- The Shift
- Posts
- Anthropic let Claude interview the world
Anthropic let Claude interview the world
Plus, 📹Mistral launches Forge to let enterprises train their own AI models, How to create a lip-synced video with LTX Studio, and more!
In partnership with
Welcome back to The Shift. Let’s get straight to what matters in AI today…
Today we have:
🌍 Anthropic Surveyed 81,000 Users on What They Hope and Fear About AI
📹How to Create a Lip-Synced Video With LTX Studio
🔧 Mistral Launches Forge to Let Enterprises Train Their Own AI Models
🔨Tools and Shifts you cannot miss
🌍 Anthropic Surveyed 81,000 Users on What They Hope and Fear About AI
Anthropic used Claude Interviewer to run open-ended conversations with 81,000 users across 159 countries in 70 languages. It's being called the largest qualitative AI attitudes study to date.
The Shift:
What People Want Most From AI - Professional excellence came out on top. After that, freeing up time, financial independence, and better life management were the most frequently mentioned hopes across the global sample.
What Scares People Most About AI - AI getting things wrong ranked as the number one fear, above job loss, losing personal agency, and growing over-reliance. Accuracy and trust are still the biggest barriers to comfort.
Sentiment Varies Sharply by Region - India and South America skewed above average in AI optimism. The US, Europe, Japan, and South Korea ran neutral or below. How people feel about AI depends heavily on where they are.
If you're building products or messaging around AI, this data is useful framing. People want AI to make them better at work, not replace them. And their biggest concern isn't losing jobs, it's trusting the output. That should shape how you position AI in any customer-facing context.
Together with Mintlify
AI Agents Are Reading Your Docs. Are You Ready?
Last month, 48% of visitors to documentation sites across Mintlify were AI agents—not humans.
Claude Code, Cursor, and other coding agents are becoming the actual customers reading your docs. And they read everything.
This changes what good documentation means. Humans skim and forgive gaps. Agents methodically check every endpoint, read every guide, and compare you against alternatives with zero fatigue.
Your docs aren't just helping users anymore—they're your product's first interview with the machines deciding whether to recommend you.
That means:
→ Clear schema markup so agents can parse your content
→ Real benchmarks, not marketing fluff
→ Open endpoints agents can actually test
→ Honest comparisons that emphasize strengths without hype
In the agentic world, documentation becomes 10x more important. Companies that make their products machine-understandable will win distribution through AI.
📹How to Create a Lip-Synced Video With LTX Studio
One reference image. One audio clip. A cinematic talking head that looks shot on location.
Step 1: Generate your character with Nano Banana Pro: Create your character inside LTX Studio's Gen space. Keep lighting, framing, and clothing consistent across every scene, it directly affects lip-sync quality.
Step 2: Record clean dialogue: Quiet room, close to the mic. Clean audio means accurate lip-sync.
Step 3: Split audio into clips under 9 seconds: LTX has a hard 9-second limit per generation. Break dialogue into individual sentences before uploading.
Step 4: Pair each clip with a starting frame: Select Audio to Video, upload your audio clip, upload your Nano Banana Pro image as the starting frame, describe the scene, and generate.
Step 5: Stitch and export: Combine clips back-to-back for a continuous, consistent sequence.
For cinematic b-roll between scenes, switch to LTX-2 Pro mode, same workflow without the audio track. You can try LTX Studio here.
🔧 Mistral Launches Forge to Let Enterprises Train Their Own AI Models
Mistral announced Forge at Nvidia GTC, a platform that lets companies build custom AI models trained entirely on their own data. It's aimed at enterprises that need more than generic model outputs.
The Shift:
1. Your Data Becomes the Model, Not Just Context - Most tools either fine-tune existing models or use RAG to pull company data in at query time. Forge trains entirely new models using Mistral's open-weight library, including Mistral Small 4.
2. Engineers Ship With the Platform - Forge includes forward-deployed engineers who embed with your team to build evals, surface internal data, and run synthetic data pipelines that most companies can't handle alone.
3. Big Names Are Already In - Ericsson, European Space Agency, ASML, and Singapore's DSO are early partners. Target customers: governments, financial firms, manufacturers, and tech companies needing models tuned to their own systems.
4. Enterprise-Only Focus Is Paying Off - Mistral is on track to cross $1B in annual recurring revenue this year while OpenAI and Anthropic chase consumers. CEO Arthur Mensch says the strategy is working.
If generic models with bolted-on customization still produce off outputs, training from scratch solves a fundamentally different problem. Forge makes that accessible.
Together with Hubspot
The Future of AI in Marketing. Your Shortcut to Smarter, Faster Marketing.
This guide distills 10 AI strategies from industry leaders that are transforming marketing.
Learn how HubSpot's engineering team achieved 15-20% productivity gains with AI
Learn how AI-driven emails achieved 94% higher conversion rates
Discover 7 ways to enhance your marketing strategy with AI.
🔨AI Tools for the Shift
📊 TrajectoryAI - Get weekly AI-powered insights on competitors, clients, and partners with actionable summaries and historical context.
✍️ Sydium - Train AI on your writing style and automate posting so your content goes live in your voice every day.
🎬 unbg.video - Instantly remove or change backgrounds in videos and images without needing a green screen.
🧙♂️ Loreform - Create unique, print-ready tabletop miniatures for games like DnD and Warhammer in minutes with AI.
🌍 Realms of Shod - Build and play TTRPG worlds collaboratively with AI-powered notes, multiplayer tools, and persistent session memory.
🚀Quick Shifts
💻 OpenAI is merging ChatGPT, Codex, and its Atlas browser into a single desktop "superapp" to reduce product fragmentation, as the company refocuses efforts amid growing competition from Anthropic's Claude.
⚠️ A Meta AI agent posted inaccurate technical advice publicly without approval, triggering a SEV1 security incident that gave employees unauthorized access to sensitive company and user data for two hours.
🤖 Meta is rolling out AI moderation across Facebook and Instagram, announcing plans to phase out third-party human content moderators over the next few years.
🖼️ Microsoft launched MAI-Image-2, its second-generation AI image model, bringing enhanced photorealism and improved in-image text generation, now rolling out across Copilot and Bing Image Creator.
🎨 Adobe launched Firefly Custom Models in public beta, letting creators and brands train AI image generators on their own assets to maintain consistent styles, characters, and aesthetics at scale.
🧩 Prompt of the Day
Blockchain and NFT Opportunities
Explore how blockchain can unlock transparency, ownership, and new revenue models for your brand.
Paste the prompt: Drop this into ChatGPT and fill in your brand context.
Prompt to paste
Create a blockchain and NFT strategy for [Insert brand]. Include:
Use Cases: [List ideas like product authentication, digital ownership, loyalty tokens, gated access]
Development Partners: [Insert potential collaborators like blockchain platforms, NFT marketplaces, or tech providers]
Rollout Plan: [Outline phases from concept to launch and scaling]
Use Case: Build transparency, enhance customer trust, and create new revenue streams.
That’s all for today’s edition see you tomorrow as we track down and get you all that matters in the daily AI Shift!
If you loved this edition let us know how much:
How good and useful was today's edition |
Forward it to your pal to give them a daily dose of the shift so they can 👇
Reply