Netflix’s AI-powered search rolls out in test markets

PLUS: Gemini 2.5 Pro hits #1, Llama-Nemotron models, WisdomAI’s $23M round & autonomous agents for game dev

In today’s agenda:

1️⃣ Netflix unveils AI-powered search assistant for natural language queries

2️⃣ Gemini 2.5 Pro I/O Edition tops WebDev Arena leaderboard

3️⃣ NVIDIA releases Llama-Nemotron reasoning models on Hugging Face

Plus, some interesting news:

  • WisdomAI raises $23M to reduce hallucinations in enterprise LLMs

  • Game, Sett secures $27M to build AI agents for game development

MAIN AI UPDATES / 7th May 2025

🎬 Netflix unveils AI-powered search assistant 🎬
Smarter recommendations powered by natural language

Netflix has launched a new AI-powered search assistant capable of understanding natural language queries like “movies with a strong female lead and a twist ending.”
Unlike basic keyword search, the tool is powered by generative models trained on metadata, viewing behavior, and critical reception.
It’s currently rolling out in select markets before a broader release.

💻 Gemini 2.5 Pro I/O Edition takes #1 spot 💻
Coding capabilities reach new heights

Google’s latest Gemini 2.5 Pro update (05-06) has claimed the #1 position on the WebDev Arena leaderboard with a top score of 1420, improving by +147 Elo over its predecessor.
The I/O Edition excels in web development, UI generation, function calling, and agentic workflows.
It also scores 84.8% on the VideoMME benchmark and can generate mathematically correct fractal visualizations.

🧠 NVIDIA's Llama-Nemotron debuts 🧠
Efficient reasoning models with enterprise license

NVIDIA has released Llama-Nemotron, a new family of reasoning-optimized models on Hugging Face.
They are designed for enterprise applications, combining strong inference performance with open licensing.
The models perform well on complex benchmarks and are ready for production environments, integrating smoothly with deployment frameworks.

INTERESTING NEWS

🧠 WisdomAI raises $23M to combat LLM hallucinations 🧠

WisdomAI, a London-based startup, has secured $23 million in Series A funding to tackle hallucinations in large language models.
Their platform integrates verifiable, structured data into LLM pipelines, aiming to increase accuracy and trust in sensitive sectors like finance and healthcare.

🎮 Game, Sett emerges from stealth 🎮

Game, Sett, a stealth-mode startup, has raised $27 million to build AI agents capable of autonomously generating game assets, mechanics, and narratives.
The company aims to drastically reduce the time-to-market in game development while supporting creative iteration and team productivity.

📩 Have questions or feedback? Just reply to this email , we’d love to hear from you!

🔗 Stay connected: