are introducing solutions to secure AI agents by providing observability, mapping them to non-human identities, and governing their access and credentials.
  • BullFrog AI is growing its sales team to accelerate the commercialization of its AI-driven clinical trial optimization tools, such as the bfPREP module.
  • Kolmar Korea launched Kolmarpkg.com, an AI platform that significantly reduces beauty product packaging sourcing time from six months to one month.
  • Shortcut introduced Korey, an AI project manager designed to automate software development tasks from planning to progress summarization.
  • Donald Trump's team has used AI for generating viral portraits and, controversially, a debunked 'MedBed' video.
  • Zhang Peng, CEO of Zhipu AI, believes that full artificial superintelligence surpassing human capabilities across the board is unlikely by 2030.
  • NFTs on the Base network experienced high sales volume in September, primarily driven by the AI trading game DX Terminal.
  • Senator Ted Cruz introduced the SANDBOX Act, aiming to establish a framework for regulating AI development and deployment.

New AI Security Needed for Evolving Enterprise Defenses

Companies are rapidly adopting AI, but this creates new risks like uncontrolled AI tool use and data exposure. Traditional security methods are not enough for this fast-changing landscape. A new security approach is needed that includes continuous discovery, real-time monitoring, and risk assessment. Wing Security offers a platform to manage AI risks, providing visibility and control over the AI application supply chain to enable safe innovation.

Wing Security Becomes AI-Centric, Securing SaaS and AI Apps

Wing Security is now an AI-first security company, expanding its platform to protect SaaS, third-party integrations, and AI tools. The platform offers complete visibility across applications, AI-specific security controls, and real-time threat detection. This evolution builds on their existing SaaS Security Posture Management (SSPM) capabilities. Wing Security aims to help organizations govern AI usage and detect threats without slowing down innovation.

Permiso Secures AI Identities for Users, Builders, and Agents

Permiso has updated its identity security platform to protect AI users, builders, and agents, offering a single solution for all digital identities. The platform uses runtime intelligence to track AI usage patterns, providing security teams with detailed insights. Permiso categorizes AI identities into users, builders, and agents, helping organizations discover vulnerabilities and defend against attacks. This expansion aims to give enterprises full coverage and eliminate blind spots in AI security.

Permiso Enhances Identity Platform for AI Security

Permiso is expanding its identity platform to cover AI users, builders, and agents, treating AI security as part of the overall identity layer. The update uses runtime intelligence and behavioral analytics to provide live visibility into AI usage. This helps security teams understand how AI is being used, where it's connected, and what data it accesses. Permiso aims to close gaps around shadow AI and personal account use, offering unified visibility and threat detection.

Permiso Adds AI Security to Identity Platform

Permiso has added comprehensive AI security features to its identity platform, covering AI users, builders, and agents. The new features provide full visibility into AI identities, protecting against compromise and misuse. Permiso categorizes AI identities and uses runtime intelligence to track usage patterns, helping security teams understand AI access and activity. This expansion aims to give enterprises a single platform for protecting all digital identities.

Entro Security Adds AI Agent Observability to Secure AI Agents

Entro Security has launched an AI Agents Discovery and Observability solution to secure AI agents within enterprise environments. The platform provides visibility, ownership, and control by mapping agents to non-human identities (NHIs) and human creators. It discovers credentials, enforces least-privilege, and monitors agent behavior in real time. This solution helps organizations adopt agentic AI securely by managing secrets, NHIs, and AI agents.

SailPoint Unveils AI Agent Identity Security Platform

SailPoint has introduced its new AI Agent Identity Security platform, designed to secure the growing number of non-human identities, including AI agents. The platform provides discovery, governance, and security for every agent, assigning clear ownership and enforcing permissions. It links agent activity to data at a granular level, offering visibility and control for autonomous AI. This innovation aims to bring trust, compliance, and security to the era of AI.

Trump's Team Uses AI for Epic, Viral Portraits

Donald Trump's team is using AI to create and post imaginative portraits of him, depicting him in various heroic or powerful scenarios. These AI-generated images are designed to be attention-grabbing and have gained significant traction online. While not intended to deceive, these images reflect a trend of using AI to craft a desired public image. Experts note the long history of politicians using imagery to shape narratives.

Trump Posts Debunked MedBed AI Video, Sparks Outrage

Donald Trump posted an AI-generated video promoting a debunked conspiracy theory about 'MedBed' hospitals, causing online outrage. The video, styled as a Fox News segment, promised universal access to futuristic hospitals. MedBeds are linked to the QAnon movement and are not real. Trump later deleted the post without explanation, and Fox News confirmed the segment never aired on their channels.

BullFrog AI Expands Sales to Boost Clinical Trial Solutions

BullFrog AI is expanding its internal sales team to speed up the commercialization of its AI-powered clinical trial optimization tools, including the new bfPREP module. This move complements their existing partnership with Sygnature Discovery for target selection solutions. The bfPREP module helps standardize complex clinical trial data quickly. BullFrog AI aims to accelerate revenue growth by offering end-to-end solutions for drug development.

BullFrog AI Grows Sales Team for bfPREP Platform

BullFrog AI is expanding its sales and marketing team to accelerate the adoption of its AI-driven clinical trial optimization solutions, like the bfPREP module. This expansion supports a dual strategy: a partnership with Sygnature Discovery for target selection and direct sales for trial optimization tools. The bfPREP module efficiently standardizes diverse clinical trial data. This move aims to increase revenue and provide comprehensive solutions across the drug development lifecycle.

AI Predicts Jake Browning Over 219.5 Passing Yards Monday Night

A self-learning AI model predicts Bengals quarterback Jake Browning will exceed 219.5 passing yards in Monday Night Football's game against the Broncos. The AI, developed by SportsLine's Data Science team, projects Browning to throw for 261.8 yards. Browning has shown potential in previous starts, and the model rates this pick as a strong 4.5-star play, suggesting he may need to pass frequently in the game.

Kolmar Korea Uses AI for Faster Beauty Packaging

Kolmar Korea has launched an AI-powered platform, Kolmarpkg.com, to speed up the sourcing of beauty product packaging. The platform allows beauty brands to upload product images and receive AI-recommended packaging solutions from a database of 4,000 certified options. This tool reduces the product launch timeline from six months to one month. It connects clients with local manufacturers worldwide, streamlining the entire packaging sourcing process.

Hance to Demo Tiny AI Audio Software at TechCrunch Disrupt

Hance will showcase its new AI audio processing software at TechCrunch Disrupt 2025. The software is notable for its extremely small size, requiring only a kilobyte of space. This technology allows for efficient AI audio processing, potentially enabling new applications and devices. Hance's demonstration at the event will highlight the capabilities of this compact AI solution.

Shortcut Launches Korey AI for Product Management

Shortcut has launched Korey, an AI project manager designed to orchestrate software development projects from planning to coordination. Korey turns ideas into structured plans, develops specifications, breaks work into tasks, and summarizes progress. It aims to automate busywork, allowing teams to focus on core tasks. Korey learns from team operations and integrates with tools like GitHub to provide real-time project insights.

New Bill Aims to Regulate AI Experimentation

Senator Ted Cruz introduced the SANDBOX Act on September 10, 2025. This bill focuses on strengthening artificial intelligence normalization and diffusion through oversight and experimentation. The legislation aims to create a framework for regulating AI development and deployment.

China's Zhipu AI Skeptical of Superintelligence by 2030

Zhang Peng, CEO of China's Zhipu AI, believes full artificial superintelligence (ASI) surpassing human capabilities across the board is unlikely by 2030. While ASI might exceed humans in some areas, it will likely fall short in many others. Zhipu AI, a leading Chinese AI startup, recently released its latest large language model, GLM-4.6, and is focusing on enterprise clients.

Base NFTs Lead Sales, Driven by AI Trading Game

Non-fungible tokens (NFTs) on the Base network saw the highest sales count in September, largely due to the AI trading game DX Terminal. DX Terminal recorded over 1.27 million transactions, a significant increase from the previous month. While Base NFTs led in sales volume, Ethereum and Polygon collections still dominated overall trading value. DX Terminal features NFTs as AI trader agents competing in a market simulation.

Sources

AI Security Enterprise Security AI Risk Management SaaS Security AI Identity AI Agents Non-Human Identities (NHIs) AI Governance Threat Detection AI Innovation AI Tool Use Data Exposure AI Observability AI Experimentation Regulation Artificial Superintelligence (ASI) Large Language Models (LLMs) Clinical Trial Optimization Drug Development AI Audio Software Product Management Software Development AI-Generated Images AI-Generated Video Beauty Packaging NFTs AI Trading Game