AI just rolled out three major upgrades redefining creativity, productivity, and autonomous workflows. From native image generation to max-level agents and inbox-running AI, here’s what’s new:

🎨 ChatGPT Images — Native Image Generation Gets Serious
OpenAI brings GPT-Image-1.5 directly into ChatGPT with 4× faster rendering, precise text handling, conversational image edits, preset styles, and ethical C2PA metadata, turning ChatGPT into a true text-to-visual creative engine.

⚡ Manus 1.6 Max — Autonomous Agent for Real Work
Manus 1.6 introduces Max, a high-performance agent excelling in spreadsheets, web dev, and deep research, plus full mobile app development and pro-grade image editing, built to run complex workflows with minimal supervision.

📅 Google CC — Gmail & Calendar Get an AI Brain
Google’s CC agent unifies Gmail, Calendar, Drive, and the web to deliver proactive daily briefings, auto-draft emails, create events, and manage tasks, turning your inbox into an intelligent personal operator.

AI isn’t just assisting anymore, it’s creating, organizing, and executing end-to-end.

Google Launches Gemini Deep Research Agent

OpenAI unveiled "ChatGPT Images" on December 16, 2025, revolutionizing its chatbot with native image generation powered by the cutting-edge GPT-Image-1.5 model. Accessible via a sleek Images tab in the ChatGPT app and web interface, this upgrade delivers 4x faster rendering than predecessors, excelling in precise text on signs, infographics, and labels—rivaling Google's Nano Banana amid intensifying AI wars. Users enjoy conversational editing: inpaint details, tweak attributes, or remix photos while preserving intricate elements like textures and lighting. Preset styles spark creativity, a discovery feed curates inspiration from global prompts, and ethical C2PA metadata ensures transparency. Free, Plus, Pro, and Team tiers worldwide gain unlimited access (with DALL-E 3-style limits), plus API integration for developers. Responding to Gemini 3's multimodal push, OpenAI promises rapid iterations in advanced edits, multilingual prompts, and hyper-realistic outputs. This leap cements ChatGPT as a creative powerhouse, blending text and visuals seamlessly for designers, marketers, and everyday innovators—unleashing boundless imagination in seconds.

Manus 1.6: Max Agent, Mobile Dev, Pro Editing

Manus AI launched version 1.6 on December 15, 2025, featuring flagship agent Manus 1.6 Max for superior autonomy and performance. Max boosts one-shot task success, cuts supervision needs, and lifts user satisfaction by 19.2% in blind tests, excelling in spreadsheets (financial modeling, data analysis, reports), web dev (polished UIs, invoice parsing, feature matrices), and Wide Research with parallel Max sub-agents.​ New Mobile Development enables end-to-end iOS/Android app creation from prompts, expanding beyond web projects. Design View offers interactive image editing: point-and-click precision, in-image text changes, and compositing for pro-level designs powered by advanced models.​ Available now to all users; Max at 50% credit discount temporarily. This upgrade targets complex workflows, delivering reliable intelligence for pros.

Gmail + Calendar AI magic: Meet Google CC

Google Labs unveiled CC, an experimental AI agent powered by Gemini, on December 17, 2025, to supercharge daily productivity. Seamlessly integrating Gmail, Calendar, Drive, and web sources, CC delivers a proactive "Your Day Ahead" briefing—synthesizing schedules, tasks, emails, and updates into concise, actionable insights like upcoming bills, appointment prep, or priority todos. No manual queries needed; it anticipates needs from your data. Key features include auto-generated email drafts, calendar invites, and summaries for rapid execution. Users interact via email replies or custom messages to refine: teach preferences ("remind me gym Tuesdays"), log ideas, or add tasks. Privacy-focused, it processes data securely with user controls. Early access targets U.S./Canada Google AI Ultra and paid subscribers (18+); join the waitlist at labs.google. CC positions Google against rivals like Manus Max in agentic AI, emphasizing effortless organization for pros amid inbox overload.

Hand Picked Video

In this video, we’ll look at Infloq, an Influencer Marketing Operating System (OS) built for brands, influencer marketing agencies, media managers, social media managers, creators, and content creators.

Top AI Products from this week

• Claims AI - ContraVault Claims AI is a claims-specific feature that drafts claim & response letters automatically, built for long-running claims. Add the email thread (or connect inbox) to keep years of emails, letters, and attachments organized in one place.

• Grov - Most AI coding tools are single-player. Your AI forgets everything when the session ends, forcing every dev to waste time on redundant exploration. Grov turns your AI coding agent into a collective team brain. It captures the reasoning behind every solution and syncs it to your entire team.

• Everything AI Tool - Everything AI Tool is built to dominate AI discovery. With 25,000+ curated AI tools and 30K+ unique weekly visitors, we give AI products real distribution, not launch-day spikes.

• CanvAi - CanvAi is a creative workspace that turns rough sketches into polished AI art. Draw on an infinite canvas, add shapes and assets, then use frames to transform them with Nano Banana Pro.

• Bilt.me - Build and launch production-ready native mobile apps with AI. Prototype, collaborate, and deploy faster — no coding required. Build from scratch or auto deploy Web Apps into Mobile Apps. By making mobile app creation as quick and frictionless as writing a thought, we aim to unlock a wave of innovation where anyone can turn fresh ideas into working native apps the moment inspiration strikes.

• Rebrew - You decide. You refine. Then AI builds. Drop a raw idea. See 3 directions with real tradeoffs. Pick one. That choice shapes everything. Your brew is a canvas — a visual prompt you shape before AI builds anything. Rearrange it. Ask AI to review it. When it's ready, one click turns it into a landing page, pitch deck, or content plan. No more generic output.

This week in AI

• Chatterbox TTS - Resemble AI's Chatterbox: MIT-licensed open-source TTS. Multilingual, emotion control, zero-shot cloning from 5s audio. Beats ElevenLabs in blind tests (63.75% preference). Real-time, watermarked.

• Dan Fu AGI Notes - Dan Fu argues AGI feasible computationally AI underutilized, hardware lags models, near "useful AGI". Explores paths beyond scaling limits. (Stanford researcher's perspective on progress).

• TechForce.gov America’s Tech Talent Hub - TechForce.gov connects Americans to AI, cyber, and advanced tech careers, offering training, upskilling, and pathways into high-impact federal and national security roles.

• Olmo 3.1 Launch - AI2 releases Olmo 3.1 Think 32B & Instruct 32B—strongest RL scaling to 32B params. Most capable open models yet for reasoning and instruction tasks.

• EgoX AI Video Gen - EgoX converts single exocentric (3rd-person) videos to egocentric (1st-person) views. Uses LoRA-adapted diffusion models, unified priors, geometry-guided attention for coherent synthesis. SOTA on benchmarks.

Paper of The Day

ACoRN proposes a robust training recipe for RAG compressors that keeps answers intact even when retrieved passages are noisy or partially irrelevant. It generates synthetic noisy contexts offline, then fine-tunes compressors to focus on answer-supporting spans, reducing positional bias toward earlier passages and mitigating hallucinations. Experiments on multiple QA and RAG benchmarks show better end-task accuracy, stronger robustness to retrieval noise, and improved faithfulness of compressed contexts compared with standard abstractive compression baselines.

To read the whole paper 👉️ here