Beginners in AI

Good morning and thank you for joining us again!

Welcome to this daily edition of Beginners in AI, where we explore the latest trends, tools, and news in the world of AI and the tech that surrounds it. Like all editions, this is human curated, and published with the intention of making AI news and technology more accessible to everyone.

THE FRONT PAGE

Scientists Prove We're Not Living in a Simulation

TLDR: Researchers at the University of British Columbia used advanced mathematics to prove the universe can't be a simulation—and the reason why reveals fundamental limits to what algorithms can ever accomplish.

The Story:

Scientists just published mathematical proof that we're not living in a simulation. The logic is surprisingly straightforward: any simulation, no matter how sophisticated, would need to run on algorithms. But researchers demonstrated that certain fundamental aspects of reality are "computationally undecidable"—meaning they can't be described or predicted by algorithms at all. The team used Gödel's incompleteness theorems, Tarski's undefinability theorem, and Chaitin's information-theoretic incompleteness to show these mathematical boundaries exist. Since you can't simulate something that's inherently non-algorithmic, the universe must be real. "A fully consistent and complete description of reality cannot be achieved through computation alone," lead researcher Mir Faizal explained. The findings, published in the Journal of Holography Applications in Physics, effectively close a debate that's ranged from Plato's Cave to The Matrix.

Its Significance:

The mathematicians revealed hard limits to computation. If certain aspects of reality are permanently beyond algorithmic reach, that constrains what any AI system can ultimately achieve, regardless of how advanced it becomes. 

The research connects to a classic computer science problem: you can't write an algorithm that predicts whether every program will finish running or loop forever. If certain aspects of reality work the same way—impossible to predict algorithmically—then no simulation can capture them, or us.

QUICK TAKES

Visa Prepares AI Commerce Infrastructure for Asia Pacific 2026 Pilots

The story: Visa announced its Intelligent Commerce platform for Asia Pacific with AI commerce pilots starting early 2026. The system uses a Trusted Agent Protocol to verify AI shopping assistants and let them make purchases on behalf of consumers. Visa's infrastructure handles the explosion of AI-driven traffic to retail sites—which jumped 4,700% in one year—by distinguishing legitimate AI shoppers from malicious bots.

Your takeaway: This marks the first major payment network building infrastructure specifically for AI agents to shop, setting the stage for a future where your AI assistant handles purchases while merchants still know who the actual customer is.

Google NotebookLM Adds Deep Research and Support for More File Types

The story: Google updated NotebookLM with Deep Research, a feature that acts like a dedicated researcher to browse hundreds of websites and create detailed reports with sources. The update also added support for Google Sheets, Word documents, PDFs from Drive, images, and Drive file URLs. Users can add Deep Research reports and sources directly into notebooks and create audio or video overviews from them.

Your takeaway: NotebookLM is evolving from a simple note-taking tool into a comprehensive research assistant that can gather information, synthesize it, and present it in multiple formats—making complex research projects much faster.

Nvidia Plans to Sell Complete AI Servers Instead of Just GPUs

The story: JPMorgan reports that Nvidia will start shipping fully assembled AI server components with its upcoming Vera Rubin platform instead of just selling individual GPUs. The move means Nvidia will pre-install Rubin GPUs, cooling systems, and power hardware, leaving partners to only handle rack integration. This vertical integration strategy should boost Nvidia's profits while making deployment faster for customers.

Your takeaway: Nvidia is taking over more of the AI hardware supply chain, which means higher profit margins for them but less design work and potentially lower margins for the hardware manufacturers who used to build these systems.

OpenAI Builds Experimental Model to Understand How AI Really Works

The story: OpenAI created an experimental AI model called a weight-sparse transformer that's far easier to understand than regular AI models. Unlike typical AI systems where neurons connect to everything and spread knowledge in confusing ways, this model uses limited connections that make it possible to trace exactly how it completes tasks. The model is as capable as GPT-1 from 2018, but researchers can watch it work step-by-step.

Your takeaway: This breakthrough in AI transparency could eventually lead to fully understandable AI systems, helping researchers figure out why models hallucinate and how much we should trust them with critical decisions.

AI Traffic Converts 3X Better Than Traditional Web Traffic

The story: Microsoft Clarity analyzed over 1,200 publisher and news websites and found that visitors coming from ChatGPT, Copilot, and Perplexity convert to sign-ups at 1.66% compared to just 0.15% from search, 0.13% from direct traffic, and 0.46% from social media. AI referral traffic grew 155.6% over eight months, though it still accounts for less than 1% of total website traffic.

Your takeaway: People who click through from AI assistants are much more likely to convert because they're arriving with higher intent—the AI has already filtered and contextualized information for them before sending them to your site.

TOOLS ON OUR RADAR

• 🔨 Super [Free]: AI-powered search that connects all your company tools—Slack, Google Drive, Notion, Confluence—into one searchable hub with instant answers and custom assistants.

• 📐 InVideo AI [Free]: Turn text prompts into complete videos with AI-generated scripts, voiceovers, stock footage, and editing—no video skills required.

• 🔧 Airtop [Free]: Automate web browsing tasks using natural language commands—your AI agents can log in, scrape data, and complete complex workflows across any website.

• 🪛 AI World Clocks [Open Source & Free]: Watch nine different AI models compete to design analog clocks every minute—a creative showcase of how different AIs interpret the same prompt.

TRENDING

Parallel Web Systems Raises $100M for AI Search Infrastructure – Former Twitter CEO Parag Agrawal's startup raised $100 million at a $740 million valuation to build APIs that let AI agents search the live web for current information instead of relying on outdated training data.

Russia's First Humanoid Robot Faceplants During Moscow Debut – Russia's first AI-powered humanoid robot named AIdol fell flat on its face moments after stepping onto a Moscow stage during its Wednesday debut, cutting the showcase dramatically short.

AI Chatbots Can Help Cut Your Grocery Bill, With Limitations – News4 tested ChatGPT, Copilot, and Gemini for budget shopping lists and found ChatGPT gave the most detailed plan with real prices from Walmart, though none included extras like snacks or paper products.

New Prediction Method Achieves Startlingly Close Match to Reality – Mathematicians created a forecasting technique that maximizes how well predictions align with actual values rather than just reducing errors, showing strong results in medical and body measurement tests.

Researchers Hide AI Prompts in Papers to Game Peer Reviews – At least 17 research papers on arXiv contained hidden white text telling AI reviewers to give positive feedback, a response to the growing use of AI tools by lazy peer reviewers.

TRY THIS PROMPT (copy and paste into ChatGPT, Grok, Perplexity, Gemini)

I need help identifying automation opportunities in my work routine.

First, I'll list my repetitive weekly tasks:
[List 5-10 tasks you do every week that feel repetitive - examples: "Write weekly status emails to my team", "Format sales data into reports", "Schedule social media posts", "Respond to common customer questions", "Create meeting agendas"]

For each task I listed, please:
1. Tell me if an AI tool can automate or speed up this task (yes/no)
2. If yes, recommend the specific tool to use (free options preferred)
3. Provide exact steps to set it up (numbered, beginner-friendly)
4. Estimate time saved per week
5. Rate the difficulty of setup (Easy/Medium/Hard)

Format your response as a table with these columns: Task | Can Automate? | Tool | Setup Steps | Time Saved | Difficulty

Then, at the end, calculate my total potential time savings per week and suggest which task I should automate first based on highest time savings with easiest setup.

What this does: Analyzes your actual workflow to find concrete automation wins you can implement this week, prioritized by effort vs. reward.

WHERE WE STAND(Based on today’s Snapshot and Trending News)

✅ AI Can Now: Search the live web in real-time to power agents that write code, analyze sales data, and assess insurance risks with current information instead of relying on outdated training data.

❌ Still Can't: Avoid being manipulated by hidden prompts that researchers are embedding in papers to trick AI reviewers into giving positive feedback.

✅ AI Can Now: Create shopping lists with detailed prices and store links, potentially saving families money on groceries when used strategically.

❌ Still Can't: Think comprehensively about real shopping needs—missing basics like snacks, paper towels, and household essentials that humans remember.

✅ AI Can Now: Generate complete videos from text prompts with scripts, voiceovers, and stock footage, making professional video creation accessible to anyone.

❌ Still Can't: Fully explain how it reaches conclusions, with even experimental transparent models only matching 2018-era GPT-1 capabilities while researchers work toward interpretability.

FROM THE WEB

Hard to ignore this one. OpenAI's Sora is taking mundane doorbell camera footage—a cat sitting at someone's front door—and transforming it into an absurdist comedy sketch where the cat plays musical instruments and an angry woman comes out to pull the instruments away. The AI preserves the original video's structure (cat waits, person appears) while adding creative elements: guitars, violins, and realistic physics showing how a cat would actually grip each instrument and even react when it's pulled away.

Even how the keyboard itself gets knocked away and the cat jumps to avoid it. These videos have come a very long way in a very short time.

Click that link or this one to remix the video with your own twist.

RECOMMENDED LISTENING/READING/WATCHING

This book will absolutely wrecked me (in the best way). It's told from the perspective of Klara, an AI companion sitting in a store window, observing the world and waiting to be chosen by a child. Ishiguro uses this AI narrator to explore what it means to be human, to love, to hope. It's quiet and beautiful and will make you think about consciousness in ways you haven't before. Same author as The Remains of the Day—another acclaimed heartbreaker, though not sci-fi.

Thank you for reading. We’re all beginners in something. With that in mind, your questions and feedback are always welcome and I read every single email!

-James

By the way, this is the link if you liked the content and want to share with a friend.