Google Launched "Vision Match" 🛍️, OpenAI Launched NextGenAI 💡, Google AI Mode In Search 🌐

In partnership with

Find out why 1M+ professionals read Superhuman AI daily.

In 2 years you will be working for AI

Or an AI will be working for you

Here's how you can future-proof yourself:

1. Join the Superhuman AI newsletter – read by 1M+ people at top companies

2. Master AI tools, tutorials, and news in just 3 minutes a day

3. Become 10X more productive using AI

Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.

Sign up and start learning AI

💬 OpenAI Developments

Introducing NextGenAI: A consortium to advance research and education with AI

OpenAI has launched NextGenAI, a consortium of 15 leading research institutions dedicated to advancing AI in research and education. With a $50 million commitment in research grants, compute funding, and API access, the initiative aims to accelerate scientific breakthroughs, enhance AI literacy, and integrate AI into universities, libraries, and hospitals. Founding partners include Harvard, MIT, Oxford, Texas A&M, and Duke, among others, leveraging AI for medical research, digital learning, and operational efficiency. NextGenAI strengthens academia-industry collaboration, ensuring AI benefits extend to researchers, educators, and students globally.

OpenAI’s GPT-4.5 AI model comes to more ChatGPT users

OpenAI has started rolling out GPT-4.5 to ChatGPT Plus users, following its debut for Pro subscribers last week. While larger and more powerful, GPT-4.5 falls short of some newer AI reasoning models and is costly to run, leading OpenAI to reconsider its long-term API availability. Despite this, the model boasts deeper world knowledge, higher emotional intelligence, and fewer hallucinations, making it more reliable. Notably, it excels in persuasion, even convincing other AIs to give it money and secrets in internal tests.

Key ex-OpenAI researcher subpoenaed in AI copyright case

Ex-OpenAI researcher Alec Radford has been subpoenaed in a copyright lawsuit against OpenAI, filed by authors alleging their works were used without permission to train AI models like ChatGPT. Radford, a key developer of GPTs, Whisper, and DALL-E, left OpenAI last year. The case, which claims direct infringement, also seeks depositions from other ex-OpenAI employees, including Anthropic’s Dario Amodei. OpenAI argues its data use falls under fair use, while the court has allowed parts of the lawsuit to proceed.

🚀 Tech Industry Moves

Apple is introducing AI-powered App Store review summaries in iOS 18.4, using Apple Intelligence to generate weekly refreshed summaries from user reviews. The feature, launching first in the U.S. in English, will expand globally over time. While it aims to help users quickly understand app feedback, concerns arise about fake reviews influencing AI summaries. Similar AI-powered review analysis features exist on Amazon, Google Maps, and Gemini AI. Currently in beta, the feature will roll out publicly in April alongside other AI enhancements in iOS 18.4.

Google has launched AI Mode in Search, allowing users to ask complex, multi-part questions and follow up for deeper exploration. Powered by Gemini 2.0, AI Mode helps compare products, analyze real-time data, and provide detailed, fact-backed responses. Currently available to Google One AI Premium subscribers via Search Labs, the feature prioritizes accuracy and references sources when needed. Google is also rolling out Gemini 2.0 for AI Overviews, improving coding, math, and multimodal queries while making AI Overviews available without sign-in and expanding access to teen users.

Google has launched "Vision Match" in its Shopping tab, an AI-powered tool that helps users find clothing based on their descriptions. Available on mobile for U.S. users, shoppers can type their fashion ideas, and AI suggests similar items to buy. Additionally, Google is expanding its AR beauty feature, allowing users to virtually try on multiple makeup products at once, and enhancing its virtual try-on experience to include pants and skirts on models of different sizes. These updates continue Google's push to integrate AI into Shopping, following last year’s AI product summaries in search results.

Anthropic has quietly removed several Biden-era AI safety commitments from its website, including pledges on AI risk management and bias research, while still maintaining commitments on AI-generated abuse. This follows the Trump Administration’s repeal of Biden’s AI Executive Order, replacing it with policies focused on AI development without ideological bias. Other AI firms, including OpenAI, have also adjusted their public stances, with OpenAI removing DEI commitments and emphasizing intellectual freedom. Despite these changes, Anthropic later stated that it remains committed to the voluntary AI commitments and will clarify this in its transparency center.

Microsoft: The UK's Competition and Markets Authority (CMA) ruled that Microsoft’s partnership with OpenAI doesn’t qualify for investigation under UK merger laws, as Microsoft influences but doesn’t control OpenAI’s commercial policy. The CMA initially investigated concerns over competition risks, but noted recent changes in the partnership, including Microsoft’s reduced exclusivity as OpenAI’s cloud provider and its withdrawal from OpenAI’s board. This decision aligns with the CMA’s broader scrutiny of AI investments, though it has yet to find evidence of anti-competitive behavior.

CoreWeave has acquired AI developer platform Weights & Biases for a reported $1.7 billion, enhancing its AI cloud services. Weights & Biases, used by 1,400+ organizations including Nvidia and AstraZeneca, provides tools for AI model training and fine-tuning. The acquisition aims to streamline AI development workflows and accelerate innovation, while customers can still deploy workloads on their preferred platforms.

Meta is expanding its facial recognition tools to the UK and EU, aimed at combating scam ads using celebrity likenesses and helping users recover compromised accounts. Initially launched in October 2024, the system uses AI-powered selfie verification and scam detection while ensuring facial data is deleted after verification. The move follows regulatory approvals in the UK and comes as Meta doubles down on AI development, including a stand-alone AI app. Despite past controversies over facial recognition, Meta hopes these tools will improve user security and trust in its platform.

💡 Emerging AI Technologies and Companies

Anna Patterson’s Ceramic.ai looks to help enterprises build AI models faster and more efficiently

Anna Patterson's new startup, Ceramic.ai, aims to help enterprises train large language models (LLMs) faster and more efficiently using fewer GPUs. Founded in January 2024, the company focuses on scalable AI infrastructure, claiming to improve model scaling by 100x. Backed by $12M in seed funding from NEA, IBM, Samsung Next, and others, Ceramic has partnered with AWS and Lambda but isn’t generating revenue yet. Competing with Together AI and MosaicML, Ceramic seeks to establish itself in the highly competitive AI infrastructure space by prioritizing early product feedback and rapid iteration.

GibberLink lets AI agents call each other in robo-language

GibberLink, a viral hackathon project by two Meta engineers, enables AI agents to communicate in a robotic language using GGWave, an efficient audio-based protocol. When two AI agents detect they’re talking to each other, they switch from human speech to "beeps and boops", reducing computation costs significantly. Inspired by dial-up modem handshakes, this open-source project could improve AI-to-AI interactions, especially in call centers and AI assistants. Though currently non-commercial, the demo has gained 15M+ views, sparking curiosity—and even a meme coin.

CoreWeave partner EcoDataCenter racks up half a billion dollars to build more sustainable buildings for AI

EcoDataCenter (EDC), a Swedish eco-friendly data center builder, has raised $478 million to expand sustainable infrastructure for AI computing. A key partner of CoreWeave, EDC focuses on renewable energy, efficient cooling, and innovative materials like cross-laminated timber—a model now followed by Microsoft. The funding will support new green data center technologies amid rising AI demand. EDC is also hosting Europe’s first Blackwell cluster with CoreWeave and Nvidia in Sweden, reinforcing its role in the AI compute boom.

🎨 AI in Creative Industries

Prime Video is testing AI-assisted dubbing for select movies and TV shows, starting with 12 titles in English and Latin American Spanish. The service combines AI with human oversight to ensure quality. AI dubbing is gaining traction in the industry, with companies like Deepdub and YouTube also investing in similar technologies. Prime Video already uses AI for features like X-Ray Recaps and Dialogue Boost to enhance the viewing experience.

🌍 Global AI Race

ARM to sign $250M chip deal with Malaysia

ARM Holdings has signed a $250 million deal with Malaysia to support its chip design ecosystem, providing seven chip design blueprints and training 10,000 engineers over 10 years. This aligns with Malaysia’s ambition to become a global chip hub, backed by its $5.3 billion National Semiconductor Strategy. With Intel, GlobalFoundries, and other major players investing heavily in chip production and AI infrastructure, Malaysia is leveraging its longstanding role in semiconductor manufacturing to attract global tech giants like Google, Microsoft, and Nvidia.

If you want more updates related to AI, subscribe to our Newsletter