Greetings Qrew,

Whether you're diving into AI for work, side projects, or just out of curiosity — this issue of The Creators Hive brings you practical insights, hands-on learning, and tools you can actually use.

1. Thinking Machine Labs Raises $2B, Nears First Product Reveal
2. Runway’s Act-Two for AI Motion Capture
3. AI Researchers Unite on Reasoning Transparency
4. LTXV Unlocks 60-Second AI Videos
5. New ChatGPT Agents for Excel, PowerPoint
6. Tool Recommendation of the Day: MapScroll

---

Thinking Machine Labs Raises $2B, Nears First Product Reveal

Former OpenAI CTO Mira Murati announced a massive $2B seed round for her stealth AI startup, Thinking Machine Labs (TML) — now valued at $12B — and confirmed that its first product will launch within months.

Key Details:

• Record Seed Round: The $2B raise comes less than a year after TML’s creation, despite no public product or clear direction.
• Open-Source Component: Murati revealed the first release will include “a major open-source component” designed for researchers and startups building custom AI models.
• Multimodal Focus: TML is developing multimodal AI systems capable of conversational and visual collaboration with users.
• Business Tools: Reports suggest TML is working on custom AI solutions aimed at driving enterprise profitability.

Why It Matters:​
Mira Murati’s track record as a former OpenAI executive and chief architect behind models like GPT-4 has investors betting big on TML — even without a product. In an industry dominated by OpenAI, Google, and Anthropic, TML’s upcoming launch could reshape the competitive landscape if its open-source and multimodal promises deliver.

---

Runway’s Act-Two for AI Motion Capture

Runway has unveiled Act-Two, its next-generation motion capture model that transforms single performance videos into fully animated characters — capturing head, face, body, and hand movements while preserving artistic styles and backgrounds.

Key Details:

• Comprehensive Tracking: Captures subtle facial expressions, hand gestures, and full-body movements from just one driving performance video.
• Single Reference Photo: Users only need a single character reference image for the model to animate the performance with consistent art styles and preserved backgrounds.
• Major Upgrade: Runway says Act-Two brings significant improvements in fidelity, movement accuracy, and consistency over the October 2024 Act-One release.
• Hollywood Partnerships: The company is working with Lionsgate, AMC Networks, and other studios to integrate AI into film production pipelines.

Why It Matters:​
Despite public skepticism about AI in Hollywood, behind-the-scenes adoption is accelerating. Act-Two’s advancements make it easier to create high-quality, stylized character animation — potentially cutting costs and production time dramatically. This could reshape the VFX and animation industry, whether or not creatives are ready for it.

---

AI Researchers Unite on Reasoning Transparency

Top AI researchers from OpenAI, DeepMind, Anthropic, and other leading institutions have co-authored a paper advocating for greater transparency into the "chains of thought" (CoT) of AI reasoning models, emphasizing that this could be a critical safety measure for the future of AI systems.

Key Details:

• Chains of Thought (CoT): These step-by-step reasoning traces provide a unique window into how models solve problems, making them valuable for auditing and debugging AI decision-making.
• Transparency Concerns: The paper warns that as AI architectures evolve and training techniques shift, CoT visibility could diminish, making models harder to monitor and control.
• Industry Signatories: The paper is backed by OpenAI’s Mark Chen, SSI’s Ilya Sutskever, Geoffrey Hinton (Nobel laureate), and DeepMind co-founder Shane Legg, among others.
• Proposed Actions: Researchers suggest standardizing evaluations for monitorability and incorporating these metrics into the deployment readiness of future frontier models.

Why It Matters:​
This rare moment of industry-wide agreement signals a turning point in AI safety. As models grow more powerful and opaque, maintaining insight into their reasoning process could be key to avoiding catastrophic errors or misuse — but competitive pressures may soon make this transparency harder to preserve.

---

LTXV Unlocks 60-Second AI Videos

Lightricks has upgraded its open-weights LTXV model to support real-time image-to-video generation for clips over 60 seconds, with live prompt control and performance optimized for consumer GPUs.

Key Details:

• Instant Streaming: The model starts streaming the first second instantly, building video continuously without cuts.
• Live Adjustments: Users can modify poses, depth, and style mid-generation, allowing for dynamic, evolving scenes.
• Licensed Data & Integration: Trained entirely on fully licensed datasets, LTXV integrates directly with LTX Studio’s production suite.
• Open & Lightweight: Available free on GitHub and Hugging Face, with both 13B and 2B (mobile-friendly) parameter versions.

Why It Matters:​
With most AI video models limited to a few seconds, LTXV’s 60-second real-time streaming is a game-changer, transforming how creators direct and produce AI-generated video content.

---

New ChatGPT Agents for Excel, PowerPoint

OpenAI is developing ChatGPT agents that can generate and edit spreadsheets and presentations directly in chat, potentially reducing reliance on Microsoft Office and Google Workspace, per The Information.

Key Details:

• Integrated Tools: ChatGPT will add dedicated buttons below the search bar to create spreadsheets and slides using natural language prompts.
• Cross-Compatibility: Outputs will use Microsoft’s open-source formats, ensuring compatibility with major office apps.
• Early Tests: Performance has been described as “slow and buggy”, with single tasks taking up to 30 minutes.
• Collaboration Feature: OpenAI is also testing a multi-user collaboration tool, but no release details are available.

Why It Matters:​
If ChatGPT can reliably create business-ready sheets and presentations, it could challenge Microsoft and Google’s productivity dominance. But speed and reliability will be critical for mass adoption.

---

Tool Recommendation of the Day: MapScroll

🗺️ Website:https://www.mapscroll.ai​

What makes it unique?

• MapScroll isn’t just a map app — it’s like having an AI-powered storytelling copilot for maps, transforming text prompts into immersive, interactive story maps.
• You simply type ideas like “Mark James Bond movie locations” or “trace Marco Polo’s route,” and MapScroll auto-geocodes pins, fetches images & articles, and builds a shareable narrative map
• It integrates with reliable mapping backends like Mapbox and OpenStreetMap and supports live event tracking & itineraries.
• Provides context-rich pop-ups at each location—complete with media, sources, and descriptions—making exploration educational and engaging .

Special twist:

MapScroll blurs the line between map browsing and storytelling. Rather than static points, it builds dynamic, flowing storylines — perfect for history trails, pop culture pilgrimages, travel planning, or investigative journeys.

Use case magic:

✅ Plot historical routes, like Marco Polo’s journey, with rich context
✅ Discover hidden gems—rooftop cafés, active volcanoes, film locations—via natural language
✅ Create shareable story maps for teaching, travel blogs, or presentations
✅ Stay updated on real-time events or trending locations
✅ Save research time by auto-collecting images, links, and summaries

Why it’s rare:

Conventional map tools focus on directions or points of interest. MapScroll understands open-ended queries and builds story-rich, interactive maps on the fly, combining discovery, context, and narrative in a seamless workflow.

⏳ User feedback:

“It’s like Google Maps on steroids” (Arun bhatia, Product Hunt)Product Hunt​

“Such brilliant product… would be playing with this in the next couple of days.” (padmanabh, Product Hunt)Product Hunt​

🌟 Final take

MapScroll is ideal if you want to turn curiosity into compelling, shareable maps — without manual pinning or research. Whether for education, travel planning, storytelling, or cultural exploration, this tool is a game-changer for map-centric narratives.

​

---

Masterpieces from the Hive 🎨

Here’s what our community has been creating this week using the Hive Wall:

32 North Gould Street, Sheridan, Wyoming 82801
​Unsubscribe · Preferences​