One Film Cost $250 Million. The Other Tells the Same Story for Five Figures. Both Open This Week.

You don't need hundreds of winning ideas. Just one.

Inside this free guide are 200+ real ways entrepreneurs, creators, and freelancers are using AI to build new income streams. One could become your next business.

Discover actionable AI opportunities across content creation, eCommerce, digital products, freelancing, automation, and more—completely free.

1  Mira Murati's startup launched Inkling, the strongest open-weight model yet to come out of a US lab.

After 17 months of building, Thinking Machines released Inkling, a model designed to balance strong performance with efficiency and built for companies to download, modify, and run on their own infrastructure. Most competitive open-weight models this year have come from China: GLM-5.2, DeepSeek V4 Pro, Kimi K2.6. Inkling is the most competitive US answer to that trend yet. See how it stacks up against the field.

2  OpenAI shipped Codex Micro, a physical keypad that lets you command AI agents with a button, a joystick, and a knob.

Codex Micro is OpenAI's first hardware product: a customizable mini keypad that lets you prompt Codex verbally with one button, trigger common workflows with a joystick, and dial up or down the model's reasoning depth with a physical knob. It won't be the last hardware release. A separate report points to more physical products in development. Watch the release video here (2M+ views).

3  Anthropic launched a standalone enterprise deployment firm backed by Blackstone and Hellman and Friedman.

Ode with Anthropic is a new services firm built to help organizations move from running AI pilots to embedding Claude into live operations. The firm pairs Anthropic's frontier models with a team of experienced AI engineers and operators, targeting the gap that has stalled most enterprise AI programs: they can experiment with the tools but struggle to make deployment actually stick. Blackstone and Hellman and Friedman are both in as financial backers.

The coincidence. Christopher Nolan's The Odyssey opens in theaters this week with a 97% on Rotten Tomatoes, the best score of his career. On the same week, AI filmmaker Ash Koosha released a trailer for ODYSSEUS: The Fall, a 135-minute feature film covering a different take on the same Greek epic — built almost entirely with AI. The timing is not coordinated. It is just an unusually sharp illustration of where AI sits in relation to Hollywood right now.

The gap in the numbers. Nolan's version cost $250 million, required nearly 100 days of filming, and assembled an ensemble including Matt Damon, Tom Holland, and Anne Hathaway. Koosha handled the script and character voices himself. AI, primarily Kling, handled the actors, sets, and cameras. His all-in production cost landed in the mid-five figures. Koosha described the contrast as "the ultimate in human creation, compared to one man's collaboration with AI."

Hollywood's split. Directors are genuinely divided. Martin Scorsese, Peter Jackson, and George Lucas have each described AI as a tool for storytelling, comparable to any other technical innovation the industry has absorbed over the decades. Many others stay entirely away. The public's reaction to Koosha's trailer — visible in the comment sections on the YouTube release — suggests audiences are not yet ready to watch a full-length AI film the same way they watch a conventionally made one.

What the comparison actually tells you. AI video tools have improved faster in the past three years than most people outside the industry track. The gap between Koosha's film and Nolan's is not primarily a gap in tools anymore. It is a gap in craft, accumulated institutional knowledge, human performance, and audience trust. Those are harder to close than the tooling was. The more interesting question than "can AI make a film" is "what does an audience decide to watch and why." That answer has not changed as fast as the technology has.

Simpler prompts, better results. OpenAI's official GPT-5.6 prompting guide has been widely shared since the model's launch. The counterintuitive tip getting the most attention: simplifying your prompts, rather than adding more instructions, lets the model choose its own most efficient path and often improves results.

Your screen as context. Hey Clicky's latest update adds screen-aware dictation: the tool reads what is on your screen and uses that context to understand what you need before you finish asking. Watch the demo here (2M views) or try it yourself.

Vintage brand ads. Someone asked ChatGPT to reimagine modern brands as 1970s print advertisements. The results are striking, and the prompt that produced them is in the comment section. Worth saving if you want to produce retro-styled visuals for any brand.

AI vs. Microsoft Paint. Someone asked an AI to draw a person in Microsoft Paint. The results are what you would expect, but the commentary running alongside makes the video genuinely funny rather than just technically interesting. A useful reminder that there are still tasks where the human with a trackpad wins.

Websites from chat. ChatGPT can now build and deploy a complete website directly from the app, with no separate platform required. OpenAI's walkthrough of how to use it has picked up 1,000 bookmarks. For anyone who has been waiting for the barrier between describing a site and publishing one to disappear, this is it.

Write your entire cold outreach sequence in one prompt

This ChatGPT prompt generates a complete, seven-message outreach sequence for any prospect type: a LinkedIn DM, a cold email, two follow-ups, a reactivation message, an objection response, and a final breakup email. Fill in five details about your prospect and offer, set the tone, and the model builds the whole sequence personalized to the context rather than using a generic template.

Replace each placeholder with specifics. For [Prospect type], name the role you are reaching: "VP of Marketing at a Series B SaaS company". For [Pain point], name one specific problem they are likely feeling right now rather than a category: "churning enterprise accounts due to slow onboarding" lands better than "retention issues". For [Offer], describe what you are selling in one sentence including the specific result it delivers. The more precise the inputs, the less generic each message. Run the full sequence first, then go back and adjust the tone of individual messages before sending.

Ready to transform artificial intelligence from a buzzword into your personal revenue generator

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get Your Guide