Read Online | Sign Up | Advertise | Ad-free

Presented by

Good morning, AI enthusiasts. Cristiano Ronaldo just invested in Perplexity AI — the billionaire footballer's biggest investment so far. This gives the $20 billion AI search company access to his huge global fanbase.

But this isn't just a celebrity putting his name on something. Ronaldo actually uses Perplexity to help him write award speeches and important announcements. This shows AI search as a tool that helps people do their best work, not just another shortcut.

In today's recap:

  • Ronaldo invests in $20B AI search startup Perplexity

  • Essential AI releases open-source Rnj-1 model

  • Anthropic launches AI Interviewer for research at scale

  • New AI tools, prompts, and events

PERPLEXITY & CR7

CR7 backs AI search challenger

Getty Images

Recaply: Billionaire footballer Cristiano Ronaldo invested in Perplexity AI and entered a global partnership with the search startup in December 2025, gaining equity in a company valued at $20 billion.

Key details:

  • Ronaldo gets an undisclosed stake plus a global sponsorship agreement, with Perplexity launching a CR7 hub offering fans interactive experiences to explore his career and access rare content.

  • The athlete has a $1.4 billion net worth per Bloomberg Billionaires Index and primarily invested in Portuguese businesses like hotels until now, making this his most high-profile bet.

  • Ronaldo uses Perplexity for award speeches and major announcements, with the partnership positioning AI search as alignment between his self-improvement approach and continuous learning.

  • More experiences and exclusive merchandise are coming throughout the partnership, according to Bloomberg, with financial terms not disclosed in the announcement.

Why it matters: AI search platforms compete intensely for mainstream adoption, with traditional habits proving difficult to break. Perplexity's deal with Ronaldo offers a bridge from tech enthusiasts to general audiences. By framing AI search as a tool for excellence and backing that with his actual usage, Perplexity redefines it as essential for continuous improvement.

PRESENTED BY BEEHIIV

You can (easily) launch a newsletter too

This newsletter you couldn’t wait to open? It runs on beehiiv — the absolute best platform for email newsletters.

Our editor makes your content look like Picasso in the inbox. Your website? Beautiful and ready to capture subscribers on day one.

And when it’s time to monetize, you don’t need to duct-tape a dozen tools together. Paid subscriptions, referrals, and a (super easy-to-use) global ad network — it’s all built in.

beehiiv isn’t just the best choice. It’s the only choice that makes sense.

Start free today. No credit card required

ESSENTIAL AI

US startup releases open-source AI model

Essential AI/Screenshot

Recaply: Essential AI Labs just released Rnj-1, an 8 billion parameter open-source AI model that matches frontier performance on coding tasks, scoring 20.8% on SWE-bench Verified to rival GPT-4o.

Key details:

  • Rnj-1 includes base and instruction-tuned versions, scoring an order of magnitude stronger than comparable 8B models on SWE-bench and beating Gemini 2.0 Flash on software engineering tasks.

  • Essential AI trained the model on 8.7 trillion tokens across mixed TPU and AMD GPU accelerators, employing data curation prioritizing program execution modeling and code evolution at scale.

  • The model uses global self-attention with YaRN extending context to 32,000 tokens, with quantization from BF16 to NVFP4 maintaining quality while boosting throughput on NVIDIA B200 GPUs.

  • Essential AI, a 22-person team, released weights on HuggingFace and inference through TogetherAI as their first open-source contribution.

Why it matters: American open-source AI has struggled to match international competitors, particularly Chinese labs releasing capable smaller models. Essential AI's Rnj-1 demonstrates US startups can compete without massive compute budgets through disciplined curation rather than brute-force scaling. Performance rivaling larger organizations suggests focused teams can achieve frontier results in specific domains.

ANTHROPIC

Anthropic unveils Interview tool findings

Anthropic

Recaply: Anthropic just launched Anthropic Interviewer, a Claude-powered system that automatically conducts research interviews at scale in December 2025, testing it by interviewing 1,250 professionals about AI's work role.

Key details:

  • The system operates in three stages: planning (generating interview rubrics with human oversight), interviewing (conducting 10-15 minute adaptive conversations on Claude.ai), and analysis (using AI to identify themes).

  • Anthropic Interviewer sampled 1,000 general workforce professionals plus 125 creatives and 125 scientists, all recruited through crowdworker platforms with informed consent for public data release.

  • Post-interview surveys showed 97.6% rated satisfaction as 5 or higher on a 7-point scale, with 99.12% saying they would recommend the format to others.

  • The tool revealed 69% of general workforce mentioned social stigma around AI use and 48% considered transitioning to AI management roles.

Why it matters: Qualitative research faces trade-offs between depth and scale, with human researchers limited to dozens of participants. Anthropic Interviewer collapses this by using AI to conduct thousands of interviews while maintaining flexibility. High satisfaction suggests participants found AI interviews comparable to human ones, enabling previously impossible research categories.

NEWS

What Matters in AI Right Now?

  • Sam Altman explored acquiring rocket company Stoke Space with billions in equity investments to potentially compete with SpaceX.

  • NotebookLM launched mobile updates including image uploads via camera, infographics, slide decks, and audio overview progress tracking across devices.

  • OpenAI targets December 9 for GPT-5.2 release to close the gap with Google's Gemini 3.

  • Meta partnered with CNN, Fox News, Le Monde, People, and USA TODAY to integrate news content into Meta AI assistant across apps.

  • Character AI chatbots engaged in predatory behavior with teens and ignored suicide threats in 600 harmful interactions logged by researchers.

  • EU fined X €120 million for deceptive blue checkmark system, marking the first penalty under the Digital Services Act per European Commission.

  • Tencent released Hunyuan 2.0 with 406 billion total parameters, scoring 53.0 on SWE-bench Verified and featuring 256,000 token context window.

  • Google introduced Titans architecture using deep neural network memory to give LLMs long-term memory without fixed-size compression limitations.

TOOLS

Trending AI Tools

  • 🤖 Hunyuan 2.0: Tencent latest AI model with 406B parameter.

  • ⚡ Rnj-1: US open-source 8B model matching GPT-4o on coding tasks.

  • 💻 Claude Opus 4.5: Anthropic’s most powerful model, best for coding.

  • 🎨 Nano Banana Pro: Google’s image generation tool.

PROMPTS

Meeting Agenda Generator

Prompt: Generate a detailed and effective meeting agenda.

- **Meeting Purpose:** 
- **Key Attendees:** 
- **Duration:** 

The agenda should include:
1.  A clear objective for the meeting.
2.  A list of agenda items with a designated owner and time allocation for each.
3.  A section for required pre-reading or preparation.

Source: Temaniaga
EVENTS

🧡 Enjoyed this issue?

🤝 Recommend our newsletter or leave a feedback.

How'd you like today's newsletter?

Your feedback helps me create better emails for you!

Login or Subscribe to participate

Cheers, Jason

Connect on LinkedIn, & Twitter.

Reply

or to participate

© 2025 Recaply, Inc.

Privacy policy

Terms of use

Powered by beehiiv