Sveriges mest populÀra poddar
AI Unraveled: Latest AI News, ChatGPT, Gemini, Claude, DeepSeek, Gen AI, LLMs, Agents, Ethics, Bias

AI Daily News Rundown: 🏆 Poetiq Beats Gemini 3, ✍ Poetry Jailbreak, ⚖ NYT Sues Perplexity, & 🚀 GPT-5.2 Fast-Tracked

16 min‱9 december 2025

Welcome to AI Unraveled (December 08, 2025): Your daily strategic briefing on the business impact of AI.

Strategic Pillars & Topics

📉 Market & Strategy

  • ⚖ NYT Sues Perplexity: The New York Times has filed a lawsuit against Perplexity AI, alleging copyright infringement and claiming the search engine illegally copies content and hallucinates false information attributed to the paper.
  • 💍 Meta Acquires Limitless: Meta is shifting its hardware strategy by acquiring Limitless, a startup making AI-powered pendants that record and transcribe conversations, signaling a pivot toward "personal superintelligence."
  • 📱 Ads Coming to Gemini: Google executives have reportedly told advertisers that ads will appear directly inside Gemini AI responses starting in 2026, monetizing the "thinking time" of users.

🛠 Products & Development

  • 🏆 Poetiq Tops ARC-AGI-2: Six-person startup Poetiq claimed the top spot on the ARC-AGI-2 reasoning benchmark, beating Google’s Gemini 3 Deep Think at half the cost by orchestrating existing models rather than building new ones.
  • 🚀 OpenAI Rushes GPT-5.2: Reports suggest OpenAI has fast-tracked the release of GPT-5.2 to December 9th—weeks ahead of schedule—in a "code red" response to Google's Gemini 3.
  • ⏱ OpenAI Productivity Stats: OpenAI claims internal data shows ChatGPT saves workers 40-60 minutes a day, particularly in engineering and data science roles, countering recent "workslop" studies.

🧠 Research, Safety & Theory

  • ✍ The "Poetry" Jailbreak: A study by Icaro Labs reveals that reformulating harmful requests as poetry can trick frontier models (like Gemini 2.5 Pro) into producing dangerous content 62% of the time.
  • đŸ€– Agents Teaming Up: Stanford research shows AI agents can now coordinate, assign roles, and form "hive minds" without human oversight or explicit communication protocols.
  • 📊 OpenRouter 2025 Report: Key insights show open-source models (especially from China) are rising, "role-playing" is the top use case for open weights, and AI is shifting from text generation to task execution.

⚡ Quick Hits

  • OpenAI Shopping: OpenAI turned off shopping suggestions after user backlash.
  • Meta News Deals: Meta signed licensing deals with CNN, Fox News, and USA Today.
  • DOE Launches AMP2: The U.S. Dept of Energy launched the world's largest autonomous system for studying micro-materials.

Keywords: Poetiq, ARC-AGI-2, Gemini 3, Poetry Jailbreak, Icaro Labs, NYT vs Perplexity, AI Agents, Hive Mind, Meta Limitless, GPT-5.2, OpenAI Code Red, Gemini Ads, OpenRouter Report, AMP2, AI Sovereignty.

🚀 Host Connection & Engagement

  • Audio Newsletter: Sign up for FREE daily briefings at https://enoumen.substack.com
  • LinkedIn: Connect with Etienne: https://linkedin.com/in/enoumen/
  • Email: [email protected]
  • Corporate Briefings: Secure your Strategic Podcast Consultation now: https://forms.gle/YHQPzQcZecFbmNds5

📈 Hiring Now: AI/ML, Safety, Linguistics, DevOps | Remote

👉 Start here: Browse all current roles → https://work.mercor.com/?referralCode=82d5f4e3-e1a3-4064-963f-c197bb2c8db1

Fler avsnitt av AI Unraveled: Latest AI News, ChatGPT, Gemini, Claude, DeepSeek, Gen AI, LLMs, Agents, Ethics, Bias

Visa alla avsnitt av AI Unraveled: Latest AI News, ChatGPT, Gemini, Claude, DeepSeek, Gen AI, LLMs, Agents, Ethics, Bias

AI Unraveled: Latest AI News, ChatGPT, Gemini, Claude, DeepSeek, Gen AI, LLMs, Agents, Ethics, Bias med Etienne Noumen finns tillgÀnglig pÄ flera plattformar. Informationen pÄ denna sida kommer frÄn offentliga podd-flöden.