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

Today:

• OpenAI Drops For-Profit Conversion Plan

• Anthropic Launches AI Science Program

• Nvidia Releases Open Transcription Model

• Cisco, Meta Launch Cybersecurity AI

• Anysphere Raises $900 Million Funding

OpenAI Drops For-Profit Conversion Plan

OpenAI has dropped plans to become a for-profit company and will stay under the control of its original nonprofit board. This decision means the company will keep prioritizing its mission to benefit humanity, rather than maximizing profit. However, it may face new challenges raising money, since it won't offer the kind of financial returns investors expect. The move reflects OpenAI’s continued focus on long-term ethical goals over short-term gains.

Why It Matters

1. Reinforces Ethical AI Development: OpenAI staying nonprofit signals a strong commitment to developing AI for public benefit, not just profit.
   

2. Investor Implications: This decision could limit OpenAI’s funding options, potentially slowing its ability to compete with heavily funded tech giants.
   

3. Sets a Precedent: It challenges the dominant model of profit-driven AI, potentially influencing how other leading AI labs think about governance.

Anthropic Launches AI Science Program

Anthropic has launched its AI for Science program to support scientific discovery using its powerful AI models. The initiative offers free API credits to researchers working on impactful projects, especially in biology and life sciences. The goal is to help scientists analyze data, design experiments, and accelerate discoveries. This move aligns with Anthropic’s mission to create AI that serves humanity and tackles major global challenges like disease and food security.

Why It Matters

1. Pushes AI Toward Real-World Impact: It shows how AI can directly accelerate science, not just software or business.
   

2. Strengthens Open Access AI in Research: Offering free access democratizes high-performance AI for academics and institutions.
   

3. Positions Anthropic as a Responsible Leader: It reinforces the company’s image as a mission-driven AI lab focused on global good.

Nvidia Releases Open Transcription Model

Nvidia released Parakeet-TDT-0.6B-v2, a fast and open-source speech recognition model that transcribes one hour of audio in just one second on GPU hardware. It achieves a strong 6.05% word error rate, rivaling commercial models like OpenAI's. Built on 120,000 hours of audio, it supports punctuation, timestamps, and commercial use. Optimized for Nvidia GPUs, the model is free under a permissive license, ideal for developers building transcription tools or conversational apps.

Why It Matters

1. High-Speed, Open-Source ASR: It democratizes access to near-commercial-level transcription performance without licensing costs.
   

2. Benchmarks OpenAI Alternatives: Shows Nvidia is a serious contender in AI models, not just chips, challenging proprietary systems.
   

3. Boosts AI Accessibility: Enables startups, researchers, and indie devs to build advanced voice tools using free, production-ready infrastructure.

🧠RESEARCH

PixelHacker: Image Inpainting with Structural and Semantic Consistency

PixelHacker is a new image inpainting model that fills in missing parts of pictures with strong attention to both structure (like shapes and textures) and meaning (like color and object accuracy). Using a custom dataset of 14 million images, it outperforms previous methods on popular benchmarks, producing more realistic and coherent results.

Improving Editability in Image Generation with Layer-wise Memory

This paper introduces a method for better image editing across multiple changes. Unlike current tools that often forget past edits, the new system uses “layer-wise memory” to remember earlier modifications and blend new ones smoothly. With just rough user input, it keeps images consistent and realistic through many editing steps, improving ease and quality.

Llama-Nemotron: Efficient Reasoning Models

NVIDIA’s Llama-Nemotron is a new family of open-source AI models built for fast and accurate reasoning. Available in three sizes, they match top models like DeepSeek-R1 but use less memory and run faster. With a unique reasoning toggle and full code and data released, they support both research and enterprise use.

🛠️TOP TOOLS

IMGCreator.ai - ImgCreator.ai is an AI-driven platform designed to transform text descriptions into visually captivating images, making it an ideal tool for creating illustrations, anime, concept art, and photorealistic visuals.

Namelix - Namelix is an AI-powered tool designed to simplify the process of generating creative, brandable business names.

ElevenLabs - ElevenLabs is a cutting-edge AI platform that offers text-to-speech and voice generation services, allowing users to create lifelike audio content in multiple languages and accents.

Vidyo.ai - Vidyo.ai is an AI-powered video repurposing platform that transforms long-form content into short, engaging clips optimized for social media.

Skipit.ai - Skipit.ai is an AI-powered content summarization tool designed to streamline information consumption across various digital formats.

📲SOCIAL MEDIA

🗞️MORE NEWS

• At RSAC 2025, Cisco and Meta unveiled open-source AI tools built specifically for cybersecurity. These models aim to unify security efforts across companies, enabling faster, cheaper, and smarter threat detection at machine-scale speeds.

• Anysphere, maker of AI coding tool Cursor, has raised $900 million at a $9 billion valuation. Backed by Thrive Capital, a16z, and Accel, the deal signals surging investor interest in AI-powered developer tools.

• Over 250 CEOs, including Microsoft’s Satya Nadella, signed an open letter urging U.S. schools to make AI and computer science part of K–12 education, warning that America risks falling behind global competitors.

• Google accidentally revealed "Material 3 Expressive," an upcoming Android design update focused on emotional, bold, and accessible user interfaces. Set to debut at Google I/O, it aims to boost usability—especially for older users.

• UnitedHealth now runs 1,000 AI tools across its operations, aiding in claims, transcriptions, and diagnosis. While AI supports decisions, the company insists it never directly denies claims, stressing a cautious, responsible approach amid growing legal scrutiny.