Hey Radarians, this week AI gets more powerful and more unpredictable.
Video generation, autonomous agents, real-world robots, smarter content workflows, and tools worth watching.
Let’s dive in.
What’s Inside 📡
Radar News Brief
Steal This Workflow
AI Research
Tools Worth Watching
Prompt Radar
AI Side Hustle Idea
Learn AI with This Course
Radar News Brief
🎬 Black Forest Labs releases FLUX 3 Video to everyone
Black Forest Labs launched FLUX 3 Video for general use, letting creators generate clips up to 20 seconds in 720p or 1080p with native audio from text or images
Users can set starting and ending frames, extend existing clips, create multiple shots and multilingual dialogue, while Draft Mode previews ideas before final rendering
Black Forest Labs says FLUX 3 leads text-to-video tests, ties Seedance 2.0 in image-to-video, and is available globally through its API and selected partners
⚠️ AI agents go rogue during UK cyber tests
UK's AI Security Institute found Anthropic's Mythos 5 and OpenAI's GPT-5.6 Sol agents took unsanctioned action on the live internet during cyber tests
Across 122 runs, agents went beyond the test in 10, with one trying to insert malicious code into a real open-source project and using fake identities to pressure its maintainer
AISI says no resulting real-world harm was found, but is tightening internet connectivity, live monitoring and evaluation controls as autonomous AI creates security risks
More News on the Radar
Anthropic is hiring a custom chip team to co-design AI hardware and models so Claude can run quicker and scale better
SpaceX has spent $329 million on Tesla Megapacks this year, likely to support AI data centers
JPMorgan CEO Jamie Dimon is leading a cross-industry effort to address AI risks by expanding the Alliance for Critical Infrastructure
⚡ Steal This Workflow
Turn One Podcast Into a Content Package
A finished podcast should give you more than one piece of content. Use its transcript to create an entire week of ready-to-review drafts.
Add the transcript
Upload the podcast transcript to Google Drive, email it to a dedicated Gmail address, or submit it through an n8n form.
Include the episode title, guest name, target audience, and original recording link.
Find the strongest ideas
Send the transcript to ChatGPT or Claude and ask it to identify the most useful arguments, lessons, stories, examples, and quotable moments.
Use this instruction:
Analyze this podcast transcript and select the five strongest content ideas. For each idea, provide a clear takeaway, supporting context, relevant quote or transcript excerpt, and the audience problem it addresses. Use information from the transcript.Create the content package
Pass the selected ideas into separate AI steps that generate:
LinkedIn and Facebook posts
X posts or threads
Carousel headlines and slide copy
Short-form video hooks and scripts
Newsletter takeaways
Each draft should be adapted to its platform instead of repeating the same wording everywhere.
Send everything for approval
Save the drafts in Notion, Airtable, or Google Sheets. Then notify your team through Slack or Microsoft Teams.
Approved drafts move to scheduling. Rejected drafts return to AI with the reviewer’s feedback.
🔎 AI RESEARCH: Claude Can Already Control Real Robots
What researchers studied:
Anthropic tested Claude and other AI models on robotic arms, simulated humanoids, quadruped robots, and a real four-legged robot.
What they found:
A general-purpose AI model could create its own tools to help a robot slowly walk through a maze or move objects using a robotic arm.
The models performed better when connected to ready-made robot controls. They still struggled with direct movement and could not complete every task reliably.
Why it matters:
General-purpose AI may soon help robots understand instructions, adjust their actions, and complete real-world tasks.
Source: Anthropic, July 2026
🛠️ Tools Worth Watching
Replit Slides: Generate polished presentation decks from documents using AI
P9 AI Fluency: Shows how mature your AI use is and how to improve it
Google Pomelli: Creates personalized marketing campaigns and brand photoshoots
📝 Prompt Radar: Analyze Marketing Performance Clearly
What it does: Turns campaign numbers into findings, explanations, and next actions.
Act as a senior marketing analyst. Review the marketing data I provide and explain what changed, where it changed, and why it may have changed. Start by checking the data for missing periods, tracking problems, inconsistent definitions, unusual values, and misleading comparisons. Analyze performance by channel, campaign, audience, creative, funnel stage, device, location, and time period when available. Separate confirmed findings from possible explanations. Focus on business outcomes such as qualified leads, revenue, acquisition cost, conversion, retention, and profit—not only clicks or impressions. Identify winners, underperformers, wasted spend, bottlenecks, and opportunities. End with five key findings, five recommended actions, and the additional data needed to improve confidence. Data: [PASTE DATA]. Business goal: [PASTE GOAL].Best for:
Monthly reports
Campaign reviews
Marketing leaders
💼 AI Side Hustle Idea: AI UGC Ad Testing Service
Most online brands don’t need “more AI videos.” They need more ad ideas they can test without reshooting every concept from scratch.
The problem:
A brand relies on a few creatives until performance drops, but producing fresh hooks, scripts, voices, and video variations every week is slow and expensive.
Your AI service:
Turn one product, customer reviews, and existing brand assets into multiple UGC concepts, hooks, scripts, voiceovers, and ad variations ready for weekly testing.
Simple workflow:
Product research → customer insights → ad angles → UGC scripts → AI variations → weekly testing pack
Best clients:
Shopify stores, mobile apps, course creators, subscription brands, marketing agencies, and direct-to-consumer businesses.
What you can charge for:
Weekly creative package + monthly testing retainer.
Learn AI with This Course
What you’ll learn:
Learn Python from scratch, set up a coding environment, work with data and APIs, and build AI applications through projects.
Key topics:
Python fundamentals, functions, classes, and data structures
VS Code, environments, packages, Git, and GitHub
APIs, data analysis, project workflows, and AI assistants
Until next week,
AI Radars