- The Current ⚡️
- Posts
- Artists Leak OpenAI's Sora Amid Ethical Concerns & Anthropic Enhances Claude with Custom Styles
Artists Leak OpenAI's Sora Amid Ethical Concerns & Anthropic Enhances Claude with Custom Styles

⚡️ Quick Hits
🤖 Artificial Intelligence
A16z Considers Leading a $200 Million Round in Black Forest Labs - It would give the startup a valuation of more than $1 billion [Yahoo Finance].
Zoom Rebrands as AI-First Company - Now operating as Zoom Communications Inc. to expand beyond video conferencing [Zoom Blog].
AI ‘Grandma Daisy’ Fights Scammers in UK Pilot Program - Virgin Media O2 tests AI to deter fraud by imitating loved ones [CBS News].
Intel Secures $7.9 Billion Subsidy for U.S. Chip Plant - Part of government efforts to boost domestic semiconductor manufacturing [Wall Street Journal].
OpenAI and Partners Train AI Models on African Languages - Collaborative effort to preserve linguistic diversity in AI [Bloomberg].
🎨 Creative & Cultural
Tech Giants Exploit Creative Data for AI, Australian Inquiry Finds - Report warns of cultural harm from unregulated data scraping [The Guardian].
Australia Senate Committee Supports Social Media Age Limits - Proposed law aims to protect children under 13 from social media use [Reuters].
Opinion: Why Parents Should Support the Social Media Bill - A parent’s shift in perspective on regulating kids’ online access [The Guardian].
₿ Crypto & Blockchain
Bitcoin Long-Term Holders May Sell Another 163k BTC - Analysts note a potential market impact from increased sell-offs [CoinDesk].
Kraken Closes NFT Marketplace Amid Industry Challenges - The exchange cites low demand for NFTs in its decision [Bloomberg].
30 Tech Founders Were Secretly Debanked During Biden Administration - Elon Musk’s post has been confirmed by several industry heads [X/Twitter].
⚖ Legal & Regulation
Google Subpoenas OpenAI, Microsoft in Antitrust Case - Legal battle intensifies with rival AI firms called to testify [The Information].
Amazon Influencers Face Copyright Lawsuit Over ‘Clean Aesthetic’ Images - Creators allege their content was misused for marketing [The Verge].
Objection to The Onion’s Infowars Acquisition Highlights Social Media Account Ownership Issues - Debate over digital identity and account control resurfaces [404 Media].
Elizabeth Warren Pushes for Action on Internet Domain Monopoly - Urges stricter regulation of Verisign for its control over .com domains [Wired].
🎱 Random
LinkedIn Experiments With AI-Generated Influencers - Platform explores automated content creators to boost engagement [Wired].
Perplexity CEO Teases $50 Hardware AI Device - after his post gained traction, Srinivas seems intent on developing a question answering wearable [X/Twitter].
🔌 Plug In To These Details
A group of artists has leaked OpenAI's unreleased video generator, Sora, on Hugging Face, allowing users to create 10-second videos up to 1080p resolution. This action appears to be a protest against OpenAI's alleged unfair practices, including unpaid labor and misleading public relations strategies regarding the tool's development and capabilities.

The leaked project on Hugging Face allows users to generate 10-second videos up to 1080p resolution using Sora's API.
The group claims OpenAI is pressuring early testers to promote Sora as useful for artists.
Allegations state that hundreds of artists provided unpaid feedback and bug testing for Sora.
The leakers accuse OpenAI of being deceptive about which Sora-generated videos are shared publicly.
All Sora outputs reportedly need approval from the OpenAI team before sharing.
Sama waking up and realizing beta testers f***ing leaked OpenAI Sora through the API on Huggingface.. This is event of the year...
Whoever did this thank you.— AshutoshShrivastava (@ai_for_success)
4:32 PM • Nov 26, 2024
🔍 Since it’s initial announcement, Sora has been a battleground for the debate over ethical practices in content creation and video tools purportedly trained on artist work. This leak is an intriguing new development in that story, and it seems artists are willing to be more and more bold in their opposition to OpenAI.
Claude, Anthropic's AI assistant, has been updated with "custom instructions," allowing users to personalize its communication style. This new feature enables Claude to adapt its responses to different scenarios, enhancing its versatility for diverse user needs. Anthropic emphasizes that while the presentation of information can be customized, Claude's underlying capabilities and ethical training remain unchanged.

Users can upload writing samples to create custom styles tailored to their preferences.
The update aims to make Claude more versatile for different use cases, from casual to professional contexts.
Anthropic ensures that custom instructions cannot override Claude's ethical training or core knowledge.
The feature allows saved styles to be reused across conversations for consistent interactions.
Want Claude to more closely match how you communicate?
Upload writing samples and Claude can automatically generate custom styles, just for you.
— Anthropic (@AnthropicAI)
6:16 PM • Nov 26, 2024
🤔 The introduction of customizable AI personas highlights the evolving relationship between humans and AI, where user control over AI behavior is becoming increasingly important.
Elon Musk has announced plans to establish an AI game studio through his company xAI. The initiative aims to challenge the dominance of corporate-owned game studios and "make games great again."
Musk's announcement came in response to claims that game developers and journalists are "ideologically captured."
This move aligns with Musk's recent focus on artificial intelligence, including the development of the Grok AI chatbot and rapid scaling of AI infrastructure.
Musk aims to challenge game studios owned by large corporations with this new AI game studio initiative.
No details are solid, but the community is already abuzz.
Too many game studios that are owned by massive corporations.
@xai is going to start an AI game studio to make games great again!
— Elon Musk (@elonmusk)
3:55 PM • Nov 27, 2024
🎮 Musk's AI game studio announcement underscores his ambitious expansion into diverse tech sectors, but the lack of concrete details and his track record of unfulfilled promises warrant cautious optimism. Is it fair to say that the gaming industry has experienced corporate capture?
📸 Creator Corner
Sora's Supposed Features
Text-to-video generation: Sora can create high-quality videos from text prompts
Duration: Capable of producing videos up to one minute long
Resolution: Outputs videos in high definition, up to 1080p
"Turbo" variant: A faster version of Sora was reportedly part of the leak, suggesting improved processing speeds
Advanced capabilities: Sora can generate complex scenes with multiple characters, specific camera movements, and detailed backgrounds
Every OpenAI Sora Leak:
1.
— AI Central (@AICentral_)
8:06 AM • Nov 27, 2024
Key Points of The Controversy
Exploitation of artists:
The leakers claim OpenAI exploited artists for unpaid labor, using them for bug testing, feedback, and experimental work
OpenAI argues participation was voluntary, with no obligation to provide feedback
Compensation and recognition:
Artists feel undervalued, providing work for a company valued at $150 billion without fair compensation
OpenAI states they offer free access and support through grants, events, and other programs
Content control and approval:
Testers claim OpenAI requires approval for all Sora-generated content before sharing
This raises concerns about creative freedom and the true purpose of the early access program
Impact on creative industries:
Fears that tools like Sora could devalue human creativity and replace traditional artistic processes
Potential to reshape entire industries, including filmmaking and animation
Democratization vs. devaluation:
While AI tools may make content creation more accessible, there are concerns about the quality and originality of AI-generated work
Copyright and intellectual property:
Ongoing discussions about the use of copyrighted materials in training AI models
Job displacement:
Concerns that AI video generators could replace human artists and filmmakers
Broader AI Generator Debate
Quality vs. quantity: The ease of generating content with AI tools may lead to a flood of mediocre work
Skill devaluation: As AI tools become more sophisticated, there's fear that traditional artistic skills may become less valued
Ethical use: Questions about the responsible development and deployment of AI in creative fields
Transparency: Calls for greater openness from AI companies about their development processes and data usage
Artist collaboration: Debate on how AI companies should engage with and compensate artists during development
🤔 Final Thoughts
In light of the Sora leak and the ongoing debates surrounding AI-generated content, artists and content creators find themselves at a crossroads. While it's crucial to stay informed and experiment with emerging technologies like AI video generators, it's equally important to approach these opportunities with caution. Artists should consider exploring AI tools to enhance their creative processes and expand their skill sets, as these technologies are likely to play a significant role in the future of content creation. However, they must also be vigilant about protecting their rights and the value of their work. Engaging with AI companies or participating in early access programs should be approached with a critical eye, ensuring fair compensation and recognition for contributions.
"DEMOCRATIZE BILLION DOLLAR BRANDS EXPLOITING ARTISTS FOR UNPAID R&D AND PR" - OpenAI Sora leak.
AI runs on IP. It's time for a better solution.
— Jason Zhao (꧁IP꧂) (@jasonjzhao)
7:02 PM • Nov 26, 2024
#soraleak : Why are we doing this ?
😜😜😜
— Sandrine Decorde (@SandrineDecorde)
6:36 PM • Nov 26, 2024
~ JL