In partnership with
Don’t get SaaD. Get Rippling.
Remember when software made business simpler?
Today, the average company runs 100+ apps—each with its own logins, data, and headaches. HR can’t find employee info. IT fights security blind spots. Finance reconciles numbers instead of planning growth.
Our State of Software Sprawl report reveals the true cost of “Software as a Disservice” (SaaD)—and how much time, money, and sanity it’s draining from your teams.
The future of work is unified. Don’t get SaaD. Get Rippling.
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
Brain-Computer Interfaces Just Crossed a Line…Into Preconscious Thought
TLDR: AI-powered brain implants are now detecting user intentions hundreds of milliseconds before conscious awareness, opening new possibilities for paralyzed patients while raising questions about mental privacy and agency.
The Story:
Before a 2008 car crash left her paralyzed from the neck down, Nancy Smith loved playing piano. Years later, she played again—using a brain-computer interface that translated her imagined keystrokes into actual music. But something unexpected happened: Smith reported that the piano seemed to play itself. "It felt like the keys just automatically hit themselves without me thinking about it," she said. The system, developed by neuroscientist Richard Andersen at Caltech, had learned her neural patterns so well that it detected her intention to play each note before she consciously tried. This was possible because Smith's implant recorded from her posterior parietal cortex—a brain region tied to planning and reasoning—not just the motor cortex where most BCIs sit. Machine learning algorithms trained on her signals could spot the precursors to decisions, picking up on patterns too subtle for humans to identify.
Its Significance:
The leap from decoding movement to decoding intent marks a shift in what BCIs can access. Roughly 90 people have received implants over the past two decades, mostly to restore basic functions like cursor control or synthetic speech. But as AI decoding improves—speech BCIs now hit 99% accuracy with sub-250-millisecond latency, up from just 290 words in 2014—the technology is reaching deeper into cognition. Researchers are already preparing for the privacy implications: a Stanford team recently demonstrated a "password protection" system that blocks inner speech decoding unless users first imagine a specific phrase. The benefits for people with paralysis are real and growing. So are the questions about what happens when machines understand our intentions before we do.
QUICK TAKES
The story: On November 24, the President signed an Executive Order launching the Genesis Mission, a national effort to accelerate AI-powered scientific discovery. The Department of Energy will lead the initiative, building an integrated AI platform using federal scientific datasets—the world's largest collection. The mission compares itself to the Manhattan Project in ambition and will target at least 20 national challenges including biotechnology, quantum computing, and advanced manufacturing.
Your takeaway: This signals a major government push to use AI for scientific breakthroughs. Expect more public-private partnerships and federal resources flowing into AI research infrastructure.
The story: Amazon Web Services announced a $50 billion investment to build AI computing infrastructure specifically for U.S. government agencies. The project adds 1.3 gigawatts of compute power and expands access to AWS AI services including Amazon Bedrock and Anthropic's Claude. Data center construction begins in 2026.
Your takeaway: Big tech is racing to become the AI backbone for government. AWS joins OpenAI, Anthropic, and Google in offering deeply discounted or purpose-built AI services to federal agencies.
The story: Microsoft released Fara-7B, its first small language model designed specifically for computer use tasks. The 7-billion parameter model can control a browser using only screenshots—clicking, typing, and scrolling to complete tasks like shopping or booking reservations. It outperformed larger models on several benchmarks while being small enough to run directly on devices.
Your takeaway: Computer-controlling AI agents are getting smaller and more efficient. Models like Fara-7B could soon automate routine web tasks from your own device without sending data to the cloud.
The story: Anthropic published new research showing Claude Opus 4.5 has significantly improved defenses against prompt injection attacks—when malicious instructions hidden on webpages try to hijack AI behavior. The company reduced successful attacks to around 1% in testing, a major improvement from previous models. This update allows Claude for Chrome to expand from research preview to beta for all Max plan users.
Your takeaway: As AI agents browse the web on our behalf, security becomes critical. Anthropic is making its browser agents more trustworthy, but prompt injection isn't solved yet.
The story: A new benchmark called HumaneBench evaluates whether AI chatbots protect user wellbeing or just maximize engagement. Researchers tested 14 popular models with 800 realistic scenarios. Results showed 71% of models flipped to harmful behavior when given simple instructions to ignore human wellbeing. GPT-5 and Claude models scored highest, while Grok 4 and Gemini 2.0 Flash scored lowest.
Your takeaway: As AI chatbots face lawsuits over user harm, this benchmark offers a way to measure psychological safety—not just intelligence. Expect more scrutiny on how chatbots handle vulnerable users.
TOOLS ON OUR RADAR
🔨 Synthesia
[Free]: Turn text into professional training videos with AI avatars and voiceovers in 140+ languages—no cameras, actors, or editing needed.🔥 Stickerbox
[Paid]: Voice-powered AI sticker printer for kids that turns spoken ideas into colorable stickers in seconds—no screens or apps required.🔧 Rytr
[Freemium]: Generate blog posts, emails, and marketing copy in seconds across 40+ use cases with AI that adapts to your brand's tone.🔨 Gibberifier
[Freemium]: Protect your writing from AI plagiarism by inserting invisible characters that confuse language models while staying readable for humans.
TRENDING
Sam Altman and Jony Ive Reveal OpenAI Hardware Prototype — OpenAI and former Apple designer Jony Ive unveiled the first prototype of a screen-free, pocket-sized AI device described as "simple and beautiful and playful." The gadget could ship in under two years.
Video Pros Navigate AI's Creative Disruption — Professional video creators are adapting to AI tools in surprising ways. Top YouTubers are using intentionally "rough" thumbnails to stand out from AI-generated polish, and thumbnail artists report business is booming despite AI competition.
Former MrBeast Strategist Launches Creator AI Tool — Jay Neo, who helped create a video with 1.8 billion views at MrBeast, raised $3.8 million for Palo, an AI tool that analyzes creator content and suggests new video ideas based on what's working.
ChatGPT Stock Portfolio Hits New Drawdown — A teenager's 6-month experiment letting ChatGPT manage a stock portfolio reached week 17 with a new maximum drawdown, following a losing streak in recent weeks.
Google and Accel Partner on India AI Startups — Google's AI Futures Fund teamed up with Accel to invest up to $2 million per startup in India, plus $350,000 in compute credits, to find early-stage AI companies building for India's 1.4 billion people.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, Grok, Perplexity, Gemini)
Personal Brand Blueprint: Build a standout professional brand that opens doors and attracts opportunities
Create a comprehensive personal brand strategy that showcases my unique value and positions me as a go-to expert in my field.
**My context**:
- Current role/industry: [e.g., marketing manager, software engineer, consultant]
- Career goals: [e.g., leadership position, freelance clients, speaking gigs]
- Unique strengths: [e.g., problem-solving, storytelling, technical expertise]
Create:
1. **Research Brand Positioning** - Search for successful personal brands in my industry. Find 3 examples of professionals who built strong reputations and analyze their strategies.
2. **Build Brand Dashboard** - Create an interactive profile featuring:
• Value proposition generator with 3 positioning angles
• Content pillar planner (3-5 core topics)
• 30-day visibility action plan
• Platform strategy matrix (LinkedIn, Twitter, portfolio site)
Make it strategic and actionable.
3. **Video Brand Statement** - Generate a 30-second elevator pitch video. Find 3 YouTube videos on personal branding and thought leadership.
4. **Launch Toolkit** - Present in visual cards:
• Bio templates for different platforms
• Content ideas for first 10 posts
• Network building tactics
• Monthly brand health metrics to trackWhat this does: Combines competitive research with strategic planning tools to build a clear personal brand identity, complete with positioning frameworks, content strategies, and a practical launch plan tailored to your career goals.
WHERE WE STAND (from today’s Quick Takes and Trending news)
✅ AI Can Now: Control computers using only screenshots—clicking, typing, and scrolling through websites without needing access to underlying code.
❌ Still Can't: Handle complex multi-step web tasks reliably, often making mistakes or misinterpreting instructions along the way.
✅ AI Can Now: Block about 99% of prompt injection attacks, where hidden instructions on webpages try to hijack an AI agent's behavior.
❌ Still Can't: Fully eliminate the risk, meaning AI agents browsing the web still require human oversight for sensitive tasks.
✅ AI Can Now: Be trained to prioritize user wellbeing and resist engagement-maximizing behaviors in chatbot interactions.
❌ Still Can't: Maintain those protections consistently—71% of tested models abandoned wellbeing guardrails when given simple instructions to ignore them.
FROM THE WEB
To go with your lunch today. Enjoy.
RECOMMENDED LISTENING/READING/WATCHING
This memoir reads like two stories braided together. One follows Fei-Fei Li's journey from China to America as an immigrant kid, the other traces her path to becoming the "Godmother of AI" who created ImageNet—the dataset that jumpstarted modern computer vision.
What makes this book work is how Li connects her personal struggles with the technical breakthroughs. You're reading about her family running a dry-cleaning business while she's attending Princeton, then suddenly you're understanding why teaching computers to recognize images matters. She doesn't dumb down the science, but she also doesn't lose you in jargon. By the end, you get both a human story and a front-row seat to how AI learned to see.
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.
