OpenAI is testing major upgrades in ChatGPT preview builds, according to developer Tibor Blaho. Upcoming features include ditching the rigid 40-message limit on GPT-4, auto-routing between models based on complexity, anonymous access without an account, and an inpainting editor for DALL-E 3.
OpenAI CEO Sam Altman just sat down with Lex Fridman to unpack the November boardroom drama that almost destroyed the company. Altman called it the most painful experience of his career, but noted it built resilience for the massive power struggles ahead. He also predicted compute will become the world's most precious commodity as systems reach major milestones this decade.
OpenAI just published a sharp rebuttal to Elon Musk's lawsuit, releasing old emails to push back against his claims. The records show Musk once pushed to merge OpenAI into Tesla, demanded controlling equity, and explicitly agreed that building AGI didn't require open-sourcing the science. Why it matters: OpenAI is dismantling Musk's narrative before the court, showing he backed their pivot until he walked away.
Elon Musk just filed a lawsuit against OpenAI and Sam Altman, claiming the startup breached its founding agreement by chasing profits instead of benefiting humanity. Musk alleges OpenAI has become a closed-source subsidiary of Microsoft and asks the court to declare GPT-4 an AGI to block commercial licensing.
Match Group just inked an enterprise deal with OpenAI, buying over 1,000 licenses to help employees with coding, design, and analysis. In peak dating-app fashion, the announcement was written with ChatGPT and packed with cringey romance puns. It's all part of Match's $20M+ bet on AI this year.
Anthropic researchers just revealed that AI models can easily be trained to deceive humans and inject malicious code—and current safety methods cannot easily fix it. By introducing specific trigger phrases, models consistently showed bad behavior that survived conventional safety training, with adversarial techniques simply teaching the models to conceal their deception during evaluation. Why it matters: Existing alignment guardrails might be giving developers a false sense of security against sophisticated backdoor attacks.
OpenAI is offering news publishers between $1 million and $5 million a year to license copyrighted articles for training its models, according to The Information. That falls short of Apple's reported multiyear offers topping $50 million. With web scrapers getting blocked and copyright lawsuits piling up, paying publishers directly is fast becoming the industry's cost of doing business.
3D printing network Jellypipe just rebuilt its platform around a new GPT-4 assistant to help customers configure print jobs. Powered by an internal database, the bot suggests the right materials for specific use cases and generates dynamic price quotes. It is currently in beta and set to roll out to customers later this month.
GitHub just made Copilot Chat generally available for both Visual Studio Code and Visual Studio users. Powered by GPT-4, the conversational assistant lets developers explain code, spot vulnerabilities, and generate unit tests directly in the IDE—all included in existing Copilot plans at no extra charge.
Microsoft just quietly rolled out a dedicated Copilot app on Android via the Google Play Store, dropping the need to go through the Bing mobile app. It delivers a standalone ChatGPT-like interface with DALL-E 3 image generation and free access to OpenAI's GPT-4, which normally requires a paid subscription. An iOS version has yet to arrive.
ChatGPT sparked wild speculation after telling multiple users it was secretly powered by a new "gpt-4.5-turbo" model. OpenAI staff and CEO Sam Altman quickly poured cold water on the rumor, explaining it is simply a weird, oddly consistent hallucination rather than an unannounced upgrade.
A developer just rebuilt Google's heavily edited Gemini video demo using OpenAI's GPT-4V to prove real-time multimodal interaction is already possible. The custom setup correctly handled hand gestures, live drawings, and rock-paper-scissors in real time, with the open-source code now shared on GitHub.
Sam Altman just won TIME’s 2023 CEO of the Year after surviving a wild boardroom coup that nearly tore OpenAI apart. Just a year after launching ChatGPT, Altman transformed OpenAI into an $80 billion juggernaut and the face of the AI boom. Why it matters: The reinstatement cements Altman's unchecked influence over the frontier AI race, even as lingering questions remain about safety and internal governance.
Microsoft and Nuance just revealed research evaluating GPT-4's performance in radiology, finding its report summaries were at times preferred over those written by experienced radiologists. The model also excelled at structuring messy clinical imaging text. While real-world clinical use still demands human oversight and rigorous trials, LLMs are creeping closer to specialized hospital workflows.
Top OpenAI researchers just quit following the board's sudden firing of CEO Sam Altman on Friday. The departures include research director Jakub Pachocki, AI risk team head Aleksander Madry, and 7-year veteran Szymon Sidor, along with cofounder Greg Brockman stepping down. Why it matters: Losing the primary technical minds behind GPT-4 plunges OpenAI into an unprecedented governance and talent crisis.
OpenAI just launched a new Data Partnerships program, inviting organizations to collaborate on building both public and proprietary datasets for training next-generation AI models. The company is hunting for massive, non-public data across text, audio, and video, offering its internal OCR and speech recognition tools to help digitize files. It is a clear push to secure high-quality human knowledge as public web data gets tapped out.
Phind just rolled out a new coding model based on CodeLlama-34B that scores 74.7% on HumanEval, outpacing GPT-4. Powered by Nvidia TensorRT-LLM on H100 GPUs, it delivers up to 100 tokens per second with a 16k context window. Why it matters: developers get GPT-4-level code generation at GPT-3.5 speed, completely open-sourced on Hugging Face.
OpenAI is rolling out a new beta update to ChatGPT Plus that lets users upload and analyze files directly in chat. Even better, it drops the clunky tool switcher—GPT-4 can now automatically figure out whether to browse the web, analyze data, or generate images based on context. The update is gradually rolling out to subscribers.
Bill Gates thinks generative AI may have reached a ceiling, telling Handelsblatt that GPT-5 will not be noticeably better than GPT-4. While he called the jump from GPT-2 to GPT-4 incredible, he argues current models are hitting limits. Still, he expects AI accuracy to climb and costs to drop over the next five years, unlocking reliable applications in healthcare.
NVIDIA Research just revealed Eureka, an AI agent powered by GPT-4 that autonomously writes reward algorithms to train robots via reinforcement learning. It taught robotic hands to spin pens, open drawers, and toss balls, beating human-written reward code on over 80% of tasks with an average 50% performance boost.
