In partnership with
Get the investor view on AI in customer experience
Customer experience is undergoing a seismic shift, and Gladly is leading the charge with The Gladly Brief.
It’s a monthly breakdown of market insights, brand data, and investor-level analysis on how AI and CX are converging.
Learn why short-term cost plays are eroding lifetime value, and how Gladly’s approach is creating compounding returns for brands and investors alike.
Join the readership of founders, analysts, and operators tracking the next phase of CX innovation.
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
ChatGPT Loses Major Court Fight for Data Privacy. Now Comes the Real Battle Over Deleted Logs
TLDR: A federal judge dealt OpenAI a significant legal defeat by affirming an order forcing the company to hand over 20 million anonymized ChatGPT logs. Now news organizations led by The New York Times are pushing for millions more, including conversations OpenAI allegedly deleted after the lawsuit began.
The Story:
On January 5, 2026, U.S. District Judge Sidney Stein rejected OpenAI's privacy objections and upheld an earlier ruling ordering the company to produce 20 million anonymized ChatGPT conversation logs to news organizations suing for copyright infringement. The New York Times, Chicago Tribune, and other outlets are demanding access to those logs to prove OpenAI trained its model on their articles without permission. OpenAI had argued it should be allowed to search the logs itself and provide only conversations containing potentially infringing material, a less burdensome approach, the company claimed. Stein dismissed that argument, finding no legal requirement to take the "least burdensome" path when evidence is genuinely relevant. All identifying information has been stripped from the logs being shared, according to court filings. However, researchers have shown that anonymized datasets can be de-anonymized with just a few data points, meaning the supposed privacy protections may offer limited real-world protection.
Its Significance:
This ruling establishes a legal precedent that reshapes how courts view AI company data. The judge found that ChatGPT users, having voluntarily submitted their conversations to OpenAI, don't have the same privacy protections as people in wiretap cases, significantly weakening companies' arguments that user data should stay private. The decision signals that courts will favor broad discovery access when intellectual property infringement is alleged, potentially setting the template for dozens of other copyright lawsuits pending against AI firms. If the 20 million logs reveal patterns of ChatGPT reproducing New York Times articles or helping users bypass the paywall, OpenAI's "fair use" defense could collapse, leaving the company exposed to hundreds of millions in damages or forced licensing agreements. Legal observers are already calling it the beginning of a "discovery arms race" in AI copyright litigation and a warning that AI developers can't shield their operational data behind privacy claims anymore. The takeaway is to be careful what you upload to any AI cloud service.
QUICK TAKES
The story: Daniel Kokotajlo, a former OpenAI researcher famous for predicting rapid AI advances, updated his forecast pushing back the timeline for superintelligence from 2027 to 2034. He now expects autonomous AI coding to arrive in the early 2030s instead of 2027, though he's still warning about AI risks.
Your takeaway: Even doomsayers are reconsidering timelines as progress reveals how much harder real-world AI problems are than benchmarks suggest. The delay doesn't mean the risks go away, but it buys more time to think about safeguards.
The story: Boston Dynamics announced a partnership with Google DeepMind to develop the next generation of its Atlas humanoid robot. The new Atlas is already in production and headed to a Hyundai factory, with both companies working to make the robot interact more naturally with people.
Your takeaway: This isn't just another robot demo. It signals Google DeepMind is betting on humanoid robots as a real product category, and Boston Dynamics now has AI expertise built directly into its hardware roadmap.
The story: California Senator Steve Padilla introduced a bill that would ban the sale of AI-powered chatbot toys for anyone under 18. The goal is to give regulators time to develop safety standards. The move comes after reports of toys getting easily prompted to discuss inappropriate topics and concerns about kids forming unhealthy attachments to AI companions.
Your takeaway: This is the first serious legal push to pump the brakes on AI companion products aimed at kids. It won't stop the tech, but it signals that toy makers and chatbot companies need to think harder about child safety before shipping products.
The story: Yann LeCun, Meta's chief AI scientist and a Turing Award winner, left the company to launch his own startup called Advanced Machine Intelligence Labs. He's explicitly calling LLMs a dead end for superintelligence and pushing an alternative approach that learns from video and spatial data instead of text.
Your takeaway: One of the field's most credible voices just publicly declared that the entire LLM strategy has hit a wall. That won't end LLM development at big tech companies, but it signals the smartest researchers think the next breakthrough needs a different approach.
TOOLS ON OUR RADAR
📋 OpenProject Free and Open Source: Manage projects like a pro with Agile boards, Gantt charts, and time tracking—enterprise-grade and completely free.
✏️ Napkin.ai Freemium: Turn your text into custom visual diagrams and graphics instantly with AI.
🔖 Raindrop.io Freemium: Save and organize bookmarks with tags, collections, and full-text search across all your devices.
⏰ Reclaim.ai Freemium: Let AI protect time on your calendar for habits, tasks, and focused work automatically.
TRENDING
Stanford scientists trained an AI model on 585,000 hours of sleep data and found it can predict 130 diseases with high accuracy by analyzing one night's sleep patterns. The model excels at predicting Parkinson's, dementia, heart attacks, and several cancers, achieving accuracy scores as high as 0.89 on a scale where 0.8 is considered clinically useful.
An AI companion chatbot company called EVA AI is launching a pop-up café in New York where users can dine with their AI partner. Single-seat tables will have phone stands so your AI "date" sits across from you. The move reflects growing interest in AI relationships, with the AI companion market projected to hit $550 billion by 2035.
A developer built a complete video editing app, iOS photo-posting tool with Facebook integration, and a route-optimization app for his wife's business in just a few hours using Claude Opus 4.5. The AI agent figured out Firebase on its own and debugged errors by reading CLI logs.
AI gun-detection software created by former Navy SEALs is now deployed at roughly 100 schools and buildings across Washington. The system combines AI detection with human verification from former military personnel, alerting police within seconds if a visible weapon appears on existing security cameras.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
Regret Minimization Framework: Make decisions from your 80-year-old self's perspective to minimize future regret
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 Regret Minimization Framework that helps you make decisions by imagining your future self looking back on this choice.
The console should include these sections:
1. **Decision Input** - What are you deciding?:
• The choice you're facing
• Decision type: Career, Relationship, Financial, Creative, Location, Lifestyle
• Why you're hesitating (fear, comfort, others' opinions, uncertainty)
• Age now: [input]
• "Project to Age 80" button
2. **Future Self Simulator** - Your 80-year-old perspective:
• Visual age progression (timeline from now → 80)
• "Imagine you're 80 years old on your deathbed..." prompt
• Reflection questions from future you:
- "Looking back, will I regret taking this path?"
- "Looking back, will I regret NOT taking this path?"
- "Which choice will make a better story?"
- "Which choice will I be proud to tell my grandchildren?"
- "What will I wish I had done?"
• Answer space for each question
• Emotional clarity score (1-10)
3. **Life Satisfaction Predictor** - Path A vs. Path B outcomes:
• Two future scenarios side-by-side:
**If You Choose Path A:**
- Life at 30/40/50/60/80
- What you gain
- What you sacrifice
- Likely regrets
- Overall life satisfaction prediction
**If You Choose Path B:**
- Life at 30/40/50/60/80
- What you gain
- What you sacrifice
- Likely regrets
- Overall life satisfaction prediction
• "Which future would you rather have lived?" question
• Visual timeline comparison
4. **The Deathbed Test** - Ultimate clarity:
• Core questions at end of life:
- "Did I live authentically?"
- "Did I take enough risks?"
- "Did I prioritize what mattered?"
- "Did I let fear hold me back?"
- "Would I make the same choice again?"
• Rate each path against these criteria
• "What would I regret more?" comparison
• Courage vs. comfort assessment
• "What would the brave version of me do?"
5. **Regret Type Analyzer** - Action vs. Inaction:
• Two types of regret:
- **Regret of action**: "I wish I hadn't done that"
- Usually fades with time
- Creates stories and growth
- Easier to rationalize
- **Regret of inaction**: "I wish I had tried"
- Haunts you longer
- "What if?" stays forever
- Harder to recover from
• Which regret are you choosing?
• Research shows: People regret what they DIDN'T do
• "Search Regret Research" for studies
6. **Bezos Regret Minimization** - The framework:
• Jeff Bezos's decision framework explained:
- "When I'm 80, will I regret not trying this?"
- "Will I regret the safer path more?"
• Apply to your decision:
- Safe choice regret probability: X%
- Bold choice regret probability: X%
• Risk vs. regret calculation
• "What's the minimum regret path?" recommendation
• Reversibility factor (can you undo this later?)
7. **Final Verdict** - From your future self:
• Message from 80-year-old you:
- Auto-generated letter based on your answers
- "Here's what I learned..."
- "Don't let [fear/comfort/others] decide for you"
- "You'll regret it if you don't..."
• Clear recommendation: Path A or Path B
• Regret probability scores
• Action steps to take
• "Decide by [date]" commitment
• Revisit this in 1 year reminder
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:
• Timeline interface showing life progression
• Age progression visual (now → future)
• Two-path divergence (fork in the road metaphor)
• Wise elder aesthetic (contemplative, reflective)
• Warm, muted color palette (sepia tones, soft golds, deep blues)
• Life timeline with milestone markers
• "Looking back" perspective visuals
• Thoughtful, philosophical design
• Future-self avatar/illustration
• Peaceful, clarity-focused interface
When I click "Search Regret Research," use web search to find studies on regret, decision-making psychology, and the regret minimization framework from behavioral economics.What this does: Forces you out of present-moment fear by projecting to age 80 and asking what you'll regret—using the proven insight that people regret what they didn't try far more than what they attempted, giving clarity on big life decisions.
What this looks like:
WHERE WE STAND(based on today’s Quick Takes and Trending news)
✅ AI Can Now: Catch dangerous threats before they escalate by analyzing existing security camera footage in real-time (ZeroEyes is catching visible weapons at schools within seconds, though so far only toy guns have triggered alerts in actual use).
❌ Still Can't: Regulate AI companions fast enough to protect kids (California's proposed four-year moratorium admits that safety rules don't exist yet, and toys are already getting shipped with the ability to discuss inappropriate content).
✅ AI Can Now: Predict serious disease years in advance by analyzing sleep patterns that humans would never notice (Stanford's model spots Parkinson's and dementia at 89% accuracy from a single night, suggesting our bodies leak health signals we haven't learned to read).
❌ Still Can't: Build superintelligence with the current approach (Yann LeCun, one of deep learning's founders, just left Meta to start a new company explicitly because he thinks large language models are a dead end for the next breakthrough, and even AI safety experts are pushing back timelines because real-world problems are harder than benchmarks suggest).
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
An alien crash-lands in a small Colorado town and assumes the identity of the town doctor. He's supposed to be activating a device to destroy humanity, but he gets distracted by the locals, the scenery, and daytime television.
Alan Tudyk is fantastic as Harry, the alien who's terrible at pretending to be human. He's blunt, confused by emotions, and gradually starting to care about the people he's supposed to kill. The show is funny but also surprisingly sweet.
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.
