Every week, I use Claude Cowork to go through the weekly stack of emails I get and pull the most interesting news and ideas of the week. I focus on creative professionals and filter out financial and industry insider info.
Claude Fable 5 is out, and it works differently than you’d expect
Anthropic launched Claude Fable 5 this week, the first public version of its Mythos-class model. It’s available on Pro, Max, Team, and Enterprise plans through June 22; after that, you’ll need usage credits. In the API, call it with claude-fable-5.
The Neuron’s summary is the most useful framing for creatives: Fable works best on long, messy work that runs for a while; research synthesis, complex multi-step projects, vision-heavy tasks. Ethan Mollick described working with it as less like chatting with an assistant and more like commissioning a small studio: you brief it and come back.
The practical prompting lessons from The Neuron’s skill section are worth keeping:
Treat it like an operating system, not a chatbot. Give it a destination, acceptance criteria, and a definition of done.
For ambiguous requests, it assumes and proceeds rather than asking. If the exact output matters, front-load your constraints: audience, format, scope, sources, success criteria.
It doesn’t over-format by default; prose unless you ask for structure. If you want extractive output (key findings, open questions, comparison points), ask for headings explicitly.
It has a strong search bias. If you need current information, specify source priority: “Check Anthropic docs first, then primary sources.”
One tool worth knowing about from The Neuron’s “Treats” section: Shotblock lets you stage 3D scenes, plan camera angles, export annotated storyboards, and generate AI-ready prompts. Free to try, and directly relevant for photographers and directors pre-visualizing shots. And Typeahead adds autocomplete to every Mac app, learns your writing style locally, keeps data on your device. Worth a look for writers. (via The Neuron)
(OK, Y’all, this is going to take a whole different approach to prompting than what we are used to. If you give it specific, pre-designed, and planned schema, this model will surprise and delight you. It is blazing fast for what I tried it on, but it is also a bit costly at twice the tokens of Claude 4.6. For most of us, 4.6 / 4.8 is plenty. Still, if you wanted to say, build a landing page that hooked up to your Stripe account, added images randomly from a database, and could detect the area of the world the visitor was coming from to deliver that specific language, all wrapped up in a highly interactive web page, sure… this is the model.)
Package your skills: the freelance offer prompt
BagelBots published a prompt this week that turns your skills into structured service packages. It builds out pricing tiers, deliverables, an ideal client profile, positioning statement, sales messaging, and objection handling. The last two sections, a first-10-clients action plan and a 30-day roadmap, make it more operational than most prompts of this type.
For photographers building out a commercial, education, or side-hustle business, fill in: Primary Skill, Experience Level, Target Industry, Target Client, Pricing Model, and Business Goals. Everything else it generates. (via BagelBots)
(Amazing prompt. Terrific. You should most definitely check this out.)
It’s Just a One-Way Street
I help photographers and creative entrepreneurs over 40 reclaim their confidence, cut through the noise, and build bold, independent lives through no-BS insight, experience, and action. Welcome to my Substack. I am glad you are here.In The Frame is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
Nine NotebookLM prompts for reading more deeply
Lea at Excellent AI Prompts wrote one of the better pieces this week; on using NotebookLM not as a summarizer but as a thinking tool for going deeper into books and articles. The core idea: take handwritten notes first (she cites the Mueller/Oppenheimer research showing longhand forces editorial choices that laptop transcription doesn’t), photograph those notes, upload them alongside the text, and use NotebookLM to reason across both.
Her opening prompt sets the right frame:
I’m reading [Book Title] by [Author]. Find: the author’s essays, interviews, or recorded talks on this subject; writing that explores [the book’s central theme] from a different angle; and any critiques or responses to this author’s ideas.
She covers nine prompts total; building a personal curriculum, creating an anti-curriculum (what NOT to read), finding where two books contradict each other, identifying what the author left out, and generating questions for a reading group. For writers and anyone doing research-backed creative work, this prompt is the most immediately useful piece of the week. (via Excellent AI Prompts)
(NotebookLM is still one of my favorite tools, and uses for it grow every week. I have hooked up Gemini to it, and now I am working on connecting Claude CoWork with NotebookLM to develop tools faster and cleaner.)
What Will You Do in the Next Five Minutes?
I help photographers and creative entrepreneurs over 40 reclaim their confidence, cut through the noise, and build bold, independent lives through no-BS insight, experience, and action. Welcome to my Substack. I am glad you are here.In The Frame is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
The AI-or-real photo problem keeps getting worse
BagelBots’ recurring “Which image is real?” poll reported 54% wrong last week — nearly identical to the week before (55%). Both weeks the subjects were people: laughing friends, babies. When the gap between AI and real closes to coin-flip territory on human subjects, photographers who work in portraiture and documentary have a real provenance argument to make. Not just aesthetically — verifiably. (via BagelBots)
(Make work that is not AI-able. If you can make what you do with AI, then you will have to justify/defend why you do it. And that can certainly be done, but it will not be based on financial reasons or speed. That’s not possible. So how do you do it? This is the challenge. I will have more on this next week.)
Gemini 3.5 Live Translate
Google launched near-real-time speech translation this week in AI Studio, Google Translate, and Meet. Still early, but the first version of this that actually works in conversation is a practical tool for any photographer, writer, or educator working across language barriers with clients or collaborators. (via The Neuron)
(Travel photographers are going to love this. Immediate, real-time speech translation. Now all you need is immediate, real-time WiFi access. My wife and I are getting the Starlink mini for travel. Losing signal and wifi in the desert southwest is more common than not.)
I’m working on a project for a small publication and finding some quiet, little images that will go along with the poetry. Here are two.
If you’re over 40 and still hungry to make, build, and create, stick around. This space is for people who aren’t done yet (and never will be). I’ve got five decades of wins, failures, comebacks, and creative battles under my belt, and I’m sharing everything that still works—and burning the rest.
No fluff.
No hustle porn.
Just real tools for building a creative life on your own damn terms.
My Personal Website
My Photography Website.
Me on Medium.
For those who don’t know, my sites were attacked and it has been a couple of weeks of sheer hell on my end. I am desperately trying to move them to another host, and so they are breaking and problems abound. I have someone helping me now, so I am hoping it all goes well in the next week. In the meantime, you will see some glitches at my other sites, and I apologize, but such is the world that we have people who just love to destroy what others have built.
My premium members have full access to over 24 productivity and creativity tools, my books, and weekly updates to help them build a better creative business.