🤖Robots That Cook & Perform Debut at WAIC 2025

Welcome to The AI Entrepreneurs

In this Issue: Google reinvents Search, China pushes for global AI rules, Claude limits usage, and AI diagnoses Parkinson’s from earwax. Plus, Higgsfield turns product pics into brand videos—instantly.

GOOGLE

Google’s AI Mode Reinvents Search in the UK

Google is rolling out AI Mode in the UK, supercharging Search with deep reasoning, multimodal input, and dynamic follow-ups all powered by a customized Gemini 2.5 model. Whether you're planning foodie adventures or decoding bird migration, AI Mode handles your long, complex queries with web-linked answers and richer context. It’s now live in the Google app and as a dedicated tab on Search results.

AI Mode breaks questions into subtopics using Google’s “query fan-out” technique and accepts queries by voice, text, or image making it a versatile tool for exploring nuanced topics or discovering hidden content across the web. Google promises high-quality results and better visibility for diverse websites, all while improving factual reliability through new ranking techniques.

ELEVENLABS

ElevenLabs CLI Powers Conversational Agents as Code

ElevenLabs has introduced its ConvAI CLI, redefining conversational AI deployment with a “code-first” approach. Developers can version-control agents alongside app code, configure them by environment (dev, staging, prod), and instantly sync updates via terminal. With pre-built templates, voice/text options, and webhook integrations, it’s ideal for building support bots, sales agents, and docs-driven assistants no manual UI tweaking required.

Key features include secure authentication, real-time tool linking, embedded HTML widget generation, and native CI/CD integration. Whether you're designing multi-agent ecosystems or simply automating onboarding with personalized voice assistants, ConvAI CLI makes scalable, programmable conversational AI effortless.

CHINA

China Proposes AI Alliance: Seeks Global Governance Framework at WAIC 2025

At the World Artificial Intelligence Conference in Shanghai, China unveiled its vision for a globally coordinated AI governance framework just days after the US released its own plan focused on domestic dominance.

Key Highlights:

• Premier Li Qiang called for an international AI governance alliance, citing regulatory fragmentation and growing risks.

• Urged against monopolistic controls: “AI must not become an exclusive game for a few nations or companies.”

• US plan contrasts sharply: focuses on streamlining regulation and securing strategic advantage.

• Tech rivalry intensifies with AI chip negotiations: Nvidia ban lifted, DuPont probe paused.

• China has over 5,000 AI firms and leads in generative AI patents.

• DeepSeek’s and Moonshot’s models shook Wall Street with low-cost, high-performance breakthroughs.

• AI investments: $209B via state VC since 2013, plus $56B in 2025 alone.

• ASEAN and Eric Schmidt push for US-China collaboration to counter misinformation and cyber threats.

Why It Matters:

WAIC 2025 makes clear: AI is more than tech it’s a geopolitical battleground and a call for shared responsibility. China’s proposal underscores the need for transparent oversight, technical parity, and global consensus before AI systems reshape economies and societies without a map.

ANTHROPIC

Claude Tightens the Screws: Weekly Limits Kick In August 28

Anthropic is implementing stricter weekly usage limits for Claude starting August 28, 2024. In addition to the rolling 5-hour window, subscribers will now face weekly caps across all models and a separate quota for Claude Opus 4. Users leveraging Claude Sonnet 4 can expect between 40 and 80 hours weekly, with those running multiple Claude Code instances reaching the ceiling sooner.

These changes aim to curb account sharing, minimize abuse, and maintain system performance, though Anthropic says fewer than 5% of users will feel the pinch. The move signals a shift toward more structured agent workflows, possibly paving the way for monetization models around Claude's code, Sonnet, and Opus tiers.

NOTEBOOKLM

NotebookLM Teases “Summer Blockbuster” AI Drop

On July 28, NotebookLM ignited speculation across X with a cryptic teaser:

No further details yet, but the dramatic phrasing has AI watchers anticipating a major feature release possibly agentic workflows, upgraded context tools, or multi-modal integration. With NotebookLM’s recent focus on expert-mode and precision prompts, this could mark a turning point in intelligent workspace design.

AI HEALTH

AI is reshaping diagnostics - one swab, scan, and pixel at a time

In this issue of AIHealthTech Insider, we dive into scent-based AI that detects Parkinson’s through earwax, retinal scans that replace blood tests, and AI that predicts how aggressive your tumor is before it spreads.

• Earwax AI screens for Parkinson’s with 94% accuracy

• Mediwhale replaces invasive tests with retinal scans

• Stemness AI forecasts tumor behavior across 11 cancers

• EnAcuity adds hyperspectral vision to surgeries no new hardware needed

• Everlab raises $10M to bring real-time diagnostics to preventive care

The future of healthcare isn’t coming. It’s already inside the clinic.

Stay Ahead: Subscribe for free to explore health tech breakthroughs weekly.

AI CREATIVITY

Upload a Product Photo, Get a Scroll-Stopping Ad Back

With Higgsfield AI, you don’t need gear, crew, or editing skills.

Just:

1. Pick your video style (Zen Garden? Glam Shot? Retro Ad?)

2. Upload your product image

3. Click “Generate”

Boom. You’ve got a scroll-stopping brand video in seconds.

From luxury shots to nature vibes all powered by AI.
Try it now → higgsfield.ai

Pro tip: Neutral background = best results.