Good morning, it’s Thursday. Teachers are banning AI for students—while relying on it behind the scenes. LinkedIn is now Microsoft’s AI petri dish, and Grok becomes a coding sidekick.

Plus, in the latest installment of our I Will Teach You to AI series, we break down the top 10 AI and automation skills professionals need in 2025 and beyond.

Read on!

🗞️ YOUR DAILY ROLLUP

Top Stories of the Day

📟 Grok Adds Coding and Google Drive Tools
Grok Studio now lets users generate and run code in real time, supporting Python, C++, JavaScript, and more within a new “preview” tab. The update also introduces Google Drive integration, enabling seamless access to Docs, Sheets, and Slides for editing and collaboration. Users can co-create content like reports, games, and documents in a collaborative studio window. All features are now available to all free and premium users.

🤖 OpenAI Launches Smarter Models o3 and o4-mini
OpenAI has introduced o3 and o4-mini, its most advanced models yet, with powerful capabilities in coding, math, science, and visual reasoning. These models can now integrate tools like web search, image generation, and code execution directly into their reasoning processes. o4-mini offers high performance with lower latency and cost, ideal for heavy use and scale. ChatGPT users on Plus, Pro, and Team plans can access them starting today.

🚫 U.S. Considers Blocking DeepSeek from U.S. Tech Access
The Trump administration is weighing penalties to stop China’s DeepSeek from acquiring American technology, including potentially restricting access to its AI services. DeepSeek’s rise, fueled by NVIDIA chips, has alarmed U.S. officials aiming to curb China's AI and military advances. NVIDIA now faces scrutiny and a possible $5.5 billion revenue hit from tightened export rules. A congressional panel is investigating how its chips reached DeepSeek despite existing bans.

🦾 Microsoft Debuts Ultra-Efficient BitNet AI for CPUs
Microsoft has released BitNet b1.58 2B4T, a hyper-efficient 1-bit AI model that runs on CPUs like Apple’s M2, bypassing the need for GPUs. With 2 billion parameters trained on 4 trillion tokens, it outperforms or matches top models from Meta and Google in benchmarks like GSM8K and PIQA. It also runs twice as fast with less memory—but only when using Microsoft’s custom bitnet.cpp framework, which limits hardware compatibility for now.

Enjoying what you’re reading? Share it with a colleague who’d
appreciate it too. Every forward helps.

🧑‍🚀🧑‍🚀🧑‍🚀🧑‍🚀
—Matt and the Team

🎓 EDUCATION

Teachers Embrace A.I. Tools—While Restricting Student Use Over Ethical Concerns

The Recap: Educators across the U.S. are increasingly using generative A.I. to streamline tasks like grading, lesson planning, and data analysis, even as many restrict students from using similar tools to complete assignments. This dual standard is prompting ethical debates over fairness, transparency, and the future role of A.I. in education. Dana Goldstein of The New York Times explores how school leaders and teachers are grappling with these tensions amid aggressive marketing from edtech firms.

Highlights:

• Teachers are using A.I. for lesson planning, grading, and summarizing materials, while often prohibiting students from using it for writing or problem-solving.

• Middle schoolers are widely using math apps like Google-owned PhotoMath to complete homework—sparking concerns over cheating.

• Some educators, like Jon Gold in Providence, advocate for teaching A.I. literacy and ethical use, not just policing student behavior.

• In Texas, automated scoring of high-stakes student writing led to 2,000 grade reversals after manual review—raising questions about fairness and accuracy.

• The Dallas school district both criticized state test scoring algorithms and embraced A.I. for AP essay grading and internal administrative tasks.

• Venture capital investment in A.I.-powered edtech has surged to $1.5 billion in two years, with major players like Google and Microsoft promoting A.I. tutors and planning tools.

Forward Future Takeaways:
This story captures a central paradox in A.I.'s rapid adoption: it's welcomed as a productivity booster for educators but restricted as a threat to academic integrity when used by students. As generative A.I. reshapes classroom norms, schools must confront deeper questions about what constitutes learning, fairness, and the human role in education. → Read the full article here.

🧑‍🏫 FORWARD FUTURE PRO

10 AI & Automation Skills to Know in 2025

The workplace is transforming faster than you can say retrieval-augmented generation. According to the World Economic Forum, by 2025, automation may displace 85 million jobs while simultaneously creating 97 million new roles to meet rising tech-driven demands. This isn't just a shift for technical specialists – it's reshaping how professionals across all business functions work.

To thrive in this new landscape, you need practical AI and automation skills. This guide breaks down 10 skills that will help you navigate workplaces of the future, with concrete actions you can take today to start building these capabilities. → Premium subscribers: get the full guide here

💼 LINKEDIN

LinkedIn Becomes Microsoft’s AI Petri Dish — and a Business Media Powerhouse

The Recap: LinkedIn is evolving from a professional networking site into a dynamic content platform and a crucial testing ground for Microsoft’s AI ambitions. While user engagement still lags behind entertainment-driven platforms, LinkedIn is leveraging its unique audience and data to shape business-to-business advertising and AI product development. As Microsoft integrates LinkedIn more deeply into its ecosystem, the platform’s role in the AI race is becoming surprisingly strategic.

Highlights:

• Since Microsoft’s $26 billion acquisition in 2016, LinkedIn’s annual revenue has grown from $3 billion to $17 billion.

• The platform now makes the majority of its money from recruitment and subscriptions, with advertising ($7 billion/year) as a growing but secondary stream.

• Despite 1.1 billion sign-ups, LinkedIn sees relatively low engagement—Android users spend just 48 minutes/month on it versus 35 hours on TikTok (Sensor Tower).

• LinkedIn’s algorithm now favors topic-based content over network-based posts to boost time on site and appeal to advertisers.

• Video uploads and comment activity have each risen by over 35% year-on-year, driven partly by exoduses from Facebook and X.

• AI tools like writing assistants, job matchers, and training simulations are being piloted on LinkedIn before broader Microsoft rollout.

• Content moderation has become a heavier lift: LinkedIn removed nearly 500,000 posts in early 2024, up from 56,000 in 2020.

Forward Future Takeaways:
LinkedIn’s transformation into both a media outlet and AI lab underscores a broader shift in how data-rich, niche platforms are being weaponized in the AI race. Its integration with Microsoft’s enterprise tools and data strategy gives the tech giant a competitive edge that goes beyond OpenAI. But balancing user trust with AI experimentation will be critical—especially as personalization starts to feel more like surveillance than service. → Read the full article here.

🛰️ NEWS

What Else is Happening

🐪 Google Unveils CaMeL Security Fix: New AI framework splits tasks to block prompt injection—treats language models as untrusted by design.

⚠️ AI Scams Get Smarter: Microsoft blocked $4B in AI-enhanced fraud—new report reveals fake job posts, scam sites, and vishing (voice phishing) attacks on the rise.

💰 Hammerspace Raises $100M for AI Data Ops: Used by NVIDIA and Meta, the startup helps wrangle unstructured data—now valued at over $500M.

🖼️ ChatGPT Adds Image Library: New “Library” tab lets users easily browse and manage their AI-generated images—now live for Free, Plus, and Pro tiers.

🧬 Profluent Finds Scaling Laws in Biotech AI: Its ProGen3 model designs novel proteins, hinting bigger models mean better bio-tools—just like with ChatGPT.

👽 AI Hunts for Alien Earths: A new algorithm trained on 50K+ simulated systems flagged 44 real stars likely hiding Earth-like planets in habitable zones.

🔬 RESEARCH

AI Glasses Help the Blind Navigate City Streets—And Spot Obstacles Before They Arrive

A new prototype wearable uses AI-powered cameras and haptic sensors to help visually impaired users navigate their surroundings with greater speed and accuracy than traditional canes. Developed by researchers in China and tested by 20 blind participants, the system offers real-time audio and vibration cues to guide users through indoor mazes and cluttered urban settings.

Unlike canes, the system can identify objects beyond arm’s reach—recognizing doors, people, and furniture up to 40 cm away. The tech is still in development, but its potential as a "very intelligent stick" could reshape mobility for millions. → Read the full story here.

📽️ VIDEO

The Fastest "Computer Control" Agent We've Ever Seen

In case you missed it, Matt explores Ace by General Agent—an ultra-fast AI that controls your desktop and browser. Trained on behavior, it handles complex tasks faster than most humans with high accuracy. Get the full scoop! 👇

🧰 TOOLBOX

Intelligent Support, Instant Summaries, and Effortless Form Building

🤝 AG2 AI: Boosts customer service with smart replies, info retrieval, and task automation, learning from chats to improve over time.

🗐 AI PDF Summarizer: Instantly distills lengthy PDFs into concise summaries, perfect for students, researchers, and busy professionals.

📋 MakeForms: Simplifies form creation with AI, enabling doctors to build secure, HIPAA-compliant forms quickly for patient intake and data collection.

🧑‍🚀 Want to explore more AI tools?
Browse the Forward Future AI Tool Library

🤠 THE DAILY BYTE

Number of Notable AI Models by Geo

There are a bunch of ways to measure which country is “ahead” in AI — published papers, citations, patents, all that. But one of the simplest signals is: who’s shipping the models that actually move the needle? Epoch AI tracks the most influential models going back to 1950, and that’s where the AI Index pulled the data for this chart. → Read the full article here.

That’s a Wrap!

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

🛰️ 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!

The Forward Future Team

🧑‍🚀 🧑‍🚀 🧑‍🚀 🧑‍🚀