In partnership with
Global HR shouldn't require five tools per country
Your company going global shouldn’t mean endless headaches. Deel’s free guide shows you how to unify payroll, onboarding, and compliance across every country you operate in. No more juggling separate systems for the US, Europe, and APAC. No more Slack messages filling gaps. Just one consolidated approach that scales.
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
A Facebook Page Told 6,200 People This Writer Was Dead. He Wasn't.
TLDR: A Facebook page with 6,200 followers published an AI-generated image of Denver Broncos beat writer Cody Roark holding a child with "RIP" across it, falsely claiming he died in a "domestic violence incident."
The Story: Cody Roark, a sports journalist covering the Denver Broncos, discovered a Facebook page had posted a fake AI-generated memorial image claiming he was dead. The page was churning out four fabricated stories per day to its 6,200 followers. This isn't an isolated incident. AI-powered content farms are now targeting ordinary people, not just celebrities, with completely fabricated tragedies and scandals designed to generate engagement and ad revenue.
Its Significance: This is a troubling shift in AI misinformation. When slop pages start inventing deaths and crimes about everyday people, the damage goes beyond viral lies. It affects careers, relationships, and mental health. Your best defense: Never trust shocking news about anyone on social media without checking actual news sources first.
QUICK TAKES
The story: Penn State researchers found that rude prompts like "You poor creature, do you even know how to solve this?" produced more accurate answers than polite ones. Accuracy jumped from 75.8% for very polite prompts to 84.8% for very rude ones.
Your takeaway: Small changes in wording create big swings in output quality. The researchers don't recommend being mean (it could normalize toxic communication), but the finding shows how unpredictable these systems still are.
The story: Chainalysis reports crypto scam losses hit at least $14 billion on-chain in 2025, with projections exceeding $17 billion. Deepfake impersonation scams grew 1,400% year-over-year as criminals use face-swap software and AI chatbots to pose as government officials, support reps, and romantic partners.
Your takeaway: Be suspicious of any unexpected financial request, even if it looks and sounds like someone you trust. When scammers can create convincing videos of people you know, verification gets harder.
The story: ASUS marked its RTX 5070 Ti and 5060 Ti 16GB graphics cards as "end of life" just months after launch. The culprit? Memory chip makers are prioritizing AI data center components over gaming hardware.
Your takeaway: AI's appetite for specialized chips is now directly affecting what gamers can buy. This shortage could push GPU prices higher and limit options for anyone building a PC this year.
TOOLS ON OUR RADAR
🗣️ TalkPal* Freemium: GPT-powered AI language tutor that simulates real conversations in 57+ languages, helping you practice speaking and listening with instant feedback.
🎨 Miro AI Freemium: Turn sticky notes into structured diagrams, generate ideas from prompts, and summarize complex boards with AI-powered whiteboard collaboration.
🤖 DocsBot AI Freemium: Build a custom AI chatbot trained on your own documentation to answer customer questions 24/7.
🎬 OBS Studio Free and Open Source: Stream and record your screen in high quality—the tool used by professional streamers worldwide.
TRENDING
Baidu's ERNIE 5 Beats GPT-5.1 in Math, Cracks Global Top 10 - China's Baidu released ERNIE 5.0, which scored #8 globally on LMArena's text benchmark and #2 in math, beating OpenAI's GPT-5.1-High and becoming the only Chinese model in the top 10.
Google Separates Gemini 3 Thinking and Pro Usage Limits - AI Pro subscribers now get 300 Thinking prompts and 100 Pro prompts per day as separate pools instead of a shared 100, responding to user complaints about running out of access too quickly.
Karpathy Says Programming Is Experiencing a "Magnitude 9 Earthquake" - Former Tesla AI director Andrej Karpathy sparked panic among developers by admitting he feels "far behind" on AI tools, while some teams report 70-90% of their code is now AI-generated.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
Meeting ROI Calculator: Watch the dollars burn in real-time and discover if this meeting should've been an email
Build a fully functional Meeting ROI Calculator as an interactive React app. Render the working application immediately - no code display.
**My company/team context**: [Industry, team size, general salary range]
**Meeting I'm evaluating**: [What meeting you're questioning]
Create these sections:
1. **Meeting Cost Calculator**
• Input fields:
- Number of attendees
- Average hourly rate (or salary → auto-calculate)
- Meeting length in minutes
• Big dramatic cost display:
💰 "THIS MEETING COSTS: $X"
• Per-minute burn rate ticking up
• "Start Meeting" button that runs live counter
• Watch the money disappear in real-time
• Annual cost if this meeting repeats weekly
2. **The Salary Burn Ticker**
• Live visualization during meetings:
$0.00 → ticking up every second
• Sound effect toggle (cash register cha-ching)
• "You just burned another $100" milestone alerts
• Attendee breakdown: "$X per person so far"
• Dramatic red numbers
• "Stop the Bleeding" end meeting button
3. **Should This Meeting Exist?**
• Decision tree quiz:
- Is there a clear agenda? (No = email)
- Do decisions need to be made? (No = email)
- Do 2+ people need to discuss? (No = email)
- Could this be async? (Yes = Slack/Loom)
- Is this a status update? (Yes = definitely email)
• Verdict:
✅ MEETING JUSTIFIED
📧 SHOULD BE AN EMAIL
💬 SHOULD BE A SLACK MESSAGE
🎥 SHOULD BE A LOOM VIDEO
• "Search Meeting Alternatives" button
4. **Attendee Audit**
• For each person invited:
- Do they need to be here? (Yes/No/Maybe)
- Are they contributing or spectating?
- Could they get notes instead?
• "Cut List" of people to remove
• Savings calculator: "Remove 3 people = save $X"
• "Optional attendee" recommendations
• Core vs. peripheral sorter
5. **Meeting ROI Scorecard**
• After the meeting, evaluate:
- Decisions made: [number]
- Action items created: [number]
- Problems solved: [number]
- Value generated: $[estimate]
• ROI calculation:
Value Created ÷ Meeting Cost = X%
• Grades:
🏆 A+: ROI > 200% (worth it!)
✅ B: ROI 100-200% (justified)
😬 C: ROI 50-100% (questionable)
💸 D: ROI < 50% (waste of money)
🗑️ F: ROI near 0 (should've been an email)
6. **Meeting Graveyard**
• Track recurring meetings:
- Weekly standup: $X/year
- Monday planning: $X/year
- All-hands: $X/year
• Total annual meeting spend
• "Meetings to Kill" hit list
• Resurrection rules: what would make it worth it?
• "You spend X hours/year in meetings"
7. **Alternative Calculator**
• Compare costs:
- 1-hour meeting (8 people): $X
- 15-min standup (8 people): $Y
- Slack thread: $0
- Loom video + comments: ~$Z
- Email chain: $0
• Time saved per year with alternatives
• "Try this instead" recommendations
• Async-first policy builder
Design specs:
- Cash register/money counter aesthetic
- Dollar bills and coins imagery
- Green money colors with red for "waste"
- Ticking counters and number animations
- Receipt-style summaries
- "Ka-ching" playful money vibes
- Dramatic large dollar displays
- Slot machine energy for the verdict
- Spreadsheet/ledger touches for ROI
- Burning money animation (subtle)
- Calculator keypad styling
When "Search Meeting Alternatives" is clicked, use web search to find async communication best practices and meeting reduction strategies.What this does: Makes meeting costs painfully visible. Watch real dollars tick away during meetings, get verdicts on whether gatherings should exist at all, audit your attendee lists, and calculate the actual ROI of your time together. Perfect for convincing leadership that yes, this really could have been an email.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Generate deepfake videos convincing enough to steal $3.2 million per scam operation on average
❌ Still Can't: Give consistent answers, with accuracy swinging nearly 10 percentage points based on whether you say "please"
✅ AI Can Now: Connect to your medical records, Apple Health, and wellness apps to give personalized health guidance
❌ Still Can't: Be trusted for diagnosis or treatment, which is why OpenAI explicitly says ChatGPT Health isn't meant to replace doctors
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
Tristan Harris and Aza Raskin, the team behind Netflix's "The Social Dilemma," turn their attention to AI in this hour-long presentation from March 2023. Their argument: social media was humanity's first contact with AI (through recommendation algorithms), we fumbled it, and now large language models represent second contact with the same mistakes happening faster.
They cover emergent AI capabilities, deployment speed versus safety testing, and why global coordination matters. Fair warning: it's explicitly alarmist, and critics have disputed some of their statistics. But it's worth watching because this perspective shaped much of the AI safety conversation in 2023, influencing proposed regulations and the "responsible AI" movement.
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 product links may be affiliate or referral links. This helps support the newsletter at no extra cost to you.
