Creators Grapple With Open Access Norms In AI Era

Also, Baidu launches Ernie 4.5

Author

Jack Lajoie
March 17, 2025 • read time ~ 9 minutes

⚡️ Headlines

🤖 AI

Foxconn's Fourth-Quarter Profit Misses Expectations Despite Revenue Gain - Foxconn reports a 13% drop in net profit to NT$46.33 billion, missing expectations despite a 15% revenue increase, and forecasts strong 2025 growth driven by AI server demand. [The Wall Street Journal].

No One Knows What the Hell an AI Agent Is - The tech industry grapples with defining 'AI agents,' reflecting confusion over their capabilities and roles. [TechCrunch].

CoreWeave Serves as Bellwether for AI in Soft IPO Market - CoreWeave's planned IPO tests investor appetite for AI-focused companies amid a sluggish market. [Semafor].

The Russo Brothers Upended Hollywood Once. Now They Aim to Do It Again. - Filmmakers Anthony and Joe Russo plan to build a high-tech studio leveraging AI to empower artists and explore 'transmedia' storytelling. [The Wall Street Journal].

Manus AI Pushes the DeepSeek Moment Further - Manus AI advances the 'DeepSeek moment,' highlighting significant progress in artificial intelligence applications. [Bloomberg Opinion].

DeepSeek AI Founder Liang Wenfeng: China’s New Superhero - Liang Wenfeng's low-cost AI model propels him to national hero status in China, inspiring a local tourism boom around his origins. [The Times].

AI Nurses and Hospitals: The Future of Health Care - Hospitals integrate AI-powered nurses to enhance patient care and address staffing shortages. [Associated Press].

🦾 Emerging Tech

BotQ: A High-Volume Manufacturing Facility for Humanoid Robots - Figure announces BotQ, a facility capable of producing up to 12,000 humanoid robots annually, with plans for significant scaling. [Figure].

Crypto Daybook Americas: Recession Concerns Stifle Bitcoin Recovery Prospects, Memecoins Buzz - Economic recession fears hinder Bitcoin's recovery, while memecoins gain attention in the crypto market. [CoinDesk].

🤳 Social Media

Bluesky Users Debate Plans Around User Data and AI Training - Bluesky's community discusses strategies for handling user data in AI training contexts. [TechCrunch].

Instagram Tests AI-Generated Comments, Posts - Instagram experiments with AI-generated content to enhance user engagement. [Social Media Today].

🔬 Research

TxAgent: An AI Agent for Therapeutic Reasoning Across a Universe of Tools - Harvard researchers introduce TxAgent, an AI designed for personalized treatment recommendations through multi-step reasoning and real-time biomedical knowledge retrieval. [Zitnik Lab, Harvard Medical School].

⚖ Legal

The Court Rejects Elon’s Latest Attempt to Slow OpenAI Down - A court dismisses Elon Musk's request for a preliminary injunction against OpenAI, allowing the organization to proceed with its initiatives. [OpenAI].

Virginia General Assembly Advances AI Oversight in Court Decisions - Virginia's legislature progresses bills to oversee AI use in court decisions and facial recognition technology. [VPM].

🎱 Random

Quest's Passthrough Camera API Is Out Now for Developers to Experiment With - Meta releases an experimental Passthrough Camera API for Quest, enabling developers to access headset cameras for custom computer vision applications. [UploadVR].

🔌 Plug-Into-This

In this article, Molly White—a software engineer, writer, and prominent critic in the web3 world—explores the unintended consequences of open access initiatives in the context of generative AI. She highlights how large corporations exploit freely licensed works without contributing back to the community, leading creators to reconsider the terms of open access.

  • Creators often experience "wait, no, not like that" moments when their freely licensed work is commercially exploited without reciprocation.

  • In reaction, some creators consider adopting more restrictive licenses or withdrawing their work from open platforms, potentially undermining the principles of open access.

  • White argues that tightening licenses may be ineffective, as major AI companies have been known to train models on copyrighted material regardless of licensing terms.

  • She suggests that the focus should shift from restricting access to ensuring that open projects receive adequate support and recognition, thereby sustaining the ecosystem of free knowledge.

  • The piece warns that overemphasizing control could lead to the erosion of the commons, ultimately harming the accessibility and collaborative nature of shared knowledge.

👐 Balancing the ideals of open access with the realities of AI-driven exploitation requires innovative approaches that protect creators' interests without stifling the free flow of information.

Chinese tech giant Baidu has introduced two new artificial intelligence models—ERNIE 4.5 and ERNIE X1—to enhance its position in the competitive AI landscape. ERNIE X1 is designed as a reasoning-focused model, claiming performance on par with DeepSeek's R1 but at half the cost. ERNIE 4.5, the latest iteration of Baidu's foundational model, boasts advanced multimodal understanding and language capabilities.

  • ERNIE X1 exhibits enhanced abilities in understanding, planning, reflection, and evolution, marking it as Baidu's first deep-thinking model capable of autonomous tool usage.

  • ERNIE 4.5 demonstrates significant improvements in language understanding, generation, logic, and memory, alongside a high emotional quotient, enabling it to comprehend internet memes and satirical content effectively.

  • Baidu plans to integrate these models across its ecosystem, including its search services and the Wenxiaoyan app, aiming to bolster user experience and engagement.

  • Both models are now freely accessible to individual users, reflecting Baidu's strategy to promote widespread adoption and gather diverse user feedback for continuous improvement.

🇨🇳 China continues to ship, now seeming set to compete within their own domestic industry, beyond disrupting American Big Tech. Time will tell if there are meaningful quality reductions implied by their cost-cutting approaches, but for now, it’s making the high budget development from USA based firms look worse and worse.

Y Combinator (YC) reports that its current cohort of startups is experiencing unprecedented growth and profitability, largely attributed to advancements in artificial intelligence (AI). The accelerator's president, Garry Tan, highlights that AI's rapid integration across various sectors has enabled startups to scale operations and innovate more efficiently.

  • Many YC-backed companies are leveraging AI to automate complex tasks, reduce operational costs, and enhance product offerings, leading to accelerated market entry and expansion.

  • The surge in AI-driven innovation has attracted increased investor interest, resulting in higher valuations and funding rounds for these startups.

  • YC's focus on nurturing AI-centric startups aligns with global trends, where businesses are increasingly adopting AI technologies to maintain competitiveness.

📈 The integration of AI is not only transforming individual startups but also redefining industry standards, compelling traditional businesses to adapt to the evolving technological landscape.

 🆕 Updates

📽️ Daily Demo

🗣️ Discourse