Introducing the first AI-native CRM

Connect your email, and you’ll instantly get a CRM with enriched customer insights and a platform that grows with your business.

With AI at the core, Attio lets you:

• Prospect and route leads with research agents

• Get real-time insights during customer calls

• Build powerful automations for your complex workflows

Join industry leaders like Granola, Taskrabbit, Flatfile and more.

👉Try Attio Pro for free

Today:

• Sutskever Deposition: A Year-Long Plot to Oust Sam Altman

• Perplexity Secures Getty Images Deal to Clean Up AI Content Use

• NVIDIA & Samsung Unite to Launch Massive AI Factory with 50,000 GPUs

• Adobe Max Unveils Game-Changing Creative AI Tools

• Meta Buys Nearly 1GW of Solar for AI

• Perplexity’s Free AI Tool Makes Patent Research Effortless

BREAKING: Ilya Sutskever DEPOSED, Sam Altman firing was planned a year in advance and more…

Ilya Sutskever’s deposition in Musk v. Altman reveals that OpenAI’s 2023 boardroom drama wasn’t sudden, it was the result of a year-long plan by Ilya and Mira Murati to remove Sam Altman. Ilya wrote a 52-page memo accusing Altman of lying, pitting executives against each other, and mismanaging power. He also made similar accusations against OpenAI president Greg Brockman. The board briefly replaced Altman with Emmett Shear, and even considered merging OpenAI with Anthropic, where former co-founder Dario Amodei had moved. 

Ilya admitted to relying on secondhand claims, mainly from Murati, without verifying them. He also noted board member Helen Toner’s pro-Anthropic views and her controversial statement suggesting OpenAI’s destruction might align with its mission. 

The deposition shows deep fractures, power struggles, and questionable judgment at the highest levels of AI leadership highlighting how personal conflicts and poor communication almost brought one of the world’s leading AI companies to collapse.

WATCH THE VIDEO ON YOUTUBE

Perplexity Secures Getty Images Deal to Clean Up AI Content Use

Perplexity, an AI search startup, signed a multi-year deal with Getty Images to legally display images across its platform. This follows past criticism over image use and plagiarism. The agreement highlights Perplexity’s push to legitimize content use and strengthen attribution in AI search.

KEY POINTS

• Perplexity’s Getty deal: Grants the startup permission to use Getty Images across its AI tools, moving away from past unauthorized usage.

• Context of controversy: Perplexity faced allegations of scraping content, plagiarism, and copyright infringement, including lawsuits and public backlash.

• Emphasis on attribution: The agreement includes linking image credits, aligning with Perplexity’s stance that it provides fair use of publicly available content.

Why it matters
This deal shows a shift in how AI companies must work with content creators instead of taking their work without permission. As AI tools grow, building trust through legal partnerships and giving credit helps avoid lawsuits and supports fair use in the digital age.

NVIDIA & Samsung Unite to Launch Massive AI Factory with 50,000 GPUs

NVIDIA and Samsung are building a powerful AI factory with 50,000 GPUs to revolutionize chipmaking, robotics, and smart devices. By combining digital twins, robotics, and accelerated computing, the facility will enable predictive maintenance, AI-powered design, and real-time factory automation for next-gen intelligent manufacturing.

KEY POINTS

• AI Factory Powerhouse: The facility will use 50,000 NVIDIA GPUs and Omniverse digital twins to optimize semiconductor manufacturing, reduce time-to-market, and automate decision-making processes.

• Smart Robotics & Mobile AI: Samsung will use NVIDIA tools (Isaac Sim, Jetson Thor) to deploy intelligent humanoid robots and enhance over 400 million devices with multilingual AI features.

• Longstanding Partnership: The collaboration builds on 25+ years of joint innovation—from early NVIDIA GPUs to next-gen HBM memory and AI-foundry services—marking a new AI industrial era.

Why it matters
This AI factory is a big step forward for how we build chips, smartphones, and even robots. Instead of waiting weeks to test designs or fix problems, factories will use digital twins and AI to make faster, smarter decisions. It’s the future of global manufacturing.

Adobe Max Unveils Game-Changing Creative AI Tools

At Adobe Max, Adobe previewed experimental AI tools that can edit entire videos using just one frame, reshape lighting in photos, and change how someone sounds in a video without re-recording. These powerful “sneaks” hint at the future of intuitive, AI-driven creative editing.

KEY POINTS

• Project Frame Forward: Edits made to a single video frame (like removing or inserting objects) automatically apply to the entire clip, eliminating manual masking or frame-by-frame changes.

• Generative Light & Audio Tools: “Project Light Touch” lets users shift lighting and shadow direction in photos. “Project Clean Take” edits speech tone, replaces words, and separates background audio.

• 3D and Texture Editing: Other tools allow real-time material swaps, object rotation in 2D images, and depth-aware editing that understands object positioning within a scene.

Why it matters
These AI tools can save creators hours of tedious editing. Instead of learning complex software steps, they can describe what they want and watch it happen instantly. It's a peek into a world where creative editing is faster, easier, and more accessible to everyone.

🧠RESEARCH

The End of Manual Decoding: Towards Truly End-to-End Language Models

Most "end-to-end" language models still rely on manually set decoding parameters like temperature. Tencent's AutoDeco changes this by letting models learn and adjust these settings themselves during text generation. It performs better than traditional methods and even understands natural language prompts to steer its randomness, enabling more flexible, human-like control.

Emu3.5: Native Multimodal Models are World Learners

Emu3.5 is a new multimodal AI model trained on 10 trillion vision-language tokens from internet videos. It can understand and generate mixed image-text content, predict future states, and perform open-world tasks. With its fast, parallel inference method (DiDA), Emu3.5 excels in generation, editing, and world-modeling—rivaling top models like Gemini 2.5.

Kimi Linear: An Expressive, Efficient Attention Architecture

Kimi Linear is a new attention architecture that beats standard full attention models across various tasks and input lengths. It uses a more efficient memory and decoding method, reducing resource use while improving performance. The model also boosts speed up to 6× and is open-sourced for research and application.

📲SOCIAL MEDIA

🗞️MORE NEWS

• Meta bought nearly 1 gigawatt of solar power this week to support its growing AI needs. Critics say its reliance on energy certificates hides true emissions, urging direct support for new renewable projects instead.

• Perplexity launched a free AI tool that simplifies patent research by letting users search in plain language. It also finds related terms and prior art across patents, academic papers, and code—making complex searches more accessible.

• Xania Monet, an AI-created R&B singer, became the first AI artist to chart on Billboard through real radio airplay. Despite industry backlash, her success signals growing public acceptance of AI in mainstream music.

• Columbia University researchers achieved the first pregnancy using STAR, an AI-powered method to find rare sperm in men with azoospermia. The breakthrough may offer new hope to couples facing male-factor infertility.

• OpenAI now lets users buy extra Sora video credits as it plans to cut free allowances. The move supports Sora’s shift toward monetization, aiming to build an AI-powered creator economy with paid features and tools.