- Superhuman AI
- Posts
- AI explores origins of life
AI explores origins of life
ALSO: How to use ChatGPT Deep Research like a pro

Welcome back, Superhuman. While we might use agents for things like building a website or competitive analysis, NASA and Georgia Tech are aiming higher. They just built a family of agents that could help solve some of life’s biggest questions — like “Are we alone?” and “Where did life come from?”
Today’s Insights
Apple embraces vibe coding, Suno drops v4.5, and OpenAI makes amends
NASA uses agents to explore the origins of life
Tutorial: How to use ChatGPT Deep Research like a pro
5 new AI tools to boost your productivity
News, memes, what’s trending on socials, and more
TODAY IN AI

Suno v4.5 brings a new level of precision to the song generation platform. Source: Suno
1. Apple hops on the vibe-coding train: The iPhone maker’s coding platform Xcode could soon get an AI overhaul thanks to a new collaboration with Anthropic. The Claude-powered tool, which can “write, edit, and test code” on its own, is being tested internally with a possible public release down the line. But Amazon’s not far behind: New reports claim it’s building an AI coding agent to take on industry-favorite Cursor. Meanwhile, Google and Microsoft admit 25-30% of their code is now written by AI, while Meta’s AI output could reach 50% in the next year.
2. Music-gen app Suno gets major upgrade: AI music generation is one of those corners of the industry that quietly keeps getting better, and Suno’s new update is a reminder it’s moving faster than most people realize. The platform, which lets you turn text prompts, images, and videos into songs, just unveiled v4.5, offering crisper sound quality, richer vocals, and better prompt adherence. You can also mix and match genres, and it looks like you can get pretty niche — like neo-soul meets Gregorian chant.
3. OpenAI digs deeper into GPT-4o’s ‘sycophant’ phase: The startup recently reversed an update that led GPT-4o to back you up, even if you mentioned doing something dangerous. Now, OpenAI is doubling down on last week’s postmortem with a plan for how it’ll prevent another mishap. Since user feedback can skew toward positivity, OpenAI said it’ll add independent metrics to make sure ChatGPT’s personality stays even-keeled. It’ll also rely more on experts who can pick up on overly “agreeable” behavior before it’s too late.
PRESENTED BY GUIDDE
Remote training will never be the same

Why send a PDF full of complex info when you can create a beautiful, branded How-To video instead–with zero effort?
Using Guidde, you can turn static docs into stunning video guides in less than 5 minutes.
Guidde will create a professional video with transitions and imagery
Add logos and customize colors/fonts
Share instantly, in any language!
Track analytics and optimize engagement over time
The best part? It costs nothing to use.
FROM THE FRONTIER
NASA sends agents off to search for the origins of life

NASA will soon use agents to analyze samples from Mars. Source: NASA
Just about every week, a new science-focused agent hits the market. But most of them zoom in to investigate things like chemical reactions and protein shapes. NASA and Georgia Tech, on the other hand, are pointing their agents toward the stars.
They just built eight AstroAgents, each tasked with poring over mass spectrometry data (which shows the chemical composition of different objects) to come up with new hypotheses about the origins of life.
A promising start: The AstroAgents system has already analyzed soil samples from closer to home, including Antarctica and the Atacama Desert, to generate dozens of “plausible” hypotheses. For example, one posits that a certain combination of molecules points to life being nearby.
Another focus involves taking a close look at meteorites, seeing what they’re made of, and trying to formulate “how life’s building blocks could form in space and eventually find their way to earth,” one researcher writes.
Eventually, the researchers will use the agents to examine lakebed samples from Mars, to help determine whether they have any trace of organic material. The agents are even able to outperform human scientists, since they can analyze so much information at once and spot patterns that would otherwise go unnoticed.
THE AI ACADEMY
How to use ChatGPT Deep Research like a pro
ChatGPT’s Deep Research is the best AI product I’ve used this year and it might be the first one that surpasses humans in research ability. Watch the full tutorial here.
Go to ChatGPT and sign up.
Use ‘03-mini high’ or ‘03-mini’ reasoning models and select Deep Research.
Type a detailed, specific prompt with clear goals.
Answer any follow-up questions and wait 5-60 minutes for the generated report.
Review the report, including sources, summary, and ask follow-up questions for further details.
Note: Deep Research was opened to ChatGPT Plus users shortly after I recorded this video, so you no longer need the Pro plan to use it.
PRESENTED BY NEBIUS
Nebius research credits program

We created this program to support academic innovation by giving researchers access to GPU cloud or AI models for inference tasks.
Our goal is to help drive scientific breakthroughs by enabling advanced research, data analysis, and machine learning to tackle big challenges.
AI & TECH NEWS
Everything else you need to know today

Google teased a new NotebookLM app set to release in a few weeks. Source: Google
📱 By Popular Demand: Google will soon launch a standalone NotebookLM app centered around its beloved, podcast-like AI Overviews feature. You can join the waitlist via the App Store or the Play Store.
🖼️ Best of Both Worlds: Midjourney launched a new feature called Omni Reference that lets you seamlessly insert any character or object into a scene, or mix and match elements from different images.
🔍 Access Granted: All paid Claude users can now use the chatbot’s internet search features, while Max subscribers (the $100/month tier) are getting full access to Claude Code directly in their terminal.
🤝 Branching Out: The United Arab Emirates’ AI giant G42 is reportedly making a push into the US. That comes as the UAE vows to invest up to $1.4T in the US over the next 10 years.
🧠 Fast Track: The FDA has just given Neuralink’s brain chip breakthrough status for treating severe speech impairment, speeding up its development timeline. The Elon Musk-led startup is also adding new AI features to the device, as seen in this video.
PRODUCTIVITY
5 AI Tools to Supercharge Your Productivity
✅ Stackie: Organize effortlessly by logging, tracking, recalling, and reflecting with AI.
✅ CrePal: Automate video creation for different platforms with AI agents.
✅ Reef: Use AI to analyze trends, create visualizations, and explain insights through engaging podcasts.
✅ Cerebro: Turn videos, articles, and books into searchable, and actionable knowledge.
✅ Napkin AI: Transform text into visuals for effective business storytelling.
🎁 Want more? Check out our Top 125 AI Tools
* indicates a promoted tool, if any
SOCIAL SIGNALS
What’s trending on socials today

🎮 Trial By Fire: Google’s Gemini 2.5 Pro just conquered Pokémon Blue after several hundred hours of work. Commenters said we’ve finally reached API: Artificial Pokémon Intelligence.
🏆 Different Strokes: Can you guess which model developers on Cursor love the most? The coding assistant just revealed its five most-used LLMs, plus the fastest climbers.
💼 Competitive Edge: Founder Matt Shumer shares a handy prompt that helps you generate an actionable list of growth moves for your business. Or, try Appollonator3000’s prompt for creating a business model optimized for how you think and what you’re wired for.
✂️ Hairline Humor: Perplexity is now trolling its own boss, CEO Aravind Srinivas, and commenters are only adding fuel to the fire.
🚨 Code Red: An eerie video is going around showing a robot going “berserk” during a test. One commenter speculated it might have been caused by a glitch where the robot tries to recover its balance but can’t, since it’s suspended in the air.
PROMPT OF THE DAY
Post-interview feedback
Prompt: Act as a hiring manager and communication expert. I've just completed an interview (or conducted one), and I want to write or receive detailed, constructive post-interview feedback. If I'm the candidate, help me simulate what the interviewer might be thinking — based on my answers, communication style, and any gaps or strengths I displayed. Suggest what I likely did well, what could be improved, and how I might follow up with a thoughtful email. If I'm the interviewer, help me write feedback that is specific, respectful, and helpful — including comments on technical skills, culture fit, communication, and growth potential. Include examples of effective phrasing for both positive and developmental feedback. If possible, offer a template for summarizing strengths, weaknesses, and next steps. End with one reflection question I can ask myself to improve future interviews — whether I'm on the giving or receiving side.
🤓 Want more? Check out our Top 1,000 Prompts
AI-GENERATED IMAGES
Monochrome Majesty

Source: starikoff7016
Midjourney Prompt: David Yarrow's wildlife photography depicting … --ar 9:16 --v 7 --stylize 750 --profile 8dzbccy gvnx5n7 vkfclms
Acquire new customers and drive revenue by partnering with us
Superhuman is the world’s biggest AI newsletter for businesses and professionals with 1M+ readers and 2M+ followers on socials working at the world’s leading startups and enterprises. Companies like Amazon, Hubspot, and Salesforce feature their products in Superhuman. You can learn more about partnering with us here.
🧞 Your wish is my command
What did you think of today's email?Your feedback helps me create better emails for you! |
Got more feedback or just want to get in touch? Reply to this email and we’ll get back to you.
Thanks for reading.
Until next time!
Zain & the Superhuman AI team