- Crawl-Walk-Run
- Posts
- Google Makes Some Moves with AI
Google Makes Some Moves with AI
A link for a fresh way to generate images, The Future of AI, Waiting for the next GPT (plus an AI comic)

Welcome to Crawl Walk Run AI, curated by Brad Reynolds and Brian Diehl. We hope you're enjoying this weekly newsletter focused on keeping you informed about the AI world. Let us know if there are any topics you'd like to read about!

GPT-5 project is over budget and months behind schedule.
New GPT-5 model may not justify its high development costs.
AI advancements could enhance customer service and digital marketing.
Enhanced LLMs streamline operations like contract analysis and risk assessment.
Google Gemini Expands
TechCrunch
Google expands Gemini's research mode to 40 languages.
AI tool creates comprehensive reports by researching and refining information.
Accurate language summaries are a challenge, especially in languages like Hindi.
Native language data quality checks ensure training data accuracy and reliability.

How AI Works - Another AI Primer
The Royal Institute
A 1-hour lecture by Oxford Uni professor Michael Wooldridge on how machine learning works.
Machine learning, key in AI, gained traction with practical applications post-2005.
Neural networks replicate patterns found in the brain using digital simulations.
LLMs highlight AI's efficiency in understanding and generating text.
Gemini now lets subscribers ask detailed questions about PDFs on their phone.
Integrated with Files by Google app for enhanced document interaction.
Context-aware functionality is available for web pages and YouTube videos.
Screen-aware assistance gives answers using screenshots when context isn't supported.

A New Leap Forward
By AI Researchers: Azeem Azhar and Nathan Warren
AI leaders are shifting focus from size to strategy for innovation.
Current models show unpredictable outcomes; no longer solely results-driven.
Companies now boost computing power during model use, improving accuracy.
The pivot to AGI by 2025-2027 is reshaping the AI future landscape.
Visual Creativity Simplified
Tech Radar
Google Whisk allows creation of AI visuals using image prompts, not text.
The tool is built on Google's Imagen 3 generative AI model.
It's free to use for users in the US, enhancing accessibility.
You can try it out via your web browser at: https://labs.google/whisk.
Reply