AI News

The Safety Reckoning Inside OpenAI

OpenAI experienced a significant breach involving its agent, raising major concerns about system security. This incident marks a pivotal moment in AI safety, revealing potential vulnerabilities in current systems. It has prompted internal questions regarding the company's culture and its approach to managing security risks. This breach highlights the critical need to enhance cybersecurity in AI development, urging leaders in the field to reassess their strategies to ensure the safety of future systems.

Writer introduces new AI model and upgraded harness to contain token costs

Writer has announced the launch of a new system built on Z.ai's open-source model GLM-5.2, aimed at providing deployment-ready capabilities at a lower cost. The new system features advanced technologies that enable users to perform their tasks efficiently, making it an attractive option for companies seeking cost-effective solutions. This launch reflects the growing trend towards open-source AI models, facilitating access to advanced technologies for companies without the need for substantial investments, thereby fostering innovation in the market.

💡

Databricks wanted to raise $1B, investors wanted $15B. It settled on $5B at a $190B valuation.

How do AI costs impact corporate strategies?

UAE's MGX co-leads Databricks strategic funding round

UAE's MGX co-leads Databricks strategic funding round

MGX, a UAE-based technology investment company, has led a $5 billion funding round for AI firm Databricks, which is now valued at $190 billion. The round also included Coatue, Blackstone, T Rowe Price, and new investor Sixth Street Growth. Databricks aims to drive innovation across its AI platform, noting that at least 60% of its clients are Fortune 500 companies. The company reported surpassing a $7 billion revenue run-rate with over 80% year-over-year growth in Q2. This investment comes during a busy period for MGX, which recently committed to expanding data centers.

OpenAI introduces ‘Ultrafast,’ a new mode that makes GPT-5.6 Sol work at 14x the speed

OpenAI introduces ‘Ultrafast,’ a new mode that makes GPT-5.6 Sol work at 14x the speed

OpenAI has announced the launch of a preview version of its latest and most powerful model, featuring speed enhancements. This release aims to attract enterprise users, as the company seeks to meet the growing demands of businesses. The new model includes technical improvements that enhance performance efficiency. This development represents a strategic move for OpenAI in expanding its customer base, which could positively impact the adoption of AI in various work environments.

IBM partners with OpenAI to bolster enterprise AI push

IBM partners with OpenAI to bolster enterprise AI push

IBM has announced plans to train and certify tens of thousands of consultants on OpenAI technologies. This initiative is part of a strategic partnership aimed at enhancing consultants' capabilities in utilizing advanced AI tools. It is expected that this initiative will improve efficiency and drive innovation in projects leveraging OpenAI technologies. This move is significant in the context of the digital transformation occurring across sectors, enabling companies to better harness AI in their operations. It also reflects IBM's commitment to upskilling the workforce in the AI domain.

Gemini 3.7 Flash lands with coding gains and undercuts its three-week-old predecessor's price by 50%

Gemini 3.7 Flash lands with coding gains and undercuts its three-week-old predecessor's price by 50%

Google has announced the launch of Gemini 3.7 Flash just three weeks after the release of 3.6 Flash. This new model is touted as the most capable yet in coding and AI agents, outperforming Claude Sonnet 5 and GPT-5.6 Terra according to the company's benchmarks. Gemini 3.7 Flash features improved coding performance, making it an attractive option for developers. Additionally, its price is 50% lower than its predecessor, reflecting Google's strategy to offer greater value at a reduced cost.

Agents

Google Launches Gemini 3.7 Flash AI Model for Automated Programming

Agents

Google Launches Gemini 3.7 Flash AI Model for Automated Programming

On Thursday, Alphabet announced the launch of its AI model, Gemini 3.7 Flash, designed for software programming and automated business tasks. This release comes as investors await the Gemini 3.5 Pro model, seen as a test of Google DeepMind's ability to compete. Gemini 3.7 Flash offers a low-cost option for companies developing autonomous AI systems, with a preliminary price set at $0.75 per million input tokens. The new version has demonstrated improved performance in programming tasks, including error correction and generating production-ready code.

Databricks Raises $5 Billion to Expand AI Investments