🎬 Runway Unveils Gen-4.5 Video Model
Runway released Gen-4.5, a major upgrade to its video generation system, delivering higher fidelity, stronger motion accuracy and better prompt adherence. The model tops the Artificial Analysis benchmark with 1,247 Elo points. It retains Gen-4’s speed while improving control modes and stylistic range. Built and optimized on NVIDIA GPUs, Gen-4.5 will roll out to all users in the coming days.
🧑💼 Apple Names Former Google Exec to Succeed Retiring AI Chief
Apple announced that John Giannandrea, Senior VP of Machine Learning and AI Strategy, will step down and retire in spring 2026. Amar Subramanya, a former AI leader at Microsoft and Google, will replace him as VP of AI, reporting to Software Chief Craig Federighi. Subramanya will oversee Apple’s foundation models, ML research, and AI safety as the company moves to close the gap with rivals in generative AI.
🚀 NVIDIA Unveils Open‑Source “Digital and Physical AI” Stack
NVIDIA unveiled a broad set of open‑source AI models, datasets, and tools spanning “digital” and “physical” AI. The centerpiece: NVIDIA DRIVE Alpamayo‑R1 (AR1), the first industry‑scale open reasoning vision‑language‑action (VLA) model for autonomous vehicles — designed to reason through complex driving scenarios using “chain‑of‑thought” style planning.
🚀 DeepSeek Debuts Free GPT-5-Class Models
DeepSeek released two MIT-licensed models said to rival GPT-5 and Gemini-3.0-Pro. The Speciale variant earned gold-level results in top math and coding contests. A new sparse-attention design cuts long-context inference costs by roughly 70%. With full weights freely available, the launch pressures U.S. AI leaders. Regulatory hurdles in Europe and the U.S. remain significant.
Enjoying our newsletter? Forward it to a colleague—
it’s one of the best ways to support us.
🪧 POWERED BY MULTIVERSE COMPUTING
Cut Compute Costs By 50% With CompactifAI
CompactifAI by Multiverse Computing compresses LLMs by up to 80% while maintaining accuracy, delivering up to 50% cost and energy savings for companies like IBM, Bank of Canada, and Moody’s. Run your models on fewer GPUs or even on the edge when you deploy CompactifAI to scale your AI workloads!
Try CompactifAI for Free and Realize Immediate Savings!
📺️ FROM THE LIVE SHOW
How a 20-Year Bet on CUDA Quietly Turned NVIDIA Into an AI Superpower
NVIDIA doesn’t sell chips, it sells time.Bryan Catanzaro (@ctnzr), VP of Applied Deep Learning Research at @NVIDIAAI, on why the company’s culture has stayed the same even as it became a multi-trillion-dollar giant.“CUDA took 20 years to matter, but we kept investing because… pic.twitter.com/27H8KQ8khe
@ForwardFutureView post on X👨👦 PARENTING
An AI Pioneer Explains How He’s Preparing His Four-Year-Old for a World Filled With Algorithms
The Recap: AI developer Soulaiman “Soul” Itani wrote about raising his four-year-old son, Omie, amid increasingly lifelike AI systems — including the digital actors he builds himself. In this opinion essay, Itani describes balancing parental concerns about large language models with an educational philosophy that emphasizes critical thinking, creative problem-solving, and real-world judgment. His core argument: parents must help shape both children’s habits and broader AI regulation to ensure kids grow up valuing human creativity while using AI responsibly.
Highlights:
Soulaiman ‘Soul’ Itani — CTO of Incantor AI and a former lead engineer on Google’s LaMDA — recounts his four-year-old mistaking an AI avatar for a real person.
He cites survey data indicating about 61% of parents worry AI will undermine children’s critical-thinking skills and warns that instant AI answers can short-circuit real problem-solving.
Itani plans to allow his son to use tools like ChatGPT eventually, but only with guidance on fact-checking and understanding how models can “learn” incorrect information.
He urges parents to push for mandatory parental controls in AI systems and stronger oversight of how children’s data is used in model training.
Forward Future Takeaways:
Itani’s essay captures a tension many families and educators now face: AI will be unavoidable in children’s lives, but thoughtful use — grounded in curiosity, verification, and human judgment — is still teachable. His call for parental involvement in AI governance signals a shift toward embedding family-centric considerations into product design and regulation. → Read the full article here. (Paywall)
💼 WORK
Why Coders and Call-Center Agents Are First in AI’s Firing Line
A growing share of employers are attributing layoffs to artificial-intelligence adoption, even though evidence of broad labor-market impact remains limited. New data shows AI uptake is already strongest in two fields: computer programming, where tools like GitHub Copilot now reach 26 million users, and customer service, where 85% of managers plan AI pilots in 2025.
These roles share traits that make them ripe for automation — their tasks are repetitive, low-context, easily verified, and backed by abundant training data such as code repositories and call-transcript archives. Large workforces and clear cost savings further attract both startups and big firms to invest. As AI becomes cheaper and more specialized, adjacent white-collar roles such as junior banking and legal work may face similar pressure. → Read the full article here. (Paywall)
🗳️ POLITICS
AI-driven Project Targets 2026 House Races for Independent Breakthroughs
The Independent Center, a nonprofit focused on independent voters, is using an AI system to pinpoint 2026 House districts where non-party candidates could realistically win. Led by veteran political strategists Adam Brandon and Brett Loyd, the group has identified 40 competitive districts and aims to field about 10 candidates by spring 2026.
Their AI analyzes real-time online sentiment, turnout patterns, and demographic shifts — especially among younger voters — to locate both viable districts and “diamond in the rough” candidates, sometimes by scanning digital footprints such as LinkedIn activity. Even a handful of independent victories could deny either party a House majority for the first time in decades, reshaping how Congress operates. → Read the full article here.
🛰️ NEWS
What Else is Happening
💰 Data-Center Boom Boosts Pay: Construction workers shifting into AI-driven data-center projects are seeing 25–30% raises, with some earning over $200K amid a 439,000-worker shortage.
⚡ Data Centers’ Power Use Set To Triple: Data center electricity demand is forecasted to reach 106 gigawatts by 2035, up from 40 gigawatts today. New facilities are far larger, with nearly a quarter expected to exceed 500 megawatts.
📢 Code Hints At ChatGPT Ads Incoming: Newly discovered code in a beta Android build suggests OpenAI is preparing to add ads to ChatGPT. References include an ads feature, search ads, and a carousel.
🫧 AI Bubble’s Real Risk: A Guardian analysis warns that debt-driven AI investments may mirror past bubbles, risking financial strain if returns lag while core tech value remains uncertain.
🛒 OpenAI Takes Stake in Thrive: OpenAI buys into Thrive Holdings to speed enterprise AI adoption and expands its reach with an Accenture deal rolling out ChatGPT to tens of thousands.
𝕏 TWEETS
You are 200x more likely to get into Harvard University than you are get a job at ElevenLabs. 180,000 people applied for jobs at ElevenLabs in the first 6 months of 2025. Just 0.018% of people who apply, get a job. Insane 🤯
@HarryStebbingsView post on X🧰 TOOLBOX
Trending AI Tools
🗣️ Murf.ai: Create lifelike AI voiceovers with customizable tones, accents, and multilingual support.
⚙️ Flowith: Automate workflows with AI-driven tools, customizable templates, and real-time analytics.
📚 Galactica: Meta’s open LLM for research, offering model access, scientific focus, and community-driven evaluation.
🗒️ FEEDBACK
Help Us Get Better
What did you think of today's newsletter? |
That’s a Wrap!
📢 Want to advertise with Forward Future? Reach 600K+ AI enthusiasts, tech leaders, and decision-makers. Let’s talk—just reply to this email.
Thanks for reading today’s newsletter—see you next time!
Matthew Berman & The Forward Future Team
🧑🚀 🧑🚀 🧑🚀 🧑🚀
