Welcome, humans.

Want to make better AI videos but have absolutely no idea what you’re doing? Perfect. Neither do we TBH. 😹

That’s why later today at 1pm ET, 10 AM PT, we’re going LIVE with the team at LTX for a TOTAL beginner’s guide to video prompting using the all new LTX-2.5. We’ll cover how video prompts differ from image prompts, then show you how to control shots, motion, subjects, and scenes. And of course, we’ll put LTX-2.5 through a live demo + Q&A. Bring your weirdest prompts.

Speaking of the power of prompts, apparently one Claude user just gave Opus 5 access to Unreal Engine for 24 hours and told it to build GTA 6 (he used AAABench for it). The result is more what I'd called “GTA 4.5” than “GTA 6,” but waaaaaaay better than anything you'd imagine AI could create entirely on its own a year ago.

Here’s what happened in AI today:

• 😺 Grok 4.6 launched for long-running AI agents.

• 📰 Lovable raised $400M at a $13.3B valuation.

• 📰 Hot tea on why Demis DIDN’T leave Google.

• 🍪 Preview is a full workspace for team video generation.

• 🎓 Turn Claude Voice into a morning briefing.

... and a whole lot more that you can read about here.

😺 Grok 4.6 Is Here, and it’s GPT 5.6 Level?!

Nuff said in the headline, am I right?

Here’s what happened:

• SpaceXAI released Grok 4.6 for long-running agents, coding, research, and more ambitious interactive work.

• Artificial Analysis scored it 61, up five points from Grok 4.5 and level with GPT-5.6 Sol overall.

• It is available now in Cursor, Grok Build, the API, OpenRouter, Vercel, and Cloudflare, with 2x included usage in Cursor and Grok Build for the first week.

AI agents get expensive fast. A model can look brilliant in one answer and still be a bad deal if a multi-hour task takes twice as many turns, tokens, and retries. Grok 4.6's pitch is that it can stay near the frontier without the frontier bill.

On price, xAI’s current pricing page includes Grok 4.6 on the $30/month SuperGrok plan, while API pricing starts at $2 per million input tokens and $6 per million output tokens. Artificial Analysis (linked above) measured it at $0.84 per task and placed it on its cost-performance frontier across every agentic evaluation in the index.

The efficiency matters A LOT on long jobs. On AA-Briefcase, a Grok reached Fable 5-tier while averaging about 53 turns and 0.5B input tokens, versus roughly 103 turns and 2.0B for Claude Opus 5 Max. Fewer turns and less context can compound into much lower agent costs over hours of work.

Why this matters: If Grok can stay near frontier quality while burning fewer turns and cheaper tokens, teams running agents all day definitely care more about cost per completed task, and will more than likely make the switch.

Also: Musk says Grok 4.7, due in 3 to 4 weeks, is already significantly better than 4.6, and is being supplemented with a massive amount of SpaceX company data.

FROM OUR PARTNERS

Unwrap’s customer intelligence platform brings all your customer feedback (surveys, reviews, support tickets, social comments, ect.) into a single view, then uses AI + NLP to surface the most actionable insights and deliver them straight to your inbox.

Unwrap works with companies like Perplexity, Stripe, DoorDash, WHOOP, Southwest Airlines, lululemon and others to help product and CX teams cut through thousands of pieces of feedback and ensure no customer voice gets lost.

If your team is still relying on time consuming manual processes (or even a mix of manual work and AI), there's a much better way to aggregate and analyze feedback.

With Unwrap you get:

• All your public feedback automatically categorized

• The ability to ask questions about your feedback using natural language with Unwrap’s Assistant, directly in Slack, or in your favorite tools via Unwrap’s MCP

• Real-time alerts from your public feedback as they arise

• A clear view of customer sentiment

As a subscriber of The Neuron you can try Unwrap for free and join teams like WHOOP, Perplexity, Stripe, and DoorDash using it at scale.

Try Unwrap for free

🎓 AI Skill of the Day: Build a Hands-Free Morning Brief

Your first hour disappears when your inbox, calendar, and project updates live in separate tabs. Anthropic's updated Claude Voice guide says Voice can use connected Gmail, Google Calendar, Google Docs, and Slack tools.

Connect the tools you trust, then ask for the brief in small steps. Anthropic recommends splitting complex questions and notes that several tools at once can add delay.

Check today's calendar first. Then summarize urgent email and Slack items tied to those meetings. Finish with three priorities and flag anything that needs my approval before action.

Have a specific skill you want to learn? Request it here.

🍪 Treats to Try

1. ChatGPT for Linux now has an official OpenAI signup page so Linux users can get notified when the desktop app becomes available.

2. Claude in Chrome now turns the browser side panel into a Cowork session that keeps history, Skills, connectors, and multi-step tasks synced across Chrome and Claude’s apps.

3. Click gives agents live context ordinary web search misses, including YouTube transcripts, LinkedIn reactions, flight fares, and financial data through an MCP connection (like a Claude / GPT Plugin).

4. Infisical lets you sandbox Claude or another agent behind a fake API key while its proxy swaps in the real credential only when requests leave the agent.

5. Preview gives AI video teams one workspace to storyboard, generate, direct, compare models, and keep characters and locations consistent across shots.

Advertise to 700K+ readers here!

P.S: Yesterday, we said something about there not being pricing visible for Grokbot. That was Grantbot missing it; actually, website says its $200/ month Billed monthly with Cursor Ultra, or $120 / seat / month Billed monthly via Cursor Premium Teams.

📰 Around the Horn

• Demis Hassabis apparently pitched his U.S.-led AI oversight body weeks before stepping down as Google DeepMind's CEO to become Alphabet's chief scientist… which apparently was a PR move to save Google’s stock.

• Amazon/Twitch said streamers' content will now be used to train generative AI by default unless creators opt out.

• Cognition reportedly entered fundraising talks at a $40B valuation after reaching a $1B annualized revenue run rate.

• Google unveiled Pixel 11 devices with Gemini features, sign-language transcription, and a more natural voice-input system.

• Lovable raised $400M at a $13.3B valuation, more than doubling its valuation since December.

• Apple is discussing a nine-figure budget for multiyear news deals that would pay publishers when Siri AI uses their content.

• Mistral launched regional inference, third-party open-model hosting, and a European compute coalition targeting up to 1 GW by 2030.

• Dream Security detailed a multi-agent system that autonomously mapped and compromised Asian government entities over four days.

• OpenAI says frontier firms now generate 8.3x more output tokens per active user than typical enterprises, up from 2.6x.

FROM OUR PARTNERS

Your AI agents can send a year's worth of compliance work in a day. Try the firewall built for it with $1,000 in credits

ZeroDrift checks every message agents send against regulations and policies before leaving the server. Clean messages pass. Anything that breaks policy is flagged or blocked. Command provides real-time visibility into what was blocked and why, along with recommended fixes.

See Command in action and connect with the team to test it out.

🧩 Thursday Trivia

You know the drill: One is AI, and one is real. Which is which? (Vote in the poll below!)

A.

B.

New from The Neuron: AI Explained

AI models can now read and write DNA. In our latest episode, Radical Numerics CEO Eric Nguyen explains how genomic AI helped design CRISPR systems and viral genomes, why “deepfake viruses” create a new biosecurity problem, and how the same technology could help detect them.

Watch on YouTube by clicking the thumbnail above, or listen on Spotify.

THIS EPISODE WAS BROUGHT TO YOU BY…

Special Shout to Beyond Trust for sponsoring this episode!

A Cat’s Commentary

Trivia answer: A is AI, and B is real.

Btw: We just launched a robotics newsletter! Sign up for it here.

And before you go… have you subscribed to our YouTube Channel? If not, can you?

P.S: Love the newsletter, but only want to get it once per week? Don’t unsubscribe—update your preferences here.