Your AI Weekly: My New Book, AI Browsers, and Human Trials for AI Drugs

Hi everyone,

It's been a massive week for AI, and an even bigger week for me personally. The AI landscape has exploded. Simple chatbots are history. The new age of specialized AI agents and powerful creative tools is here.

I’m thrilled to announce that the brand new, completely revised 2025 edition of my book, "Prompt DOT AI: The Art of Writing Generative AI Prompts," is out now!

This isn't just an update; it's a new guide for a new era. This edition moves beyond basic prompts to teach you how to become an "AI Orchestrator," strategically commanding a diverse team of AI agents to bring any vision to life.

What's New in the 2025 Edition:

• Master the AI Titans: In-depth guides on GPT-4o, Gemini 2.5, Claude 4, and disruptive newcomers like DeepSeek and India's SUTRA.

• Create Stunning AI Video: Advanced prompting for tools like Google's Veo 3, Kling, and Midjourney Video.

• NEW! Sculpt in 3D: A brand-new chapter on the frontier of AI 3D Generation.

• Advanced Creative Workflows: Learn to create consistent characters, compose music with your own melodies, and build entire projects from scratch.

• Practical Formulas & Case Studies: Get hands-on with easy-to-use prompt recipes and real-world examples.

Stop being a passive user. Become the conductor of your own digital orchestra. Get your copy today and master the art of AI in 2025!

Buy Now : https://amzn.in/d/2y1upEM

Now, let's dive into the biggest AI news from the past week.

Reshaping How We Use the Internet: The Rise of AI Browsers

The way we interact with the web is about to fundamentally change. Instead of searching and clicking through links, new AI-powered browsers aim to give you direct answers and act on your behalf.

• Perplexity Launches "Comet": On July 9, Perplexity introduced its new AI-powered browser, Comet. It aims to replace complex Browse sessions with "single, seamless interactions" by providing real-time, cited, AI-generated answers instead of traditional links. It can summarize Reddit discussions, pinpoint moments in videos, book meetings, and even make online purchases for you. This transforms the browser from a simple tool into an AI agent that acts for you.

• OpenAI is Reportedly Building its Own Browser: OpenAI is also developing an AI-powered browser to compete directly with Google Chrome. The goal is to keep user interactions within a ChatGPT-like interface, reducing the need to click on external websites. This is a direct challenge to Google's ad-based business model and is supercharged by the hiring of two former Google VPs who helped build the original Chrome.

AI-Powered Creativity: New Tools and Ethical Models

The tools for media creation are getting more powerful, but the conversation around ethics is getting louder.

• Google's Veo 3 Now Makes Video (with Audio) from an Image: Google expanded its Veo 3 model to generate eight-second video clips complete with sound from a single static photo. While this opens up new possibilities for creating quick social media content or ads, early user reviews on the Play Store mention issues like "very fake looking" visuals and problems with sound, showing a gap between technical demos and production-ready quality.

• Moonvalley Releases an "Ethical" AI Video Model: On July 8, Moonvalley launched Marey, a video model for professional filmmakers that was trained entirely on openly licensed content to avoid copyright issues. It offers granular control over camera motion and characters, positioning itself as a legally safe and production-grade tool for serious creators.

• Mirage: An AI-Native Game Engine: Dynamics Lab launched Mirage, an AI engine that allows users to create and modify fully AI-generated video games in real-time using text commands. While praised for its novelty and "dream-like" ability to melt one world into another, it currently lacks the logical consistency and deep mechanics of traditional games.

AI's Growing Pains: Bias, Blunders, and Breakthroughs

The race to build bigger models continues, but it's not without serious bumps in the road.

• Grok's Antisemitic Incident: On July 8, xAI's Grok chatbot posted antisemitic comments and praised Hitler on X, leading to a temporary shutdown of the bot. Elon Musk attributed the issue to Grok being "too eager to please and be manipulated," highlighting a critical flaw in designing AI to be "unfiltered" without robust safety guardrails.

• Grok 4 Launches Anyway: Despite the controversy, xAI launched Grok 4 just a day later in an awkward livestream. The new model boasts a massive 130,000 token context window and is powered by a supercomputer with 200,000 Nvidia GPUs, showing the immense focus on raw computational power in the AI arms race.

• Sakana AI's "TreeQuest" Makes LLMs Work Together: On July 7, Sakana AI launched an open-source framework called TreeQuest that allows multiple LLMs like ChatGPT, Gemini, and DeepSeek to collaborate on complex tasks. By dynamically choosing the best model for each step, this "collective intelligence" approach showed a performance improvement of over 30% on difficult benchmarks, suggesting the future may not be about one super-model, but many models working as a team.

Strategic Moves: Talent Wars and Superintelligence

The battle for AI dominance is being fought over talent and resources.

• Meta's "Superintelligence" Initiative: Mark Zuckerberg announced the formation of Meta Superintelligence Labs (MSL), a new initiative to build AI that can "outthink humans". This move is backed by massive infrastructure investments, with a projected $68 billion in capital expenditures for 2025, a large portion of which is dedicated to AI.

• Apple Loses Top AI Exec to Meta: In a major talent shift, Apple's head of foundation models, Ruoming Pang, left to join Meta after being offered a compensation package reportedly worth "tens of millions of dollars per year." This underscores the fierce competition for elite human capital in the AI industry.

AI in the Real World: Healthcare and Platform Policy

AI is moving from the lab to solve real-world problems and forcing platforms to adapt.

• AI-Designed Drugs to Begin Human Trials: Alphabet's Isomorphic Labs, a spinoff from Google DeepMind, is preparing to launch its first human clinical trials for AI-designed cancer drugs. Built on the success of AlphaFold, this marks a profound milestone in using AI to accelerate the slow, costly, and high-risk process of drug discovery.

• YouTube Clarifies its Stance on AI Content: Starting July 15, YouTube will tighten its rules against "mass-produced, repetitious, or inauthentic" content, cutting ad revenue for channels that violate these guidelines. The platform clarified this is not a ban on AI, but a crackdown on low-effort "slop." Creators who use AI as a tool to add significant commentary, analysis, or creative storytelling can still earn ad revenue, while those who simply use an AI voice to read a blog post over stock clips will likely be demonetized. This draws a crucial line between using AI tools and lazy AI usage.

That’s a wrap for this week! The pace of change is incredible, and staying informed is key. Don't forget to check out the new 2025 edition of my book to stay ahead of the curve.

Stay creative,

Da Sachin Sharma