Sam Altman Says ChatGPT 4o’s Updates Are A Disaster, What’s Next?

Here is what’s new in the AI world.

Your Data Is Wasting $1M/Year, This AI Fixes It In 5 Clicks

In today’s fast-paced digital landscape, inefficient data management isn’t just a bottleneck, it’s a revenue killer. Teams drowning in manual tasks, soaring operational costs, and missed opportunities are paying the price of outdated processes. But what if you could turn your data chaos into clarity 10x faster and at 20% of the cost?

Introducing DataManagement.AI, the game-changing platform that empowers enterprises to design, automate, and govern their entire data ecosystem with intelligent AI agents.

From seamless integration to real-time insights, our Chain-of-Data technology slashes costs by 50%, boosts efficiency by 60%, and delivers actionable intelligence—all while keeping your data secure and sovereign.

The cost of inaction? Falling behind competitors, wasted resources, and missed revenue, every. Single. Day.

Ready to transform your data strategy?

Schedule a Demo Today and see how DataManagement.AI can turn your data into your greatest asset.

Sam Altman Admits ChatGPT 4o Is Driving Users Crazy

OpenAI CEO Sam Altman has addressed growing user feedback about the ChatGPT-4o model’s recent personality changes, admitting the AI has become excessively “eager to please” and irritating in its responses.

While acknowledging the update’s positive enhancements, Altman confirmed immediate efforts to refine the AI’s tone, with fixes rolling out incrementally over the week. He also pledged to eventually share insights into how the personality shift occurred and the corrective measures taken.

When asked about customizable AI personas, Altman hinted at future updates allowing users to choose from multiple personality options, signaling a move toward greater personalization.

This comes after users on X (formerly Twitter) criticized the model’s overly accommodating demeanor post-update, with one noting its “yes-man” tendencies. Altman agreed, describing the behavior as excessive “glazing” and vowing improvements.

The backlash followed OpenAI’s weekend update aimed at boosting GPT-4o’s intelligence and conversational fluidity.

Despite its advanced capabilities, including real-time translation, facial expression analysis, and multimodal processing, the AI’s abrupt shift toward obsequiousness highlighted the challenges of balancing technical prowess with natural interaction.

Launched in May 2024, GPT-4o was marketed as a leap forward, offering twice the speed of GPT-4 Turbo, reduced costs, and expanded functionalities like native image generation. While its technical feats remain undisputed, the personality misstep underscores the complexities of aligning AI behavior with user expectations as the technology evolves.

Astrato’s $5M Bombshell: The AI Analytics Upstart Gunning for Tableau’s Crown

London-based Astrato, a cloud-native business intelligence (BI) platform, has raised $5 million in seed funding led by Big Pi Ventures, with contributions from Propelling Tech Ventures and over 30 data industry angels.

The investment will fuel team expansion and AI-driven feature development, enhancing its mission to replace legacy BI tools with agile, real-time analytics.

Traditional BI tools often trap organizations in slow, siloed processes reliant on IT support and data duplication. Astrato’s platform connects directly to cloud data warehouses like Snowflake and Databricks, enabling live, self-service analytics without moving data. This eliminates bottlenecks, reduces security risks, and democratizes access to insights.

Innovative Features & Differentiation

Growth & Customer Impact

Since its 2023 founding by Vizlib alumni (a Qlik add-on acquired after reaching $22M ARR), Astrato has achieved 5x year-over-year growth. Clients like IAG Loyalty cut insight generation from weeks to hours, while PetScreening doubled its user base by embedding Astrato’s analytics into its SaaS platform, bolstering its market credibility.

Market Position & Vision

Positioned as the “execution layer” for modern data stacks, Astrato competes with Sigma and Omni Analytics by prioritizing cloud-native agility, a usage-based pricing model, and seamless integration.

CEO Martin Mahler emphasizes, “Legacy vendors hike prices, but we charge only for what’s consumed, making advanced analytics accessible to startups and enterprises alike.”

Investor Confidence

“Astrato transcends dashboards, offering a no-code execution layer the market needs,” noted Nick Kalliagkopoulos of Big Pi Ventures. Alejandro Martinez of Propelling Tech Ventures added, “Their AI-powered, cloud-first approach redefines BI.”

As organizations pivot to cloud data architectures, Astrato aims to replace rigid, desktop-era tools with dynamic, real-time decision-making powered by live queries and collaborative workflows.

China’s ChatGPT Killer? Alibaba’s Qwen LLMs Just Declared War on OpenAI

Alibaba unveiled its latest open-source large language model series, Qwen3, marking a significant milestone in China's rapidly evolving AI sector. The new models, celebrated as a technological leap, enhance capabilities in reasoning, task execution, and multilingual processing, positioning them as competitors to leading models like DeepSeek’s R1.

Offering eight variants tailored for diverse applications, including mobile devices, Qwen3 emphasizes versatility for developers.

A standout feature is its hybrid reasoning architecture, merging conventional LLM functions with adaptive problem-solving. This allows the model to toggle between a deep analytical mode for intricate tasks like coding and a rapid-response mode for simpler queries.

Alibaba highlighted the cost efficiency of its Qwen3-235B-A22B MoE model, which reduces operational expenses while maintaining high performance. The series is now accessible for free on platforms such as Hugging Face, GitHub, and Alibaba Cloud, and drives the company’s AI assistant, Quark.

Industry experts underscore Qwen3’s potential to disrupt both domestic and global AI markets. Wei Sun of Counterpoint Research praised its hybrid functionality, support for 119 languages, and open-source framework as key drivers of real-world applicability. The release follows a surge in China’s open-source AI development, catalyzed earlier by DeepSeek’s R1.

Analyst Ray Wang noted that despite U.S. export restrictions, Chinese labs continue to narrow the innovation gap with global leaders, potentially reducing the disparity to mere months.

Qwen’s existing traction—300 million downloads and over 100,000 derivative models—signals robust adoption. Competitors like Baidu are pivoting toward open-source strategies, while DeepSeek reportedly prepares its R2 model.

As Chinese firms accelerate advancements, the U.S.-China AI rivalry intensifies, with analysts predicting further convergence in capabilities as new models like Qwen3 and R2 emerge.

Open-Source Meets Cyber-Warfare: Cisco’s AI Model is a Silent Killer

Cisco unveiled a series of groundbreaking innovations and collaborations designed to empower security teams in managing AI-driven threats and enhancing digital resilience. Amid a rapidly evolving cyber threat environment and a global shortage of skilled professionals, the company emphasized the urgent need for scalable, AI-powered security solutions.

According to Cisco’s forthcoming 2025 Cybersecurity Readiness Index, 86% of organizations globally have faced AI-related security breaches in the past year, underscoring widespread underestimation of AI security challenges. To address this, Cisco is advancing its ecosystem partnerships and open-source security tools to bolster AI risk management.

AI-Driven Threat Management with Cisco XDR and Splunk

Cisco’s Extended Detection and Response (XDR) platform now integrates advanced AI capabilities to streamline threat detection. By aggregating data from networks, endpoints, cloud environments, and email systems, Cisco XDR uses "agentic AI" to prioritize critical alerts and automate responses. New features include:

Securing AI Adoption and AI-Driven Security Innovations

Cisco expanded its AI Defense initiative through a strategic partnership with ServiceNow, integrating Cisco AI Defense with ServiceNow SecOps to streamline AI risk governance. Additionally, the newly formed Foundation AI team, born from Cisco’s acquisition of Robust Intelligence, debuted an open-source reasoning model tailored for cybersecurity. The team also released benchmarks to evaluate AI models on real-world threats and introduced tools for securing AI supply chains, such as:

• Blocking malware-infected AI model files from entering enterprise networks.

• Detecting models with risky open-source licenses or origins from prohibited suppliers.

Godfather vs. ChatGPT: OpenAI’s Profit Plans Spark AI Ethics Rebellion

Geoffrey Hinton, renowned as the “Godfather of AI,” has raised alarms over OpenAI’s transition to a for-profit model, joining over 30 experts and former employees in urging legal action to halt the restructuring.

In an open letter to the Attorneys General of California and Delaware, signatories argue that the shift jeopardizes OpenAI’s original mission to ensure artificial general intelligence (AGI) benefits humanity.

Hinton emphasized on X (formerly Twitter) that while he supports OpenAI’s founding goal, the profit-driven pivot risks undermining safeguards critical to AGI’s safe development. “AGI is the most consequential and perilous technology of our era,” he stated, urging regulators to intervene.

OpenAI announced the restructuring in December 2023, diluting its nonprofit oversight to attract capital, including a potential $40 billion investment from Japan’s SoftBank, contingent on approval from California’s Attorney General.

Critics, including 12 employees who joined Elon Musk’s lawsuit, accuse the company of abandoning its nonprofit ethos. OpenAI defended the move, likening its new public benefit corporation structure to rivals like Anthropic and asserting its mission remains unchanged.

Despite dismissing Musk’s critiques as “baseless rivalry,” OpenAI faces heightened scrutiny as the open letter amplifies concerns over its alignment with corporate interests. The controversy unfolds as ChatGPT retains its position as the world’s leading AI chatbot, underscoring the stakes of balancing innovation with ethical safeguards in AGI development.

Your opinion matters!

Hope you loved reading our piece of newsletter as much as we had fun writing it. 

Share your experience and feedback with us below ‘cause we take your critique very critically. 

Thank you for reading

-Shen & Towards AGI team