google, openai and salesforce Updates

The artificial intelligence sector is experiencing intense competition and rapid development, with major players like Google and OpenAI at the forefront. OpenAI CEO Sam Altman reportedly issued a "code red" as Google's new AI model, Gemini 3, demonstrates superior performance, prompting Salesforce CEO Marc Benioff to switch to it. Despite having 800 million weekly ChatGPT users, OpenAI, valued at $500 billion with funding from Microsoft, is now redirecting resources to enhance its chatbot and plans a massive $1.4 trillion investment in data centers over eight years. Meanwhile, Apple is also bolstering its AI talent, having hired Amar Subramanya, a former Google Gemini head, as its new AI vice president. Beyond the competitive race for foundational models, AI applications are expanding across various industries. New AI browsers, such as Perplexity's Comet and ChatGPT Atlas, are emerging but still require users to craft very specific prompts to yield good results, making them less intuitive than traditional search engines like Google. In cybersecurity, Hack The Box (HTB) launched its HTB AI Range on December 3, 2025, a platform designed to test AI agents and human defenders against cyberattacks in realistic simulated environments. HTB is also partnering with Google to offer an AI Red Teamer Certification in Q1 2026, aiming to close a significant skills gap in AI security. Healthcare is another area seeing substantial AI integration. Healthify upgraded its Ria AI assistant, which uses OpenAI's technology, to include real-time conversation features and support over 50 languages, including 14 Indian languages. Ria can now log meals via camera, even through Ray-Ban Meta smart glasses, and Healthify plans a U.S. AI plan for $20 per month. Google is also investing in European healthcare AI, with Google.org providing $5 million to Bayes Impact for "Impulse Healthcare," an initiative empowering medical professionals to develop their own AI solutions, which has already shown success in reducing emergency room wait times by over an hour. Enterprises are actively working to integrate AI and train their workforces. Davenport University's Institute for Professional Excellence (IPEx) partnered with Business Connect to deliver custom AI training programs, featuring 15 hours of instructor-led sessions using real company data. General Dynamics Information Technology (GDIT) President Amy Gilliland emphasizes human mastery of AI, noting a tenfold increase in AI training courses for GDIT staff in 2025. GDIT also opened its Mission Emerge Center near Washington, DC, a 5,000-square-foot facility for government customers to securely test AI products. Furthermore, AWS is simplifying AI model customization with new features in Amazon Bedrock and Amazon SageMaker AI, making advanced techniques like Reinforcement Fine Tuning (RFT) more accessible and improving model accuracy by an average of 66 percent. As AI adoption grows, so does the focus on ethical deployment and regulation. Senator Ed Markey and Congresswoman Yvette Clarke reintroduced the AI Civil Rights Act of 2025. This bill aims to apply existing civil rights laws to AI systems, making it illegal for developers or users to deploy algorithms that cause "disparate impact" and unfairly harm specific groups. The act mandates regular evaluations and assessments of AI systems to ensure transparency and accountability, placing responsibility on developers and deployers to prevent civil rights violations. Google Photos also introduced new features for its 2025 Recap, offering more customization, shareability, and integration with Gemini features for U.S. users to highlight top hobbies.

Key Takeaways

• Google's Gemini 3 model is intensely competing with OpenAI's ChatGPT, leading Salesforce CEO Marc Benioff to switch to Gemini 3.
• OpenAI, valued at $500 billion with Microsoft funding, plans to invest $1.4 trillion in data centers over eight years to enhance its chatbot.
• Apple hired Amar Subramanya, a former Google Gemini head, as its new AI vice president, signaling a push in its AI strategy.
• New AI browsers like Perplexity's Comet and ChatGPT Atlas still require users to create very specific prompts, making them less intuitive than traditional search engines.
• Hack The Box (HTB) launched its HTB AI Range for cyberattack training and partnered with Google to offer an AI Red Teamer Certification in Q1 2026.
• Healthify upgraded its Ria AI assistant (using OpenAI technology) with real-time conversation, 50+ language support, and meal logging via Ray-Ban Meta smart glasses.
• Google.org is providing $5 million to Bayes Impact for "Impulse Healthcare" in the EU, empowering medical professionals to create AI solutions and already reducing ER wait times by over an hour.
• AWS introduced new features in Amazon Bedrock and Amazon SageMaker AI, simplifying advanced model customization and improving accuracy by an average of 66 percent.
• General Dynamics Information Technology (GDIT) increased AI training courses tenfold in 2025 and opened a Mission Emerge Center for government customers to test AI products.
• The AI Civil Rights Act of 2025 was reintroduced to protect against AI discrimination, making "disparate impact" illegal and requiring regular AI system evaluations.

AI Browsers Still Need Smart Users Today

Victoria Song reported on December 3, 2025, that new AI browsers like Perplexity's Comet and ChatGPT Atlas are not yet better than humans at web surfing. She tested five browsers, including Chrome and Edge, to see if they improved the internet experience. The main issue is that users must still create very specific prompts for these AI tools. This is different from traditional search engines like Google, which can understand less precise queries.

New AI Browsers Struggle with User Prompts

Victoria Song reported on December 3, 2025, that five different AI browsers, including Perplexity's Comet and ChatGPT Atlas, share a common issue. These new browsers aim to change how people use the internet by embedding chatbots. However, users must still work hard to create the right prompts to get good results. This makes them less intuitive than traditional search engines like Google.

Hack The Box Launches AI Cyber Training Platform

On December 3, 2025, Hack The Box HTB introduced its new HTB AI Range. This platform helps companies test AI agents and human defenders against cyberattacks in a realistic setting. It simulates complex business environments and follows cybersecurity guidelines like MITRE ATT&CK. While AI agents solved many basic challenges, human teams performed better in more complex tasks. HTB also plans to offer an AI Red Teamer Certification early next year to improve AI defense skills.

HTB Unveils AI Cyber Range and New Certification

On December 3, 2025, Hack The Box HTB launched the HTB AI Range, a new platform for businesses to test AI security agents and human defenders. This "live-fire" training ground simulates complex cyberattacks and aligns with standards like MITRE ATT&CK. HTB also announced an AI Red Teamer Certification, set for Q1 2026, created with Google to address a major skills gap in AI security. CEO Haris Pylarinos stated HTB is defining how cyber defense evolves with AI.

Davenport University Offers Custom AI Training

On December 3, 2025, Davenport University's Institute for Professional Excellence IPEx announced a partnership with Business Connect. IPEx creates custom AI training programs for businesses to help employees use AI effectively. Business Connect chose IPEx after a wide search because IPEx designed a program specifically for their staff and company culture. The training included 15 hours of instructor-led sessions and hands-on exercises using real company data. Dan Rundhaug, IPEx executive director, emphasized their unique approach to make training relevant and practical.

OpenAI Faces Challenge from Google Gemini 3

OpenAI CEO Sam Altman issued a "code red" as ChatGPT faces strong competition from Google's new AI model, Gemini 3. Gemini 3 has shown better performance than rivals, leading some, like Salesforce CEO Marc Benioff, to switch to it. OpenAI, which has 800 million weekly ChatGPT users, is now focusing more resources on improving its chatbot and delaying advertising plans. The company, valued at $500 billion with funding from Microsoft, plans to spend $1.4 trillion on data centers over eight years. Meanwhile, Apple also hired Amar Subramanya, a former Google Gemini head, as its new AI vice president.

Healthify Upgrades Ria AI Assistant for Real-Time Talk

On December 3, 2025, Healthify upgraded its AI assistant, Ria, to include real-time conversation features. Ria uses OpenAI's technology and now supports over 50 languages, including 14 Indian languages. Users can ask Ria for health summaries, get insights from various trackers, and log meals by pointing their camera, even through Ray-Ban Meta smart glasses. Healthify plans to integrate Ria into user onboarding and create a memory layer for personalized advice. The company, with over 45 million users, is launching a new AI plan in the U.S. for $20 per month.

Google Funds New AI Health Initiative in Europe

At the European Health Summit in Brussels, Google's Greg Corrado announced new funding for health AI. A Google-commissioned report revealed that AI is boosting scientific productivity in European healthcare and reducing costs. For example, AI has already cut emergency room wait times by over an hour. Google.org is providing $5 million to Bayes Impact to start "Impulse Healthcare," an EU project. This initiative will empower nurses, doctors, and administrators to create and test their own AI solutions, aiming to improve patient care across Europe.

Google Photos 2025 Recap Offers New Features

Tommy Meaney, Senior Product Manager at Google Photos, announced new features for the 2025 Google Photos Recap. This year's Recap is more customizable and shareable, highlighting memorable photos with fun graphics and cinematic effects. Users can now see a selfie count and hide specific people or photos from their Recap. Those in the U.S. with Gemini features can also see top hobbies. The Recap can be easily exported to CapCut for editing with special templates and shared directly to social media or WhatsApp Status.

AWS Simplifies AI Model Customization for Developers

AWS introduced new features in Amazon Bedrock and Amazon SageMaker AI to simplify advanced model customization. These updates help customers build faster and more efficient AI agents while reducing costs. Reinforcement Fine Tuning RFT in Amazon Bedrock makes it easier for any developer to tailor models, improving accuracy by an average of 66 percent. Amazon SageMaker AI now reduces complex customization workflows from months to just days. This makes advanced techniques, previously requiring deep machine learning expertise, accessible to more organizations.

GDIT President Gilliland Focuses on Human AI Mastery

Amy Gilliland, President of General Dynamics Information Technology GDIT, believes human employees will master artificial intelligence. She reported that GDIT's investment focuses on adapting Large Language Models and training staff, with AI training courses increasing tenfold in 2025. Employees are eager to use AI for routine tasks, allowing them to do more creative work. GDIT also launched an AI tool for internal job matching and unveiled its new Mission Emerge Center near Washington, DC. This 5,000-square-foot facility allows government customers to test AI products in a secure, unclassified environment, aiming to speed up military technology upgrades.

New Bill Protects Civil Rights from AI Harm

Senator Ed Markey and Congresswoman Yvette Clarke reintroduced the AI Civil Rights Act of 2025 to protect people from AI discrimination. The bill aims to apply existing civil rights laws to artificial intelligence systems, which are used in important areas like housing, employment, and healthcare. It makes it illegal for AI developers or users to deploy algorithms that cause "disparate impact," meaning they unfairly harm certain groups. The act also requires regular evaluations and assessments of AI systems to ensure transparency and accountability. Developers and deployers will be responsible for preventing civil rights violations by their AI models.

Sources

• One day, AI might be better than you at surfing the web. That day isn’t today.
• Five different AI browsers have the same problem
• HTB AI Range offers experiments in cyber-resilience training
• Hack The Box Launches AI Cyber Range for Enterprise Readiness
• Davenport University’s IPEx provides custom AI training partnership with Business Connect
• Sam Altman issues ‘code red’ at OpenAI as ChatGPT contends with rivals
• Healthify upgrades its AI assistant Ria with real-time conversation capabilities
• We’re announcing new health AI funding, while a new report signals a turning point for health in Europe.
• Look back on your 2025 with Google Photos Recap
• AWS simplifies model customization to help customers build faster, more efficient AI agents
• GDIT chief Gilliland bets her humans will master artificial intelligence
• AI is Infringing on Your Civil Rights. Here’s How We Can Stop That