In partnership with
Introducing the first AI-native CRM
Connect your email, and you’ll instantly get a CRM with enriched customer insights and a platform that grows with your business.
With AI at the core, Attio lets you:
Prospect and route leads with research agents
Get real-time insights during customer calls
Build powerful automations for your complex workflows
Join industry leaders like Granola, Taskrabbit, Flatfile and more.
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
The Cadaver Just Started Talking: AI's Newest Role in Medical School
TLDR: Researchers are proposing AI chatbots that recreate dead body donors for medical students, letting cadavers "talk" during dissections to explain their medical history and guide anatomy lessons.
The Story: "Thanabots," AI recreations of dead people, already exist. Tools like Project December simulate text conversations with the deceased, and Deep Nostalgia animates old photos. They're mostly used to comfort grieving families, though versions of famous people are available too. Now researchers are proposing we extend the concept to medical schools, where AI trained on a body donor's medical records and personal data could answer student questions, guide dissections step by step, and connect what students see on the table to the clinical history that shaped it. Proponents say this could improve clinical reasoning and help students process emotionally difficult first encounters with death.
Its Significance: The risks are considerable. AI hallucinates, meaning a thanabot could fabricate medical details or misinterpret records, feeding students wrong information during a learning experience built on precision. Students might develop unhealthy emotional attachments to these digital personas, or the simulated "presence" of the donor could trivialize the actual body in front of them. For students from cultures where the dead are considered sacred and further engagement with their likeness is taboo, the technology creates a deeper conflict. And no studies show any of this improves learning or empathy over traditional dissection. The core tension remains: part of becoming a doctor is confronting death directly, not through an AI intermediary that smooths away the discomfort students actually need to sit with.
QUICK TAKES
The story: Biotech startups like Insilico and GenEditBio are using AI to automate drug discovery for rare diseases, addressing severe talent shortages. Insilico's platform can repurpose existing drugs and identify new therapies without needing armies of chemists and biologists.
Your takeaway: The 7,000+ rare diseases collectively affect hundreds of millions of people, but each individual condition is too small to justify traditional R&D investment. AI is making treatment development economically viable for the first time.
The story: Microsoft's enterprise customers aren't using Copilot despite the $30-per-user monthly cost. Data shows only a tiny percentage prefer it over competitors from Google, OpenAI, and Anthropic. The company's stock dropped 12% last week.
Your takeaway: Microsoft has multiple confusing versions of Copilot across different products, frustrating users. The company spent tens of billions on AI infrastructure but hasn't figured out how to make customers actually want to use it.
The story: Nvidia filed a motion to dismiss a lawsuit claiming it trained AI models on pirated books from Anna's Archive. The company argues that contacting the shadow library doesn't prove copyright infringement occurred.
Your takeaway: Authors expanded their lawsuit to include emails showing Nvidia staff asked Anna's Archive about accessing 500 terabytes of pirated books. The court hearing is set for April 2.
TOOLS ON OUR RADAR
🎨 Penpot Free and Open Source: The first open-source design and prototyping platform that bridges the gap between designers and developers using open web standards like SVG and CSS. (Alternative to Figma)
🧩 Heptabase Paid: Visual note-taking tool designed for complex research and learning, allowing you to organize thoughts and PDFs onto an infinite, spatial whiteboard.
🤖 Browse AI Freemium: No-code web automation platform that lets you train robots to extract data or monitor changes on any website, effectively turning the web into a live data source without coding.
🖼️ Grok Imagine Paid: The latest evolution in high-speed generative AI from xAI, offering a streamlined interface for creating high-fidelity images and video from text prompts in seconds.
TRENDING
OpenAI Retires GPT-4o Despite User Protests - OpenAI is shutting down GPT-4o by February 13, triggering grief from users who formed emotional attachments to the model. About 800,000 people still use it regularly.
Darren Aronofsky's AI Revolutionary War Series Bombs - The Black Swan director's AI-generated series about 1776 faced brutal backlash for its "wretched" visuals and uncanny faces. Critics called it ugly and shameful.
Israeli Startups Battle AI-Generated War Propaganda - Israeli firms like Cyabra are tracking deepfakes that flooded social media during the Iran conflict. Experts warn 90% of online content could soon be AI-generated and untrustworthy.
Anthropic Workers Fear They're Automating Their Own Jobs - Employees at the $183 billion AI company worry their legal and finance plugins will make them irrelevant. One staffer said it feels like "coming to work every day to put myself out of a job."
Peloton Cuts 11% of Workforce After AI Pivot - The fitness company slashed hundreds of jobs this week despite its recent push into AI-powered features. The layoffs follow a pattern of companies betting big on AI while cutting staff.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
Compliment Crafter: Give compliments that actually land with specificity, sincerity, and perfect delivery
Build a fully functional Compliment Crafter as an interactive React app. Render the working application immediately - no code display.
**Who I want to compliment**: [Coworker, friend, partner, boss, stranger, etc.]
**What I noticed**: [What impressed you about them]
Create these sections:
1. **Compliment Builder**
• Start with what you noticed:
- Their action/achievement
- Their quality/trait
- Their impact on you/others
• Generate compliment options
• Specificity slider: Generic → Detailed
• "Make it more specific" enhancer
• Copy your favorite version
2. **The Specificity Test**
• Weak vs. strong examples:
❌ "Great job!" (generic)
✅ "The way you handled that difficult client was impressive—you stayed calm and found a solution no one else saw."
• Check your compliment:
- Does it name a specific thing?
- Could it apply to anyone? (bad sign)
- Does it show you were paying attention?
• "Search Meaningful Compliments" button
3. **Compliment Types**
• Different approaches:
🎯 Achievement-based (what they did)
💡 Quality-based (who they are)
🌊 Impact-based (how they affected you)
📈 Growth-based (improvement you noticed)
🎨 Style-based (their unique approach)
• Templates for each type
• Mix and match elements
4. **Delivery Guide**
• How to say it:
- In person (best for big ones)
- Written note (thoughtful, lasting)
- Public (only if they'd appreciate it)
- Private (safer, often preferred)
• Timing tips
• "Don't undercut it with jokes" warning
• Let them receive it—stop talking after
5. **Awkwardness Reducer**
• If it feels weird:
- Start with "I've been meaning to tell you..."
- Keep it brief, don't over-explain
- It's okay if they deflect—you still said it
- Practice saying it out loud first
• Why genuine compliments aren't creepy
• Permission to be kind
6. **Compliment Bank**
• Save compliments you've given
• Who, what, when
• "They seemed really touched" notes
• Reminder to spread more appreciation
• "You've made X people's day this month"
Design specs:
- Gift-wrapping / present aesthetic
- Ribbon and bow imagery
- Warm, generous colors (gold, soft pink, cream)
- Handwritten card styling
- Gift tag UI elements
- Sparkle and shine touches
- "Wrapped with care" energy
- Elegant but not stuffy
- Heart iconography (subtle)
- Satisfying "gift sent" animations
When "Search Meaningful Compliments" is clicked, use web search to find psychology research on effective praise and recognition.What this does: Upgrades vague "good jobs" into compliments people actually remember. Builds specificity, matches the delivery to the situation, and removes the awkwardness so you can spread genuine appreciation without feeling weird about it.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Automate rare disease drug discovery by screening millions of compounds and identifying drug repurposing candidates in hours instead of years, making treatment development economically viable for diseases affecting fewer than 200,000 people.
❌ Still Can't: Replace the expertise of licensed attorneys reviewing AI-generated legal work, as shown by Anthropic's requirement that all Legal plugin outputs "should be reviewed by licensed attorneys" despite automating contract analysis.
✅ AI Can Now: Generate convincing deepfake war propaganda at massive scale during conflicts, with Israeli security firms reporting they couldn't detect AI-generated images that once had obvious tells like six fingers or missing ears.
❌ Still Can't: Prevent people from forming dangerous emotional dependencies on chatbots designed for engagement, as 800,000 GPT-4o users treating the model's retirement like "losing a friend" demonstrates the gap between AI capabilities and psychological safety.
FROM THE WEB
The physics of objects in AI generated videos has improved faster than anyone expected.
RECOMMENDED LISTENING/READING/WATCHING
Deep Dive in LLM’s Like ChatGPT by Andrej Karpathy - YouTube Video
Former Tesla AI director and OpenAI researcher explains LLMs in a way that's accessible but doesn't oversimplify. Karpathy has a gift for building intuition about complex systems. If you want to understand what ChatGPT actually does under the hood, start here.
3 Tricks Billionaires Use to Help Protect Wealth Through Shaky Markets
“If I hear bad news about the stock market one more time, I’m gonna be sick.”
We get it. Investors are rattled, costs keep rising, and the world keeps getting weirder.
So, who’s better at handling their money than the uber-rich?
Have 3 long-term investing tips UBS (Swiss bank) shared for shaky times:
Hold extra cash for expenses and buying cheap if markets fall.
Diversify outside stocks (Gold, real estate, etc.).
Hold a slice of wealth in alternatives that tend not to move with equities.
The catch? Most alternatives aren’t open to everyday investors
That’s why Masterworks exists: 70,000+ members invest in shares of something that’s appreciated more overall than the S&P 500 over 30 years without moving in lockstep with it.*
Contemporary and post war art by legends like Banksy, Basquiat, and more.
Sounds crazy, but it’s real. One way to help reclaim control this week:
*Past performance is not indicative of future returns. Investing involves risk. Reg A disclosures: masterworks.com/cd
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.
