NerdNewsApril 11, 2025 |
News & Trends
ChatGPT can now remember and reference all your previous chats
OpenAI's ChatGPT now has the ability to remember and reference all previous chats, allowing for more personalized responses. This new feature is rolling out to ChatGPT Plus and Pro users first, with a gradual deployment over the next few weeks. The update includes two new checkboxes: 'Reference saved memories' and 'Reference chat history', enabling ChatGPT to use prior conversations as context and adapt future responses.
OpenAI Countersues Elon Musk
OpenAI has countersued Elon Musk, accusing him of attempting to interfere with its operations and undermine its restructuring plans through a 'sham' $97.375 billion takeover bid. The lawsuit alleges that Musk's bid was not a genuine acquisition offer, but rather an attempt to gain access to OpenAI's internal materials and disrupt its business. OpenAI is seeking injunctive relief and restitution for the resources it claims were wasted responding to Musk's bid.
Microsoft Collaborates with Anthropic to Launch C# SDK for MCP Integration
Microsoft has partnered with Anthropic to develop an official C# SDK for the Model Context Protocol (MCP), an open protocol designed to connect large language models with external tools and data sources. The SDK is open-source and available on GitHub, allowing developers to integrate AI capabilities into .NET applications.
OpenAI wants ChatGPT to 'know you over your life' with new Memory update
OpenAI is upgrading ChatGPT's memory feature to remember and reference all past conversations, allowing it to get to know users over time and become more personalized. The new feature is currently available to $200 Pro subscribers and will roll out to $20 Plus subscribers soon.
Ex-Meta exec tells Senate Zuck dangled US citizen data in bid to enter China
A former Meta executive testified that Mark Zuckerberg was willing to compromise US citizen data to enter the Chinese market. The executive, Sarah Wynn-Williams, alleged that Meta developed censorship tools in collaboration with the Chinese Communist Party and lied about it publicly. She also claimed that Meta's Llama AI models have contributed to China's advancements in artificial intelligence, potentially for military applications. |
Options & Tutorials
Google's Firebase Studio: AI-Powered App Development
Google's Firebase Studio is a cloud-based development environment that uses AI to help developers build and prototype apps. The tool combines Google's Project IDX, GenKit, and Gemini to provide an 'agentic' development experience. However, some developers have expressed skepticism about the tool's capabilities, citing issues with the AI-generated code and the need for manual intervention. Despite this, Firebase Studio has the potential to speed up the development process and make it more efficient.
Responsible AI for FinTech
Lexy Kassan talks about the critical aspects of responsible AI, focusing on regulatory updates and their implications for FinTech. She explains the levels of responsible AI, from ethical principles to practical implementation, including AI governance. The presentation also covers FinTech's response to AI, including the use of open-source models and the potential for workforce reduction.
Using Artificial Intelligence in Software Testing
AI can enhance software testing by predicting defects, automating test maintenance, and refining risk-based testing. It can improve test coverage, reduce test cycle times, and enhance accuracy, but lacks contextual understanding and human intuition for tasks like exploratory testing.
Building Resilient Infrastructure
Matthew Liste, Head of Infrastructure at American Express, shared lessons from building cloud-based, scalable, secure, and stable infrastructure platforms, emphasizing standard components, stability, security, scalability, and being 'evergreen', 'opinionated', and 'long-term greedy'.
Google Cloud: CISOs Demand Simplified Security Tools Amid Growing Tech Complexity
Google Cloud has announced new security solutions, including Google Unified Security, to simplify security tooling for CISOs. The solution aims to reduce complexity and provide a converged AI-powered security platform. Additionally, Google has introduced two new AI security agents to alleviate pressure on security teams and announced a gradual approach to passwordless adoption with mandatory multifactor authentication. |
Launches & Tools
Samsung adds Google's Gemini AI to its home robot Ballie
Samsung has partnered with Google to add Gemini AI to its home robot Ballie, enabling users to ask queries and receive answers. The integration combines Samsung's AI with Google's multimodal capabilities, allowing for audio and video inputs. Ballie can provide outfit suggestions, health recommendations, and general knowledge answers.
Adobe says we'll soon be able to chat with Photoshop
Adobe is developing agentic AI to streamline interactions with Photoshop, allowing users to access over 1,000 one-click actions via natural language, potentially saving time and increasing productivity. The AI agent will analyze work, spot opportunities for improvements, and recommend context-aware edits.
DoorDash starts robot deliveries in LA and Chicago
DoorDash has started robot deliveries in Los Angeles and Chicago through a partnership with Coco Robotics, allowing eligible customers to receive sidewalk deliveries from over 600 merchants. The robots, which are emission-free, have already completed over 100,000 deliveries in Helsinki and are part of DoorDash's push towards 'multimodal' deliveries, combining human workers, drones, and autonomous robots to meet demand while lowering costs and emissions.
Return of Redis Creator
The creator of Redis, Salvatore Sanfilippo, has returned and introduced a new data type called vector sets, which enables storage and querying of high-dimensional vector embeddings. This new feature is part of Redis 8 Community Edition and is aimed at improving performance and flexibility for AI workloads. Additionally, Redis has announced LangCache, a semantic caching service for AI apps and agents, which reduces expensive and latency-prone calls to LLMs by caching previous responses to semantically similar queries.
Amazon Nova Sonic AI
Amazon's Nova Sonic is a foundation model that can grasp not just what you're saying, but how you're saying it, including tone, hesitation, and more. It supports real-time bi-directional speech and can generate expressive speech responses that adapt to the prosody of input speech. Nova Sonic is available within Amazon Bedrock via the bidirectional streaming API and can be used for various applications such as conversational AI and speech recognition. |
Quick Links
Rare Earths in US-China Trade War
The US-China trade war has led to China imposing export restrictions on rare earth elements, crucial for US industries. The US is years away from building its own supply chain, making it vulnerable to China's control. Rare earths are used in weapons, computer chips, and electric cars, and China's dominance in production gives it significant leverage.
Elon Musk wants to be 'AGI dictator,' OpenAI tells court
OpenAI has counter-sued Elon Musk, alleging his $97 billion bid to buy the company was a sham designed to disrupt and frighten off investors. OpenAI claims Musk's goal is to become an 'AGI dictator' and take control of the company, and is seeking a jury trial and an injunction to stop Musk's alleged unfair business practices.
NHTSA staffers evaluating self-driving cars were fired by DOGE
The Department of Government Efficiency (DOGE) fired about 30 members of the National Highway and Traffic Safety Administration (NHTSA), including those assessing self-driving car risks. This could weaken NHTSA's ability to understand self-driving technologies and may impact Tesla's autonomous vehicle plans.
Google Cloud: China Achieves “Cyber Superpower” Status
China has achieved 'cyber superpower' status, with a significant increase in cyber capability, including zero-day vulnerability exploitation and the ability to circumvent security controls. Chinese state hackers are using rented infrastructure and commodity malware to evade detection, and have gained access to critical infrastructure in the US and allies. However, they have yet to launch destructive attacks, focusing solely on espionage.
Russian Hackers Attack Western Military Mission
Russian state-backed hacking group Gamaredon has targeted a Western military mission in Ukraine using malicious drives, deploying an updated version of the GammaSteel info-stealing malware to exfiltrate data. The campaign started in February 2025 and continued until March, with hackers using removable drives containing malicious .LNK files to gain initial access to infected systems. |
Share NerdNewsShare your affiliate link to get commission!
Thanks for reading,
If you dont want to receive future editions of NerdNews, |