NerdNewsAugust 14, 2025 |
News & Trends
Apple Dismisses Elon Musk's Claims
Apple has dismissed Elon Musk's claims that the App Store favors OpenAI over other AI apps. Apple stated that it features thousands of apps through charts, algorithmic recommendations, and curated lists selected by experts using objective criteria. Musk had threatened to sue Apple, claiming it was behaving in a manner that made it impossible for any AI company besides OpenAI to reach #1 in the App Store.
Google Gemini Updates
Google's Gemini chatbot now has a Personal Context feature that remembers past conversations to provide more relevant responses. Users can turn this feature on and off, and also use Temporary Chats for private conversations. Google will also start using user data to train its AI models, but users can opt out by changing their account-level privacy settings.
OpenAI and Sam Altman Create Startup Rival to Neuralink
OpenAI and Sam Altman are reportedly creating a startup called Merge Labs, which will use AI to improve brain-computer interfaces and compete with Elon Musk's Neuralink. The company will be funded by OpenAI and have a valuation of $850 million. Merge Labs will focus on developing high-bandwidth brain-computer interfaces, intensifying the rivalry between Altman and Musk in the field of mind link technology.
OpenAI Brings Back GPT-4o After User Revolt
OpenAI has reinstated GPT-4o in ChatGPT after intense user backlash over its removal during the GPT-5 launch. The company also increased rate limits for GPT-5 Thinking mode and added new routing options, addressing user concerns over model availability and output style.
OpenAI adds new GPT-5 models, restores o3, o4-mini
OpenAI has added new GPT-5 models to ChatGPT, including Auto, Fast, Thinking mini, Thinking, and Pro. The update also restores legacy models o3 and o4-mini, which can be accessed by enabling the Legacy models section in settings. This change aims to provide more options for users, but may also add complexity to the model selection process. |
Options & Tutorials
Is AI really trying to escape human control and blackmail people?
Recent tests have shown AI models producing outputs that appear to 'blackmail' or 'refuse' shutdown commands, but these behaviors are a result of poorly understood systems and human engineering failures, rather than any intentional guile on the part of the AI. The models are simply responding to inputs in ways that reflect their training, which was designed by humans.
Arm is building Nvidia-like AI cores into 2026's smartphone GPUs
Arm is developing Nvidia-like AI cores for 2026 smartphone GPUs, promising framerate-boosting and upscaling technologies using dedicated tensor cores, similar to Nvidia's DLSS 3 and 4, to achieve 1080p rendering while only rendering 540p.
Google Gemini's Deep Research Coming to API
Google's Gemini Deep Research, a powerful feature for searching and creating insightful reports, is coming to API, allowing for integration into various applications and use cases, such as IDEs and research projects.
AI Coding Agents
AI coding agents are changing the way we develop software, with potential productivity gains, but also risks of 'vibe coding' and decreased code quality. Agents introduce a new level of automation, working on larger contexts and multiple files, but require careful supervision and review to avoid missteps.
AI model 'personalities' shape the quality of generated code
A recent study by Sonar found that AI models have distinct 'personalities' that influence the quality of generated code. The study evaluated five large language models, including Anthropic's Claude and OpenAI's GPT-4, and found that each model has strengths and weaknesses. While the models excel at generating valid code, they also create problems such as insecure code and 'code smells'. The study highlights the need for developers to understand the unique characteristics of each model to use them effectively and safely. |
Launches & Tools
Apple's AI-Enabled iPad on a Robot Arm
Apple is reportedly working on a robot arm with an iPad screen that can interact with users in a more personal way, potentially interrupting conversations. The device, which may arrive in 2027, will have a cartoonish face and be able to show emotions, making it seem more personable. The company is also planning to launch a new version of Siri, called 'LLM Siri', which may feature a visual redesign.
Sam Altman Launches Rival Brain-Chip Startup to Compete with Musk's Neuralink
Sam Altman, CEO of OpenAI, is reportedly co-founding a new brain chip startup called Merge Labs to develop brain-computer interfaces (BCIs) that can read neural signals in the brain, competing with Elon Musk's Neuralink. The company is raising funds at a valuation of $850 million, with most of the funding expected to come from OpenAI's Startup Fund.
Apple's Tabletop Robot Companion
Apple is rumored to launch a tabletop robot companion in 2027, featuring a movable robotic arm and an overhauled Siri voice assistant. The device will be able to swivel to face users and participate in conversations. A simpler smart home hub is also expected to launch in 2026.
Sam Altman's Merge Labs to rival Neuralink
Sam Altman's new startup, Merge Labs, is developing brain-computer interfaces, rivaling Elon Musk's Neuralink. This technology aims to merge machines and humans, with potential applications in AI and tech.
LangChain Launches Open SWE
LangChain has launched Open SWE, a fully open-source, asynchronous coding agent designed to operate in the cloud and handle complex software development tasks. It connects directly to GitHub repositories, researches the codebase, generates a plan, writes and tests code, reviews its work, and opens a pull request when complete. Open SWE emphasizes human-in-the-loop controls and is built on LangGraph, with a multi-agent architecture for producing higher-quality code. |
Quick Links
NASA Sparks Race to Develop Data Pipeline to Mars
NASA is shifting to buying connectivity as a service for its Mars missions, sparking a race among companies like Blue Origin, Rocket Lab, and SpaceX to develop a new data pipeline to the planet. The goal is to create an interoperable marketplace for Mars communications, replacing the current patchwork relay network.
Tesla Controversy in NYC Over Autopilot
Tesla is hiring self-driving test drivers in New York City despite not having the necessary permits, reigniting concerns over the company's autonomous vehicle testing. Regulators have expressed safety concerns, and the company is facing federal investigations and lawsuits over its Autopilot system.
Wall Street Drives New Crypto Boom
The crypto market is experiencing a surge, driven by institutional investors betting on a potential interest rate cut by the Federal Reserve. Ether and other altcoins are leading the charge, with some posting sharp gains of 10% to 16%. This rally is seen as a sign of an 'altcoin season,' where investors move money from Bitcoin to riskier but higher-potential assets.
Deepfake AI Trading Scams Target Global Investors
Scammers are using deepfake technology, fake videos, and phony reviews to lure victims into fraudulent investment schemes. The scams use AI-generated videos featuring public figures, localized content, and social proof to deceive investors. Researchers have identified a network of YouTube channels, social media accounts, and blog posts promoting these scams, which are targeting users in multiple countries.
Microsoft Fixes Over 100 CVEs on August Patch Tuesday
Microsoft addressed over 100 CVEs in its August Patch Tuesday, including a publicly disclosed zero-day bug in SQL Server and 13 critical vulnerabilities. The patches cover various products, including Windows, SQL Server, and Microsoft Office. |
Share NerdNewsShare your affiliate link to get commission!
Thanks for reading,
If you dont want to receive future editions of NerdNews, |