Smarter CX insights for investors and founders
Join The Gladly Brief for insights on how AI, satisfaction, and loyalty intersect to shape modern business outcomes. Subscribe now to see how Gladly is redefining customer experience as an engine of growth—not a cost center.
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 edited, and published with the intention of making AI news and technology more accessible to everyone.
THE FRONT PAGE
Nvidia's New AI Chips Promise 10x Lower Costs. Here's What That Means for You
TLDR: Nvidia unveiled its next-generation AI platform at CES 2026, promising 5x more performance than current chips while slashing costs by 90%, and announced that a car powered by its new "thinking" self-driving AI ships to U.S. roads this quarter.
The Story:
Jensen Huang took the stage in Las Vegas to reveal Vera Rubin, Nvidia's successor to the Blackwell chips that currently power most AI data centers. The numbers are staggering: 50 petaflops of AI inference performance, 5x what Blackwell delivers, with a promised 10x reduction in the cost per token for AI models. Microsoft, AWS, Google Cloud, and Oracle are already lined up to deploy it. But the consumer news stole the show: Nvidia announced Alpamayo, a 10-billion-parameter open-source AI model for self-driving cars that doesn't just react to the road. It reasons through decisions and explains why it made them. The first car running this system, the Mercedes-Benz CLA, ships to U.S. customers in Q1 2026.
Its Significance:
Here's the strange position Nvidia now occupies: it sells the picks and shovels to every major AI company, then uses that money to build competing products. OpenAI, Anthropic, Google, Meta, and Tesla all depend on Nvidia GPUs to train their models. Tesla specifically relies on Nvidia hardware to develop Full Self-Driving. Now Nvidia is shipping its own self-driving software that directly competes with Tesla's FSD, and it's giving it away for free on Hugging Face. Any automaker can grab Alpamayo and run it on Nvidia chips, potentially commoditizing what Tesla has spent years trying to perfect. The data center dominance funds the expansion into cars, robots, and healthcare AI. The 10x cost reduction in Rubin chips means cheaper AI infrastructure across the board, which accelerates adoption everywhere. But it also means Nvidia's customers are funding the R&D for products that may eventually compete with their own. Huang called this "the ChatGPT moment for physical AI."
QUICK TAKES
The story: A viral Reddit post claiming to expose fraud at a food delivery app got 86,000 upvotes and 36 million views on X. When Platformer's Casey Newton investigated, the "whistleblower" sent an AI-generated employee badge and an 18-page fake technical document created by an LLM.
Your takeaway: Hoaxes that once took days to fabricate now take minutes. Newton caught this one because Gemini flagged its own watermark on the fake badge, but next time we might not be so lucky.
The story: Polymarket partnered with real estate data firm Parcl to let crypto traders bet on home prices in major U.S. cities like Austin, Miami, and San Francisco. Markets settle against Parcl's daily housing indices, so there's no debate about outcomes.
Your takeaway: Real estate is the world's largest asset class, worth over $650 trillion globally. Now anyone can trade on housing trends without buying property or taking on leverage.
The story: Amazon launched Alexa.com at CES, letting Early Access users chat with Alexa+ through any browser. The company also redesigned its mobile app to put the AI chatbot front and center.
Your takeaway: Amazon sold 600 million Alexa devices, but that's not enough anymore. To compete with ChatGPT and Gemini, Alexa needs to work everywhere you are, not just in your kitchen.
The story: Families suing OpenAI over wrongful deaths say the company is selectively sharing chat logs in court cases. OpenAI's terms don't address what happens to conversations after a user dies, and the company retains them indefinitely unless manually deleted.
Your takeaway: Google and other tech companies have policies for managing accounts after death. OpenAI doesn't, which leaves families with no way to access or control their loved ones' AI conversations.
TOOLS ON OUR RADAR
🔐 Standard Notes Free and Open Source: Take encrypted notes that sync across devices with a 100-year commitment to longevity and zero-knowledge architecture.
🍅 Focus To-Do Freemium: Combine Pomodoro timer with task management to power through focused work sessions.
📧 Clean Email Freemium: Bulk clean your inbox with smart views that help you unsubscribe and organize thousands of messages at once.
🎙️ Notta Freemium: Transcribe meetings in 58 languages with AI-generated summaries and action items.
TRENDING
Google TV Getting Gemini Features at CES - Google previewed AI features for TVs including voice-controlled settings ("the dialogue is lost"), photo search, and "deep dives" that create narrated overviews of topics for the whole family.
AI Chatbots Misinterpret Medical Risk Terms - Vanderbilt researchers found ChatGPT, Gemini, Grok, and Claude define "rare" as affecting up to 4% of people on average. Doctors mean 1 in 1,000.
MIT Tests Whether Clinical AI Leaks Patient Data - Researchers developed tests to check if AI models trained on medical records might reveal individual patient information when prompted by bad actors.
Wisconsin Bill Would Criminalize AI Deepfake Scams - A proposed law would make it a misdemeanor to create deepfakes for harassment and a felony to use them for fraud. The bill is modeled after similar legislation that passed in Pennsylvania.
AI Triples Noise Removal in Brain Scans - Boston College researchers built a tool called DeepCor that removes image distortions from fMRI data 200% better than previous methods, which could sharpen research on brain disorders.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
Voice Tone Analyzer: Improve your speaking delivery with pace tracking, filler word detection, and energy level monitoring
Build this as a fully functional interactive web application that renders immediately. Do not show code - create the actual working app.
Build me an interactive Voice Tone Analyzer that helps you practice and improve speaking delivery for presentations, podcasts, and important conversations.
The console should include these sections:
1. **Practice Setup** - What are you working on?:
• Speaking context: Presentation, Podcast, Interview, Sales pitch, Teaching, Casual conversation
• Length: 2 min, 5 min, 10 min, 15+ min
• Goal: Sound more confident, Reduce filler words, Better pacing, More engaging, Professional tone
• Practice mode: Live feedback, Record & analyze
• "Start Practice" button
2. **Live Voice Monitor** - Real-time feedback:
• Simulated audio waveform display (visual representation)
• Current metrics dashboard:
- **Speaking pace**: Words per minute (target 140-160)
- **Pause frequency**: Natural breaks vs. rushed
- **Volume consistency**: Too quiet, good, too loud
- **Energy level**: Monotone → Engaging → Too hyped
• Visual indicators (meters, gauges)
• Color-coded feedback (green = good, yellow = adjust, red = issue)
• Timer showing elapsed time
3. **Filler Word Counter** - Reduce verbal tics:
• Track common fillers in real-time:
- "Um" / "Uh" count
- "Like" count
- "You know" count
- "So" (at start of sentences)
- "Actually" / "Basically"
• Total filler count with frequency (per minute)
• Visual alert when filler detected
• "Filler-free streak" timer
• Goal: Under X fillers per minute
• Alternative: "Pause instead" reminder
4. **Pace & Rhythm Tracker** - Speaking speed analysis:
• Words per minute calculator (simulated)
• Pace zones:
- Too slow (<120 WPM): Risk losing attention
- Ideal (140-160 WPM): Natural, engaging
- Too fast (>180 WPM): Hard to follow
• Speed variation graph (good speakers vary pace)
• Pause quality assessment:
- Strategic pauses (emphasis, thinking)
- Awkward silences (filler word replacement)
- Rushed speech (no breathing room)
• "Breathe here" reminders
5. **Energy & Emotion Detector** - Engagement level:
• Vocal energy meter (1-10):
- Monotone (1-3): Sounds bored/tired
- Conversational (4-6): Natural, authentic
- Energetic (7-9): Engaging, enthusiastic
- Over-the-top (10): Might be too much
• Emotion indicators detected:
- Confident vs. Uncertain
- Calm vs. Nervous
- Warm vs. Cold
- Authentic vs. Scripted
• Variation tracking (avoid monotone)
• "Where to add emphasis" suggestions
6. **Session Analysis** - Post-practice review:
• Session summary:
- Total time spoken
- Average pace (WPM)
- Filler word count and breakdown
- Energy consistency score
- Pause quality rating
• Improvement areas highlighted:
- Top issue (pace too fast, too many fillers, etc.)
- What you did well
- Specific practice exercises
• Compare to previous sessions (getting better?)
• "Record yourself" recommendation for actual analysis
7. **Practice Exercises** - Targeted improvement:
• Pace drills: Read passage at target speed
• Filler elimination: Conscious pause practice
• Energy variation: Emphasize key words
• Breathing exercises: Support better pacing
• "Search Speaking Tips" for expert advice
• Video examples of great speakers
• Weekly practice goal tracker
Create this as a complete, working application with all functionality implemented. The app should render immediately when generated, not display as code.
Visual design specifications:
• Audio waveform aesthetic (sound wave visualizations)
• Real-time meters and gauges (podcast/radio studio vibe)
• Dark theme with neon accent colors (green, blue, orange)
• Pulsing visual feedback
• Sound engineer control panel style
• Clean, modern audio interface
• Live monitoring dashboard
• Recording studio inspiration
When I click "Search Speaking Tips," use web search to find voice coaching techniques, public speaking best practices, and exercises for improving vocal delivery.What this does: Provides real-time feedback on speaking delivery—tracking pace, counting filler words, monitoring energy levels, and identifying improvement areas so you can practice becoming a more confident, engaging speaker.
What this looks like:
WHERE WE STAND(based on today’s Quick Takes and Trending news)
✅ AI Can Now: Bet on whether your city's home prices will rise or fall next quarter, with outcomes verified against daily data feeds instead of monthly government reports.
❌ Still Can't: Tell you what "rare" means in a medical context. When doctors say a side effect is rare, they mean 1 in 1,000. AI chatbots think it means 1 in 25.
✅ AI Can Now: Generate fake employee badges, technical documents, and whistleblower accounts convincing enough to fool Redditors and rack up millions of views before being discovered.
❌ Still Can't: Be trusted to keep your conversations private after you die. OpenAI has no policy for deceased users' data, leaving families locked out while the company decides what to share in court.
RECOMMENDED LISTENING/READING/WATCHING
David is a robot child programmed to love his human mother unconditionally. When she abandons him, he sets out to find the Blue Fairy from Pinocchio, convinced that becoming a real boy will make her love him back.
Spielberg took over this project after Stanley Kubrick died, and the film feels caught between their two sensibilities. The first half is Kubrick: cold, disturbing, raising questions about what we owe beings we create. The second half is Spielberg: emotional, fantastical, reaching for hope. The ending is divisive, but the film's questions about love and consciousness linger. Haley Joel Osment is heartbreaking as David, a child who can't stop loving even when it destroys him.
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.
Some links may be affiliate or referral links. This helps support the newsletter at no extra cost to you.
