✨ New ChatGPT Images Launches
OpenAI released a new ChatGPT Images experience powered by its flagship image model. The update delivers more precise edits that preserve details like lighting, composition, and likeness, with generation speeds up to four times faster. It supports advanced editing, creative transformations, and improved text rendering. The model is rolling out to all users as GPT-Image-1.5.

🥞 DoorDash Debuts Zesty, an AI App for Restaurant Discovery
DoorDash has launched Zesty, an AI-powered app that helps users find local restaurants through conversational prompts and social sharing. Now live in San Francisco and New York, Zesty uses a chatbot to recommend spots based on user preferences and context. It pulls data from DoorDash, Google Maps, and TikTok, adding social features like saves and follows.

😎 Meta AI Glasses Get Smarter
Meta rolled out a v21 software update for its AI glasses, adding conversation focus and Spotify integration. The update lets users amplify a speaker’s voice in noisy settings to better hear conversations. A new multimodal feature allows Meta AI to play Spotify music that matches what the wearer is looking at. The update is rolling out gradually, starting with Early Access users.

📑 Disney–OpenAI Deal Limits Exclusivity
Disney’s licensing deal with OpenAI runs for three years but is exclusive for only the first year. During that time, OpenAI’s Sora can legally use over 200 characters from Disney, Marvel, Pixar, and Star Wars. After year one, Disney can strike similar deals with other AI companies. CEO Bob Iger framed the move as testing generative AI while embracing disruption.

Vultr is empowering the next generation of generative AI startups with access to the latest AMD and NVIDIA GPUs.

Try it yourself and use promo code "BERMAN300" for $300 off your first 30 days.

Learn reinforcement learning at home with a step-by-step tutorial using NVIDIA RTX and Unsloth to train a local AI from scratch to master the game 2048 using verifiable rewards.

Zoom Video Communications said its AI system scored a record 48.1% on Humanity’s Last Exam (HLE), surpassing Google’s Gemini 3 Pro at 45.8%, according to a December 2025 announcement by CTO Xuedong Huang. The result drew attention because Zoom did not train a new large language model (LLM); instead, it orchestrated outputs from multiple existing models using a proprietary “federated AI” framework.

Critics argue Zoom is taking credit for the work of OpenAI, Google, and Anthropic, while supporters counter that ensemble approaches are a proven way to advance performance. The debate highlights a broader shift in AI: whether progress should be measured by standalone models or by systems that combine them effectively. → Read the full article here.

OpenAI introduced FrontierScience, a new benchmark designed to evaluate AI systems on expert-level scientific reasoning across physics, chemistry, and biology, according to a December 16, 2025 publication. The benchmark includes two tracks—Olympiad and Research—aimed at measuring both constrained problem-solving and open-ended research tasks written and verified by domain experts.

In initial results, GPT-5.2 topped both tracks, scoring 77% on Olympiad questions and 25% on Research tasks, outperforming other frontier models. OpenAI positions FrontierScience as a response to saturation in earlier benchmarks and as a tool to better forecast how AI might accelerate real scientific workflows. → Read the full article here.

Lightspeed Venture Partners announced it raised $9 billion in new capital, the largest fundraise in the firm’s 25-year history, as limited partners concentrate money in established venture firms. The raise comes after weaker returns from the 2021 VC boom pushed endowments, pension funds, and sovereign wealth funds to favor managers with proven exits.

Lightspeed has benefited from recent IPOs by portfolio companies including Rubrik, Netskope, and Navan, despite a sluggish public markets environment. The firm is also doubling down on AI, positioning itself to make large, capital-intensive bets in the sector. → Read the full article here.

🤔 Oracle Reveals $248B AI Leases (Paywall): New filings show massive off-balance-sheet data-center commitments through 2028, nearly $150B more than previously disclosed, underscoring the scale—and risk—of its AI push.

🚪 OpenAI Comms Chief Departs: Hannah Wong told staff she will leave the company in January. Wong joined in 2021 and helped guide OpenAI’s public image through ChatGPT’s rapid growth and the 2023 leadership crisis.

🤪 “Slop” Is 2025’s Word: Merriam-Webster crowned “slop” for capturing the flood of low-quality AI content, reflecting a year defined less by fear of AI than by mockery of it.

📬 Google Tests Email AI Assistant: New Labs experiment “CC,” powered by Gemini, sends daily email briefs from Gmail, Drive, and Calendar, aiming to automate planning for paid AI users.

🎬 Adobe Firefly Levels Up Video AI: New prompt-based editor lets users tweak scenes, colors, and camera angles, while added third-party models expand image and video generation options.

📣 Anti-AI Marketing Gathers Steam: As AI “slop” floods media, brands and creators signal “100% human” content, betting consumer trust and authenticity will matter more in 2026.

Forward Future Jobs is where the AI-fluent find their next big opportunity. Our mission: cut through the noise of generic job boards and connect talented professionals with roles that value AI fluency, even when it isn’t explicitly required.

❤️ Love Forward Future? Share your unique referral link with friends and colleagues to unlock exclusive Forward Future perks! 👉 Get your link here.

Thanks for reading today’s newsletter—see you next time!
Matthew Berman & The Forward Future Team

🧑‍🚀 🧑‍🚀 🧑‍🚀 🧑‍🚀