- Work With AI
- Posts
- Leaked internal emails from Musk’s lawsuit against OpenAI reveal major conflicts
Leaked internal emails from Musk’s lawsuit against OpenAI reveal major conflicts
Plus Perplexity introduces shopping feature for pro users
Today’s Highlights:
📰 News: Perplexity introduces shopping feature for pro users + Leaked internal emails from Musk’s lawsuit against OpenAI reveal major conflicts
💰 Funding: SuperAnnotate raises $36M Series B for managing large-scale multimodal AI datasets
⚡️ Top News Stories:
1. Perplexity has launched a shopping feature for U.S. Pro users, allowing product recommendations and one-click purchases with tools like detailed product cards, reviews, and quick checkout options, while partnering with Shopify and other merchants to provide free API access, enhance search indexing, and streamline the purchase flow.
2. Internal emails from Musk’s lawsuit against OpenAI reveal major conflicts, with leadership battling over AGI control, pay scales, and Microsoft partnerships.
Major conflicts between Elon Musk, Sam Altman, and Ilya Sutskever highlight the major tensions during its early years, including over control of AGI, Musk’s desire for unilateral power, and philosophical differences about OpenAI’s mission.
Musk backed pay hikes of $100K–$200K aimed to counter DeepMind’s poaching attempts and retain AI talent, reflecting the intense competition between the two organizations
Musk opposed early deals with Microsoft, including a rejected $60M Azure offer, fearing OpenAI would become their marketing pawn, and even offered $50M to prevent the partnership.
Musk criticized OpenAI's shift towards commercialization and for-profit structures, expressing concerns that it deviated from its original nonprofit, safety-first mission for AGI.
Sutskever and others warned against potential “AGI dictatorship” under Musk and questioned Altman’s commitment to OpenAI’s founding principles, exacerbating internal mistrust.
3. Mistral’s Le Chat platform introduces multiple new features, including web searches with citations, a “canvas” tool for in-place content modification, and the ability to analyze large PDFs and images, including charts and equations.
Mistral unveiled Pixtral Large, a 124-billion-parameter multimodal model that excels in understanding documents, charts, and natural images, and Mistral Large 24.11, a text-only model with improved long-context capabilities.
Le Chat now hosts shareable workflows, or “AI agents,” for automated tasks like processing expense reports and invoices, showcasing Mistral’s focus on productivity enhancements.
4. TikTok for Business has launched Symphony Creative Studio, an AI-powered platform that generates TikTok-ready videos in minutes from simple inputs like URLs or product details, offering features like customizable digital avatars, video translation in 30+ languages, and advanced editing tools, while ensuring transparency by labeling AI-generated content and implementing safeguards to protect intellectual property and ensure ethical use.
5. TSMC has secured $11.6Bn in U.S. funding under the Chips Act, including $6.6Bn in grants and $5Bn in loans, to support its $65Bn Arizona project— the largest foreign manufacturing investment in U.S. history—focused on producing advanced 3nm AI chips across three fabs.
6. Stripe has introduced a developer guide for integrating its payment solutions into agentic workflows powered by LLMs, enabling seamless, API-driven transaction flows where AI agents can request, confirm, and process payments, with use cases like virtual assistants managing e-commerce purchases or subscription payments autonomously.
7. Final Cut Pro 11 introduces cutting-edge AI tools, including the Magnetic Mask for subject isolation without a green screen and Transcribe to Captions for automatic captioning.
8. Scientists at the Arc Institute have developed Evo, a groundbreaking AI model trained on 2.7M microbial genomes, capable of simultaneous learning from DNA, RNA, and protein sequences, with early tests showcasing its ability to design genetic editing tools, predict bacterial DNA changes, and generate synthetic genome-length sequences over 1M base pairs.
9. Stanford Medicine researchers have developed SEQUOIA, an AI tool that analyzes tumor biopsy images to predict gene activity linked to cancer growth, metastasis, and treatment response, bypassing the need for molecular testing which streamlines diagnosis and personalized treatment planning. It has demonstrated strong predictive performance across cancers like breast, lung, and colon.
10. MIT Jameel Clinic has launched Boltz-1, the first open-source biomolecular interaction model matching AlphaFold3-level accuracy; with training data, model weights, and code released under the MIT license it aims to democratize advanced biomolecular modeling and foster global collaboration in drug design and structural biology.
11. The NIH-developed AI algorithm TrialGPT streamlines clinical trial enrollment by analyzing patient summaries and generating annotated, ranked lists of relevant studies from ClinicalTrials.gov, matching clinicians' accuracy while reducing screening time by 40%, with the potential to boost enrollment efficiency, accelerate medical research, and improve access for underrepresented populations.
12. The University of Chicago's SAND Lab developed Glaze and Nightshade, tools to protect artists' work from AI scraping by masking images and injecting "poison" to mislead generative AI, with over 6M and 1.6M downloads respectively, the aim is help pressure AI companies into negotiating fair compensation for using copyrighted material.
13. Alibaba, ByteDance, and Meituan are aggressively hiring AI talent in California to advance their generative AI capabilities, despite US export bans on high-end Nvidia chips.
14. Nvidia's new Blackwell AI chips are facing overheating issues, especially in high-density server racks with up to 72 chips, prompting multiple redesigns from suppliers and potentially delaying data center deployments for cloud customers like Meta, Google, and Microsoft who depend on them for scaling AI infrastructure.
15. HarperCollins has partnered with an unnamed AI company to offer authors $2,500 for licensing select nonfiction backlist titles for AI training over three years, requiring authors to opt in voluntarily under "clear guardrails" to protect their rights and preserve their works' value.
16. NVIDIA is partnering with Google Quantum AI to accelerate next-generation quantum computer design using its CUDA-Q platform and Eos supercomputer, enabling simulations of quantum devices with up to 40 qubits, achieving weeks-long simulations in minutes while addressing noise, a key challenge in scaling quantum hardware.
17. A University of Pittsburgh study involving over 2,300 participants found that AI-generated poetry often outperformed works by iconic poets like Shakespeare and Dickinson, receiving higher ratings on rhythm, beauty, and emotional impact across 13 qualitative measures, with readers accurately distinguishing AI from human poetry only 46.6% of the time; however, when told poems were AI-generated, participants rated them lower, despite preferring AI poems in blind evaluations for their accessible emotional communication and a “more human than human” effect.
💰 Top Funding News:
1. Google’s philanthropic arm is committing $20M in cash and $2M in cloud credits to support around 15 global research organizations by 2026, leveraging AI to address interdisciplinary challenges in fields like disease research, materials science, sustainability, and experimental biology.
2. SuperAnnotate, which provides enterprise-grade tools for creating and managing large-scale multimodal AI datasets, raised a $36M Series B, led by Socium Ventures, w/ NVIDIA, Databricks Ventures, Play Time Ventures (Lionel Messi's fund), Glynn Capital, and existing investors.
3. Moonvalley, which is developing generative AI models for video creation with a specific focus on ethical practices, raised a $70M Seed round, co-led by General Catalyst and Khosla Ventures, w/ Bessemer Ventures.
4. Juna.ai, which utilizes AI to revolutionize industrial process control by integrating machine learning with real-time operational data, raised a $7.5M Seed round, led by Boldstart Ventures, w/ Root Ventures, Flybridge, and angel investors.
That's all for today's email! If you want more please follow us at the social channels linked below, or check out our website!
How'd you like today's email? |
Share our newsletter: If you like our work please share/forward this email with your friends, colleagues, and family. It's the best way to support us!
If this email was forwarded to you please sign up here to continue receiving them.
Want your content, product, jobs, or event featured in our newsletter? Reply to this email with the details, and our team will reach out to you.
Do you use AI for work? Tell us how, and you could be featured in our newsletter!
Check out our website for more resources, including a list of AI investors, products, events, and twitter follows.
For an archive of all our posts, click here.
We'd love to hear from you! You can always leave us comments or feedback by replying to this email!