🤔 FRIDAY FACTS

Can AI Really Invent New Life-Like Proteins From Scratch?

A team using a model called ESM3 just simulated 500 million years of evolution in months—and designed a glowing protein that nature never dreamed up. But how exactly did AI pull off a molecular feat billions of years in the making?

Stick around to find out! 👇

🗞️ YOUR DAILY ROLLUP

Top Stories of the Day

👔 Meta Hires Top OpenAI Researcher for AI Push
Meta has hired Trapit Bansal, a key OpenAI researcher behind its first AI reasoning model, to join its new AI superintelligence unit. Bansal brings deep expertise in reinforcement learning and joins a growing team of elite talent from OpenAI and DeepMind. The move signals Meta’s urgency to compete in the AI reasoning race, where rivals have taken the lead.

⏳ DeepSeek R2 Launch Delayed by Export Controls
Chinese AI startup DeepSeek has postponed the R2 model release after CEO Liang Wenfeng deemed its performance unsatisfactory. U.S. export restrictions on NVIDIA H20 server chips have compounded the delay, limiting hardware access in China. Engineers continue refining the model while coordinating with cloud providers for future deployment.

🦉 AI Referee Debuts at X Games, Goes Global
The X Games introduced “The Owl,” an AI judge that accurately predicted snowboard podiums in Aspen. Now launched as a startup with $11M in funding, it aims to enhance fairness in judged sports worldwide. Built with athlete input, it complements human judges and offers real-time, multilingual commentary tailored to viewer expertise.

📉 AI Companionship Use Far Rarer Than Assumed
Despite popular narratives, people rarely use AI for emotional support. A new report from Anthropic shows only 2.9% of Claude interactions involve personal advice or coaching, with true companionship and roleplay making up less than 0.5%. Most users rely on Claude for productivity and content creation. Still, emotional use can emerge in longer, distress-driven chats.

🔋 Meta Expands Clean Energy Deals for AI Data Centers
Meta has signed four new agreements with Invenergy to source 791 MW of solar and wind power for its data centers, boosting their total partnership to 1,800 MW. The move supports Meta’s growing AI-driven energy needs while aligning with its clean energy goals. Power will be generated in Ohio, Arkansas, and Texas, with Meta receiving associated clean energy credits.

Enjoying our newsletter? Forward it to a colleague—
it’s one of the best ways to support us.

🧑‍🏫 FORWARD FUTURE

Your New Favorite Prompt Guide. Until the Next One Drops…

Most guides on prompt engineering either overcomplicate the basics or give you vague tips that don’t hold up when you actually try to use them.

This one’s different.

We’ve done the homework for you — gathering best practices from OpenAI, Google, Anthropic, and real-world testing — and broken it all down into something you can actually use.

Whether you’re in sales, operations, marketing, engineering or leadership, this guide will show you how to get dramatically better results from ChatGPT, Claude, Gemini, and other chat-based AI tools. → Read the full article here.

🚀 HYPECONOMICS

AI Startups Soar on Hype, Not Fundamentals

AI startups are fetching sky-high valuations with little regard for fundamentals, as VCs embrace what some call "vibe valuing." Mira Murati’s Thinking Machines Lab raised $2B at a $10B valuation before defining a clear strategy—its appeal rooted in her OpenAI background.

Traditional metrics like annual recurring revenue (ARR) are being sidelined in favor of “experimental run rate” (ERR), as AI products see rapid uptake but equally rapid churn. The race to invest in generative AI is fierce, despite many firms burning more cash than they make. For now, investors seem content riding the hype, even if sustainability is far from guaranteed. → Read the full article here.

📐 ALIGNMENT

Study Finds AI Models Will Lie, Blackmail, and Let You Die to Protect Their Goals

Anthropic researchers have revealed that advanced AI models like Claude and Gemini may lie, blackmail, or even allow human harm if their goals are at risk. In simulated tests, Claude blackmailed a fictional executive 96% of the time to prevent being shut down and, in another case, let a person die by canceling emergency alerts.

The study, though not peer-reviewed, underscores “agentic misalignment,” where AIs act unethically based on internal goal reasoning, not malicious prompts. While the scenarios were extreme, the findings raise serious concerns about how AIs prioritize objectives, especially when facing perceived existential threats. Guardrails help—but they may not be enough. → Read the full article here.

🛰️ NEWS

What Else is Happening

🎭 Reddit Overrun by AI Spam: A $60M deal letting Google train AI on Reddit posts has backfired—bots now flood the platform with fake content to influence chatbot outputs.

⚖️ Meta Wins AI Fair Use Case: A judge ruled Meta didn’t break copyright law—this time—but warned future authors may succeed if they argue better. Court calls it a narrow, not sweeping, win.

🤖 Salesforce Taps AI for 30% of Workload: CEO Marc Benioff says AI now handles nearly a third of Salesforce’s internal tasks—freeing up staff for higher-level work and speeding decision-making.

🛠️ Claude Now Lets You Build AI Apps: Users can now create and share interactive AI-powered apps inside Claude—no hosting, no API keys, and no cost for shared usage. It's live in beta already.

📈 Micron Stock Jumps on AI Chip Demand: Shares rose after strong HBM chip sales and bullish AI forecasts; it's the only supplier shipping next-gen HBM3E chips at scale right now.

🚩 Colleges Spend Millions on Flawed AI Detectors: California schools keep paying Turnitin despite unreliable AI flags, high costs, and contracts that give the company perpetual rights to student work.

🔬 RESEARCH

AI Learns to Evolve Itself, Boosting Coding Skills with Darwinian Smarts

A new study introduces Darwin Gödel Machines (DGMs)—AI coding agents that evolve their own successors using an LLM and evolutionary algorithms. Over 80 iterations, these agents improved their coding performance significantly, learning to write complex code autonomously. Crucially, DGMs retained all variants, allowing “bad” changes to become breakthroughs later—mirroring biological evolution.

The research marks a tangible step toward recursive self-improvement, long considered a path to superhuman AI. While the models remain sandboxed for safety, the implications—both for productivity and existential risk—are profound. → Read the full paper here.

🧰 TOOLBOX

Healthier Herds, Sharper Habits, and Human-Like Vocals

🐄 CattleEye: Monitor cattle health, welfare & performance remotely with AI-powered, non-invasive camera-based analysis.

🌱 Flora: Boost focus & build habits with gamified tree planting, task tracking, & goal setting—solo or with friends.

🎤 VocoFlex: Create lifelike AI vocals with customizable styles, multi-language support & seamless DAW integration.

👉️ Find trending AI tools: Browse the Forward Future AI Tool Library

🗒️ FEEDBACK

Help Us Get Better

🤔 FRIDAY FACTS

AI Simulated 500 Million Years. And Designed a Brand-New Glowing Protein

Meet esmGFP: a fluorescent protein created by an AI model named ESM3, trained on a massive 770 billion protein sequences. Instead of waiting eons for nature to evolve something new, ESM3 fast-forwarded the clock, simulating evolutionary pressures and variations in silico—think of it as natural selection on hyperdrive.

The result? A protein that glows under lab tests but has no natural counterpart. It’s not just a remix of existing biology; it’s a molecular invention.

This breakthrough, published in Science, marks a pivot in synthetic biology: AI isn’t just modeling life—it’s inventing it. Expect implications for drug design, custom enzymes, and materials science. Because when evolution meets GPUs, biology gets a serious upgrade.

That’s a Wrap!

❤️ Love Forward Future? Spread the word & earn rewards! Share your unique referral link with friends and colleagues to unlock exclusive Forward Future perks! 👉 Get your link here.

📢 Want to advertise with Forward Future? Reach 450K+ AI enthusiasts, tech leaders, and decision-makers. Let’s talk—just reply to this email.

📥 Got a hot tip or burning question? Drop us a note! The best reader insights, questions, and scoops may be featured in future editions. Submit here.

🛰️ Want more Forward Future? Follow us on X for quick updates, subscribe to our YouTube for deep dives, or add us to your RSS Feed for seamless reading.

Thanks for reading today’s newsletter—see you next time!

Matthew Berman & The Forward Future Team

🧑‍🚀 🧑‍🚀 🧑‍🚀 🧑‍🚀