In this edition we’ll be covering…
Google's breakthrough Gemini 2.5 Computer Use model that navigates browsers like a human
The three-step workflow for transforming CSV files into publication-ready infographics
A tutorial on implementing Claude's new persistent memory tool in your API
5 trending AI signals from Anthropic, Atlassian, Google, and the courtroom
3 AI tools for brand monitoring, web fiction writing, and market research
And much more…
The Latest in AI
Google’s AI Can Now Use Your Browser Like a Human
Google DeepMind just dropped Gemini 2.5 Computer Use, and it's an AI that can actually navigate your browser, click buttons, and complete tasks just like you would.
Think of it as having a digital intern who can actually use the internet…
Available now via the Gemini API in both Google AI Studio and Vertex AI, this specialized model builds on Gemini 2.5 Pro's visual understanding and reasoning capabilities to power agents that interact with user interfaces directly.
I thought these other guys had this too???
Gemini 2.5 Computer Use ranks number one on multiple web and mobile control benchmarks, outperforming both OpenAI's Computer Agent and Anthropic's Claude 4.5 Sonnet in accuracy and latency.
The model can handle browser-based tasks like filling forms, manipulating dropdowns, clicking buttons, and operating behind logins – all the messy UI work that APIs can't touch.
Google's already using it in production for Project Mariner, the Firebase Testing Agent, and agentic capabilities in AI Mode Search. Early testers report 50% faster performance than competitors and up to 18% better accuracy on complex data parsing tasks.
So What?
Computer use capabilities represent the next evolution of AI agents – from systems that only interact through structured APIs to ones that navigate the messy, visual world of human interfaces. Google's betting that the future isn't about building more APIs, but about AI that can use the web as it already exists.
This is the clearest signal yet that we're moving toward AI agents that can handle entire workflows autonomously, not just answer questions or generate text. When Google's payments team reports that Computer Use rehabilitates over 60% of previously broken UI tests, we're looking at real productivity gains in ways most people haven't even considered yet.
Tool Spotlight
Turn CSV Chaos Into Magazine-Quality Infographics in Under 2 Minutes
Did you know you can turn 500 rows of raw sales data into a magazine-quality infographic in under 2 minutes (without Canva, Figma, or a design team)?
Just Google Gemini and a three-step prompt sequence most people don't know exists.
Here's the exact workflow:
Step 1: Upload a CSV to Gemini and run this prompt:
Analyze this CSV data and identify the key patterns, trends, and insights. What are the most important takeaways?
Gemini immediately surfaces your top performing regions, seasonal buying patterns, and high-value customer segments. Data you would've spent 30 minutes manually reviewing.
Step 2: Extract the metrics that matter:
Create summary statistics from this analysis. Give me the top 5 key metrics that would work well in an infographic.
Step 3: Generate the infographic with Nano Banana:
Create a professional infographic displaying these 5 key metrics: [paste stats]. Use a modern, clean design with data visualization elements.
Two minutes later, you have a presentation-ready visual that would've cost you 2 hours in Canva.
Proof that you can go from CSV to final asset before your morning coffee gets cold!
🧠 You're reading this newsletter because you know AI matters…
But reading about AI ≠ actually using AI.
Here's the gap: You consume AI content, but you're not building meaningful products or workflows with it yet.
Inside our learning platform, you’ll have access to:
100+ hands-on lessons (not just theory)
A complete roadmap to take you from 0 → 1 with AI
Prompts you'll actually use today
Step-by-step automations for your workflows
(PLUS) weekly updates as AI evolves
(PLUS) 20+ new lessons released each month
All created by industry veterans. 3-day free trial. Only $5.99/month after.
Innovation Showcase
How to Use Claude’s New Memory Tool In The API
Anthropic just released one of the most underrated features in the Claude API: persistent memory that actually works across conversations.
The memory tool lets Claude store and retrieve information through a client-side file directory that persists between sessions. Unlike keeping everything in the context window, Claude can build knowledge over time without burning through tokens.
Here's how to get started:
Include the beta header
context-management-2025-06-27
in your API requestsAdd the memory tool to your request
Implement client-side handlers for memory operations
To handle memory operations, you'll need to implement handlers for each memory command. Anthropic's SDKs provide memory tool helpers that handle the interface, you can subclass BetaAbstractMemoryTool
(Python) or use betaMemoryTool
(TypeScript) to implement your own memory backend (file-based, database, cloud storage, encrypted files, etc.).
Here's a basic example:
curl https://api.anthropic.com/v1/messages \
--header "x-api-key: $ANTHROPIC_API_KEY" \
--header "anthropic-version: 2023-06-01" \
--header "anthropic-beta: context-management-2025-06-27" \
--data '{
"model": "claude-sonnet-4-5",
"max_tokens": 2048,
"messages": [{
"role": "user",
"content": "Help me debug this Python function..."
}],
"tools": [{
"type": "memory_20250818",
"name": "memory"
}]
}'
The memory tool supports six commands your client needs to handle:
view - Shows directory contents or file contents with optional line ranges
create - Create or overwrite a file
str_replace - Replace text in a file
insert - Insert text at a specific line
delete - Delete a file or directory
rename - Rename or move a file/directory
Quick Bites
Stay updated with our favorite highlights, dive in for a full flavor of the coverage!
Anthropic and Deloitte expanded their alliance to make Claude available across Deloitte's global network and develop industry-specific AI solutions for regulated industries like financial services and healthcare.
Atlassian launched the Software Collection at Team '25 Europe, featuring Rovo Dev – an AI agent that handles code reviews, CLI workflows, and automated debugging across the entire software development lifecycle.
Google introduced CodeMender, an AI agent that automatically detects and patches security vulnerabilities in code, moving beyond simple static analysis to actually fix the problems it finds.
More people are using ChatGPT as legal counsel in court, with some self-representing litigants winning cases while others face sanctions for filing AI-hallucinated citations to non-existent cases.
Cisco unveiled the 8223 routing system powered by the new Silicon One P200 chip, designed to efficiently connect AI data centers across vast distances with Microsoft and Alibaba as launch customers.
Trending Tools
📊 PromptSignal - Track how leading AI models mention, rank, and describe your brand across ChatGPT, Claude, and Gemini to measure your visibility in AI-generated answers.
✍️ WebNovel AI - Structured planning and drafting tool for web fiction writers that helps you build worlds, design characters, and organize your story from concept to chapters.
🔍 Extruct AI - Automate company research and market intelligence with AI agents that scan and extract relevant data for B2B businesses in structured, easy-to-use formats.
Community Corner
Where we spotlight the practical AI skills our members are mastering right now!
We just launched our comprehensive Everything You Need to Know About Claude program – and our members are absolutely loving it.
This isn't another surface-level tutorial. We're talking 13 deep-dive lessons covering everything from advanced prompting and extended thinking to MCP integration and complete workflow automation.
Each lesson includes hands-on examples you can implement immediately. The feedback has been incredible – members are already seeing massive improvements in their Claude workflows.
Head over to our learning platform and level up your Claude game.
Until we Type Again…
Thank you for reading yet another edition of Digestibly!