MIT's new image gen | Gemini with Deep Think achieve gold at IMO | YouTube removed nearly 11,000 channels

🌟 Welcome to the latest edition of Top249 🤩 

We're here to keep you updated on AI, tech, and business news and articles, so you can save time and focus on learning and growth.

Let's get started! 😎 

Important: If you're reading this in the Promotions tab, please drag it to your Inbox and click ‘Yes’ when your email provider asks to do it for future emails.

MIT: A new way to edit or generate images

• A new method allows users to edit or generate realistic images by simply describing the desired changes in natural language.

• Unlike previous models, it can understand spatial relationships and fine-grained object details, leading to highly accurate edits.

• The technique is more efficient and interpretable, making it potentially useful for creative industries, design, and medical imaging.

Advanced version of Gemini with Deep Think officially achieves gold-medal standard at the International Mathematical Olympiad

• An advanced version of Gemini with the “Deep Think” reasoning system has officially achieved gold-level performance on past International Mathematical Olympiad (IMO) problems.

• It solved over 60% of challenging math problems—comparable to top-tier high school mathematicians—without prior access to solutions.

• This milestone highlights significant progress in AI’s ability to perform complex symbolic reasoning and step-by-step logical thinking.

YouTube wipes out thousands of propaganda channels linked to China, Russia, others

• YouTube removed nearly 11,000 channels in the second quarter of 2025 that were linked to coordinated state-backed influence operations.

• Over 7,700 of those channels were connected to China and promoted pro-government narratives, while more than 2,000 were linked to Russian campaigns.

• The crackdown is part of Google’s broader effort to disrupt global disinformation networks originating from countries like China, Russia, Iran, Turkey, and others.

AI as the greatest source of empowerment for all

• OpenAI emphasized its long-term vision of AI as a force that expands access to education, productivity, and human potential globally.

• The company pledged to continue building safe, broadly beneficial systems while addressing risks like bias, misuse, and job displacement.

• OpenAI also reinforced its commitment to sharing benefits of AGI with society, including support for global equity and democratic values.

Crowdstrike’s massive cyber outage 1-year later: lessons enterprises can learn to improve security

• A faulty update from CrowdStrike caused a 78-minute outage that disrupted Windows systems globally across hospitals, airports, and enterprises.

• The incident has prompted companies to re-evaluate dependency on single vendors and improve rollback strategies for critical software updates.

• Cybersecurity leaders are now prioritizing resilience, transparency in vendor relationships, and enhanced testing pipelines to avoid similar risks.

Till next time.