§ 01 · The Hook
The bait, then the rug-pull.
Matt Wolfe opens with a single load-bearing claim — ChatGPT Image 2.0 is 'actually way more useful' than what came before — and then proves it for thirty-two minutes by screen-sharing one prompt after another. The hook is the title's number (40+) plus a small promise: even if you think you know what this model can do, he'll show you something you missed.
What the video promised.
stated at 00:32“I promise you that if you're watching this and you think you know everything this model can do, I'm probably gonna show you some stuff that you had no idea that this model was capable of.”delivered at 17:55
Where the time goes.
00:00 – 00:47
01 · Cold open + thesis
Talking-head setup: new model dropped, it's actually useful, here come a ton of prompts — and a promise that he'll surprise even veteran users.
00:47 – 03:20
02 · YouTube thumbnail concept boards
First prompt: a 2x3 grid of thumbnail concepts for a meta ChatGPT-Image-2 video. Output looks Canva-ish and is locked to squares, but he uses it as concept fodder rather than finals. Then aspect-ratio fix.
03:20 – 04:23
03 · A/B/C thumbnail test sheet
Three thumbnail concepts side-by-side, each with an attached hook rationale. Aspect ratio is still rough but the per-concept hook copy is the unexpected unlock.
04:23 – 05:20
04 · Nine-frame video storyboard
He asks for a nine-panel storyboard for an intro about 'AI images are useful now' — skeptical creator, garbled text shot, readable text, infographic, comic page, surprised reaction, title card. The storyboard is the lesson: one prompt = a full visual outline.
05:20 – 06:50
05 · URL-aware Facebook ad
Gives ChatGPT his futuretools.io URL and asks for a Facebook ad. The model fetches the actual logo, an accurate screenshot, and a CTA — proving that this model can browse the web and pull live assets into the generated image.
06:50 – 07:22
06 · Channel banner ad
Same URL trick on his YouTube channel. Returns a 'Stay Ahead of AI / Watch Matt Wolfe on YouTube' banner with weekly-AI-updates value props and a made-up MW logo he half-likes.
07:22 – 09:00
07 · Social carousels (IG + LinkedIn)
Seven-slide Instagram carousel from a single prompt — hook slide, five teaching slides, CTA slide. Then a parallel LinkedIn version with a more business-flavored angle. Big point: one prompt = a whole content unit, not one image.
09:00 – 09:22
08 · Quote-card pack
Four standalone quote cards generated in one shot. 'The best AI tools do not replace your taste, they amplify it' — designed for IG posting without manual layout.
09:22 – 10:32
09 · 30-day social media calendar
Visual 30-day content calendar for a fictional local coffee shop — promo, behind-the-scenes, education, customer story, community, repeating. Frames a real use case: hand this to an intern as a month's runbook.
10:32 – 12:00
10 · Brand mood board
Brand kit for fictional AI app FlowPilot — palette with hex codes, typography, monograms, icon styles, brand personality. Hit one content-policy false-positive that resolved on re-submit with identical text.
12:00 – 13:20
11 · Logo exploration sheet
Eight logo variants for fictional brand Signal Lab, each with a one-line rationale (wordmark, monogram, lab-beaker, etc.) plus a color palette and icon-only versions.
13:20 – 14:20
12 · App launch one-sheet
Product launch sheet for app called Recall — logo, copy, mock screenshot. Tangent: this could be the visual spec a coding agent then writes the app from.
14:20 – 15:20
13 · Local-business flyer
Promotional flyer for fictional Wolf Clean mobile detailing in San Diego — with a recognizable skyline behind the copy.
15:20 – 16:20
14 · Restaurant menu
Print-ready taco menu for fictional El Lobo Loco Tacos. Acknowledges he let the AI invent the food and prices, but the layout reads like a real laminated menu.
16:20 – 17:20
15 · Event poster + schedule graphic
Creator meetup poster ('Signal and Spark') with San Diego skyline, then a printable schedule for a fictional Creator AI Summit — registration, keynote, sessions, lunch, fireside.
17:20 – 19:00
16 · Real estate flyer from a Zillow URL
The biggest 'whoa' moment: pasted a $5.3M La Jolla Zillow listing URL and asked for a flyer. ChatGPT pulled the actual address, price, details, AND the real listing photos straight off the page and laid out a clean realtor flyer.
19:00 – 20:30
17 · 3-day Tokyo travel itinerary
Visual three-day Tokyo plan broken into morning/afternoon/evening blocks. Notes he can't verify the recommendations but the document is print-ready.
20:30 – 21:50
18 · Illustrated city map
Asks for an illustrated San Diego visitor map. Honest negative: Coronado and downtown land roughly right, but North Park and Balboa Park are wrong. Maps remain the weakest use case.
21:50 – 23:20
19 · Travel + family checklists
Five-day beach packing list, then a one-day Disneyland strategy with hour-by-hour ride/rest blocks and meltdown-avoidance tips. Same pattern: structured, printable, useful as-is.
23:20 – 24:50
20 · Wedding schedule + chore chart + habit tracker
Guest wedding-weekend schedule, then a kid chore chart with reward tracker, then a printable 30-day habit tracker. The 'analog tool' lane — paper documents AI generates faster than a designer.
24:50 – 26:30
21 · Garage organization planner + infographics intro
Five-phase garage plan with budget, timeline, before/after, checklist. Then he pivots: 'this is where this model beats Nano Banana — infographic accuracy'.
26:30 – 28:00
22 · AI agents infographic
Explainer infographic of how AI agents work for a non-technical audience. First pass was boring blue/white; a follow-up 'make it more colorful' rebuild proves iterative refinement works.
28:00 – 29:40
23 · Infographic from a research-paper URL
Feeds the model the VOID-Model GitHub page (video object removal) and asks for a visual explainer. ChatGPT reads the page and lays out problem / how-it-works / training / why-it-matters. The universal pattern: any URL becomes a visual explainer.
29:40 – 31:20
24 · Timeline + mind map + process diagram
Horizontal AI-image evolution timeline, a 'ways creators can use AI' mind map with auto-fleshed-out sub-branches, and a customer-inquiry workflow process diagram. The mind map is the underrated unlock — partial input expands into structured output.
31:20 – 32:40
25 · ChatGPT vs Claude vs Gemini vs Perplexity chart
Comparison chart for non-technical users. Mostly accurate; he disputes the Claude long-doc framing in favor of Gemini. Notes the meta-irony: ChatGPT willingly lists its own weakness ('can sound confident when wrong').
32:40 – 33:40
26 · Prompt cheat sheet + mini slide deck
A printable AI-image-prompt cheat sheet (formula, ingredients, common mistakes, examples). Then a five-slide deck about chatbots-vs-agents — separate slide images you can click through.
33:40 – 35:00
27 · Sales one-pager
One-page sales pitch visual for fictional automation agency 'Mattomation' aimed at small business owners. Defaults to blue/white again — color-steering needed for brand specificity.
35:00 – 36:50
28 · Ecommerce: packaging + product image set + merch mockups
CPG packaging concept for fictional 'Wolf Power' protein snack — logo, palette, multiple SKUs. Then a six-image product set for a 'Lampinator' desk lamp (hero, lifestyle, callouts, mood, packaging). Then a merch mockup sheet — t-shirt, hoodie, sticker, mug, dad hat.
36:50 – 38:20
29 · App store screenshots + website hero
Five App Store screenshot mockups for fictional Taskflow productivity app, then a full hero section mockup for 'Draftly' AI writing assistant with menu, CTA button, screenshot, social proof, and headline.
38:20 – 40:00
30 · Course curriculum + onboarding checklist
Visual six-module curriculum for an 'AI for Everyday Productivity' course, then a client-onboarding checklist for a freelance video editor — kickoff, brand assets, footage, style, timeline, feedback, delivery.
40:00 – 41:20
31 · Messy notes → polished one-page action plan
Pastes raw stream-of-consciousness notes ('I have too many ideas, I overthink thumbnails, I want a simpler system…') and asks for a printable visual action plan. Returns 'Overwhelmed → Output: A Simple Creator Action Plan' with goal, bottleneck, priorities, stop-doing, daily process, and a seven-day action plan.
41:20 – 42:20
32 · Printable recipe card
High-protein breakfast tacos recipe card — image, ingredients, instructions, fit-for-print formatting.
42:20 – 44:30
33 · Six-panel comic page
Finale: 'The Day My AI Assistant Took Over My Inbox' — six panels with consistent character, color, dialog ('How do I have 437 emails… AI please help me clean up this inbox… These are actually good. I can breathe again.'). Comic format is where this model genuinely surprised him.
44:30 – 46:40
34 · Wrap + URL-aware verdict
Recaps the framing: Nano Banana still wins on photorealism, but ChatGPT Image 2.0 wins on text accuracy and on doing research (reading URLs) before generating. Tells viewers to check the AI news video for more.
46:40 – 50:12
35 · CTA + channel pitch
Long-form subscribe pitch: weekly AI news every Friday, ~300 stories filtered down to the 10–15 that actually matter, plus tutorials in between. 'My goal is to help prevent the overwhelm.'
Visual structure at a glance.
Named ideas worth stealing.
07:22concept
One prompt = one full content unit
Repeated thesis: a single ChatGPT Image 2.0 prompt now outputs a 7-slide carousel, a 6-panel comic, a 9-frame storyboard, a 30-day calendar, an 8-logo exploration sheet — not just one image. The unit-of-work has changed.
Steal forReframe your own workflow descriptions — instead of 'AI made me an image', show 'AI made me a deliverable I could ship to a client'.
05:20concept
URL-aware image generation
You paste a URL into the prompt; ChatGPT browses it, pulls the brand assets and copy from the live page, and bakes them into the generated image. Demonstrated on his own futuretools.io, his YouTube channel, the VOID-Model GitHub paper, and a Zillow $5.3M listing.
Steal forThis is the punchline for your own demos — any creator who ships sales pages, ads, or product pages now has a 'paste URL → generate flyer' button. Make a 60-second short of just this.
19:30list
Six axes of comparison with Nano Banana
- Photorealism — Nano Banana still wins
- Text accuracy — ChatGPT wins decisively
- Following the URL — ChatGPT wins
- Infographic accuracy — ChatGPT wins (Nano looks prettier)
- Aspect ratio compliance — both struggle
- Map / spatial accuracy — both still weak
Throughout the video he runs the same comparison without naming it as a framework — the implicit grid is the actual takeaway.
Steal forWhen you compare two tools on stream, write the axes on screen before you start. Don't make the audience reconstruct your framework.
40:00concept
Messy-notes-to-polished-document pattern
Paste unstructured personal brain-dump → ask for a one-page printable visual plan → get a structured, sectioned, action-plan-shaped document.
Steal forThis is a stronger framing than 'AI organizes your notes' — record yourself rambling into a doc, then run this prompt live. The before/after is the hook.
Lines you could clip.
00:13
“It is just actually way more useful.”
00:32
“If you think you know everything this model can do, I'm probably gonna show you some stuff that you had no idea that this model was capable of.”
07:22
“It created the entire seven slide carousel for us here with just a single prompt.”
17:55
“I didn't know if it was gonna pull in actual images from the flyer and then incorporate them into this image, but it did.”
30:20
“All you have to do is feed it the URL and say, make an infographic based on what's at this URL.”
41:35
“Overwhelmed to output — a simple creator action plan for making more videos with less friction.”
How they spent the runtime.
Hook length47s
Info densityhigh
Filler8%
Things they pointed at.
04:31toolNano Banana (Google AI image model)
03:11toolIdeogram
How they asked for the click.
48:00subscribe
“If that's something that sounds interesting to you, maybe consider liking and subscribing to this channel.”
Soft, late, and earned — comes after 33 minutes of genuine value with no mid-roll pitch. The pitch is fully aligned with the channel's identity ('I cut 300 stories down to 10–15 so you don't have to') rather than a generic ask. Weakness: it runs ~2 minutes long, longer than most viewers will tolerate.
Word for word.
HOOKopening / re-engagementCTAthe pitchmetaphoranalogystory
00:00HOOKSo last week, OpenAI released ChatGPT images two point o, and it is a really, really good image model. Honestly, I've been spending a ton of time with this model, and pretty much all I can say about this model is it is just actually . You can make slide decks. You can make full carousels for places like Instagram. You can get it to explain stuff to you in images and have it be really accurate. I went and tested tons of prompts. Like, I kind of lost count, but I'm gonna show you just a handful of the cool stuff that you could do with this model. I promise you that if you're watching this and you think you know everything this model can do, I'm probably gonna show you some stuff that you had no idea that this model was capable of. But let's start with some YouTuber use cases. I know a lot of you aren't YouTubers, but I promise this will get better and better as the video goes on. So for this first one, I gave it a prompt, create a YouTube thumbnail concept board for a video about ChatGPT images two point o. I know. Super meta. Video idea, 33 useful things ChatGPT
00:58images two point o can make. Create six different thumbnail concepts in a two by three grid. You decide the best visual metaphor, facial expression idea, layout, and short thumbnail text for each concept. Requirements, each thumbnail should feel visually distinct, use large readable text, make the concepts clickable but not misleading, avoid clutter, use modern AI tech creator aesthetic, include one concept that compares chat g p d images to nano banana. Style, bold YouTube thumbnail design, high contrast, blue purple tech accents, expressive, clean, readable at small size. Now I'm not gonna read the prompt for every image I'm gonna go to. Originally, I was gonna do 33, but I ended up doing way more than 33. I don't actually know the final count, but it'll probably be, like, in the title of the video. But here's what this prompt got me here. It made this image of six different thumbnail concepts with different facial expressions, different text, different potential images I can use, and they're horrible.
01:48I mean, they look like something that's maybe like a Canva template, but they give us a good starting point. Now one obvious issue with these is that it generated them all as squares, and we all know YouTube thumbnails are not squares. But that's easily corrected. I said, now make each thumbnail sixteen nine so they match an actual YouTube thumbnail aspect ratio, and, well, it went through and it fixed them all for me. I don't quite think these are still 16 by nine. It's still struggling with this aspect ratio a little bit, but it's a little bit closer to what you want. And, the idea here isn't to actually make the thumbnail for me. It's to give me a bunch of concepts
02:23so that I can land on the best sort of overall idea I wanna use for the thumbnail and then dial in that individual thumbnail. Going further down this thumbnail rabbit hole here, create an a b c thumbnail test sheet for a YouTube video. Chat GPT images two point o versus Nano Banana. Give me three thumbnail options, each one with a different sort of concept, some details about what I want the thumbnail to include, some requirements for the typography,
02:48and some details for the style here, and this is what it gave me. Once again, kinda screwed up the aspect ratios, but again, a decent starting point here with three completely different concepts and three hooks to go along with the videos. First hook, head to head rivalry makes people want to know which one wins. Second hook promises real world value instead of just pretty pictures being actually useful,
03:10and the third hook taps into a long standing pain point and frames this as a major breakthrough. It does text, which honestly, I don't really think is a pain point too much anymore. Nano Banana has been pretty good at text for a while, same with ideogram, but all concepts that we could play with. If I'm being totally honest, I still don't think that even chatty p t image or nano banana is very good at making YouTube thumbnails, but it isn't bad at giving me some general rough draft concepts that I could then turn into my own. It can do storyboarding for a video you're about to put out. Create a nine frame storyboard for a YouTube video intro. Video topic, AI images are useful now. I gave it each frame of the storyboard and some style details down here. As with all the prompts, I'm just gonna brush over them quickly, but I'm gonna leave them on the screen so you can pause and read the whole thing if you'd like. But here's the storyboard it gave me. Frame one, a skeptical creator. Frame two, old AI image with garbled text. Shot three, a new AI image with readable text. Shot four, menu design. Five, an infographic. Six, a comic page. Seven, YouTube thumbnails. Eight, a surprise creator going, wow. And then finally, the title card, 333
04:17things AI images can finally do. Again, this was meant to be a little bit meta about the video you're watching now. Now another thing that I think is really cool about this model is that you can give it a URL and say, make me an image that goes along with this URL over here, and it does a really good job. So maybe you run a business and you wanna make some Facebook ads. Well, you can give it the website you wanted to make a Facebook ad for and watch as it looks at the site, figures out what it's about, and then goes and makes your ad. For example, create a Facebook ad to promote my website, futuretools.io.
04:50Create something that's likely to get a lot of clicks, but is accurate to what the site promotes. And we can see here's the ad it made. It actually pulled in the future tools logo directly from the website, and it knows exactly what the site's about and even has, like, a semi accurate screenshot to the website. Find the best AI tools, discover AI tools, stay on top of AI news, and explore helpful videos all in one place. This is very accurate to what's available on the future tools website. We have a call to action, explorefuturetools.io.
05:18The URL down here built for creators, entrepreneurs, and curious professionals, and it even mentions the newsletter. Get the best AI tools and news delivered. That's a solid ad that literally took the main selling points of my website and the images from my website and put them into a decent looking image. To me, that seems super helpful to, like, any business who needs to run ads. Here's one that I did for my YouTube channel. Create a horizontal banner ad that promotes the Matt Wolf YouTube channel. I gave it the URL. Make sure it has a CTA to subscribe and tells people why they should watch. And here's the banner that it created. I didn't explain what my YouTube channel's about. I just gave it the URL. And it says stay ahead of AI. Watch Matt Wolf on YouTube. Get practical AI news, useful tools, and step by step tutorials for creators, entrepreneurs, and curious tech fans. Weekly AI updates, tool breakdowns, actionable tutorials. Subscribe now. It actually made this logo up. That's not a logo that I actually have, but I kind of like it, honestly. Maybe I'll make it something like that for my Matt Wolf logo. I don't know. But this was created by me just giving it a YouTube URL, it going and looking what that channel's about, and then making this image to represent that channel. Alright. Now let's talk about some ways that you could use this to grow on social media. Did you know that it can create a multi image carousel for you? Create a seven slide Instagram carousel for creators. Topic, seven ways to use AI images that are actually useful. You decide the best wording for each slide. I tell it slide one should be a strong hook, slides two through six should each teach one useful idea, and slide seven should be a call to action. I give it some other details, and look at this. It gave me seven different slides here. Slide number one, seven ways to use AI images that are actually useful. Slide number two, thumbnail mock ups. Now we're getting into the actual use cases, infographic carousels,
07:02HOOKstoryboards, b roll stand ins, create missing cutaways when you need a visual fast, ad creators, test promo concepts before spending money on a shoot. And then finally, a call to action. Save this carousel. Use one idea in your next post. Follow for more creator AI tips. It created the entire seven slide carousel for us here with just a single prompt. Here's another one. This time, make quote posts that I can use on Instagram. Use these exact quotes, one quote per card. I gave it four quotes to use and some design
07:32HOOKaesthetic direction. And here we have four quote cards that can be used on social media. The best AI tools do not replace your taste, They amplify it. The future belongs to people who can explain things clearly. Don't chase every tool. Chase the capability it unlocks. Your workflow is your moat. It created all four of these all in one single prompt. I don't have to go and create a new image for each slide in the carousel anymore. Alright. This time I gave it more specifics towards LinkedIn. Create a professional LinkedIn carousel about how AI images are becoming useful business tools. Once again, you decide the copy and layout. I gave it some style details here, and here we go. A LinkedIn carousel. AI images are becoming useful business tools. Fast campaign concepts, so it's more focused on businesses. We've got some ecommerce product shots, social content at scale, better sales and pitch decks, customer education and support, and the final takeaway, AI images work best as a business multiplier. Definitely feels more geared towards a LinkedIn audience than an Instagram audience from just shifting the prompt a little bit to tell it it's for a professional LinkedIn carousel. But not only can it create these various carousels and things like that for you, it can plan an entire social media calendar for you and give it back to you an image form. Check this out. Create a visual thirty day social media content calendar for a small business. Business type, a local coffee shop. You decide the post ideas. Thirty days of post concepts. A mix of promos behind the scenes, customer stories, educational posts, and community content. And check this out. It made a full thirty day calendar that I can follow along to. Day one, promo, featured drink of the week. Day two, morning prep before opening. It's a behind the scenes shot. Day three, educational. How we steam milk for lattes. Day four, community post. Ask what's your go to coffee order. Day five, customer story. Meet a regular plus a favorite drink. Day six, back to a promo, pastry plus coffee combo special. Day seven, behind the scenes, full thirty days of content ideas that you can post to your social media, and you can even, like, print this out and give it to a intern or whoever's managing your social media account and let them just follow the map, and you have a month's worth of social media posts that you can kind of duplicate every month if you want. How about for dialing in a brand? Create a brand mood board for a AI productivity app called Flow Pilot. Audience, busy creators, and entrepreneurs,
09:46you decide pretty much everything. Now this one was kind of interesting because when I first gave it the prompt, it said, we're so sorry, but the image we created may violate our content policies. If you think we got it wrong, please retry or edit your prompt. I don't know what about a brand mood board for FlowPilot would cause any issues, so I literally just gave it the exact same prompt again. I literally did not change a single word, copied and pasted it back in, and the second time, it made me my mood board. Don't know why it hiccup the first time, but here you go. It's got the various colors, but it's even got the hex code for those colors so you can get the exact colors every time. It gave us a logo. It gave us typography to use. It gave us a pilot mark, a flow monogram,
10:28a compass mark, a UI visual style, icon styles, brand personality. It pretty much designed the full brand and what it should look like all in a single image for us. Maybe you need a logo for a new brand you're working on. Create a logo exploration sheet for a fictional creator brand called Signal Lab. Create eight different logo concepts, short labels explaining each direction,
10:50a small color palette, and a few icon variations. And here's what it generated for us. Eight different potential logos that we can use as well as a quick description of why that image. So word mark focus, clean editorial word mark for authority and clarity. Monogram, geometric monogram for a modern compact brand mark. Lab mark subtle lab beaker conveys experimentation and rigor. We've got color palettes and the various icon
11:15variations sort of more by themselves without the text. But if you like one of these logos, you're off to a pretty good start. Getting AI to create logos for you used to be such a nightmare, and now it's just so simple. You can have one prompt generate multiple logo variations for you. Create a product launch one sheet for an app called Recall. What the app does, it helps creators save ideas, summarize notes, and turn messy thoughts into action plans. Gave it some requirements here, and here's what it generated for us. This could also be the starting point of a web page as well, but it even created the logo, created all the copy, created a screenshot of what the app could look like. You could even go and do this process if you're thinking about building an app and have it kind of design the app and the logos and the fonts and the color schemes and have it all figured out before you ever write a line of code. And then you can feed it this image, and then the code could actually be written based on your initial starting image, which could be kind of a cool concept. Create a promotional flyer for a fictional local business called Wolf Clean, a premium mobile car detailing service in San Diego. Here's my flyer for Wolf Clean. If you need your car detailed, well, Wolf Clean seems like a pretty good option. We've even got the San Diego skyline here in the background. Pretty good flyer, honestly. Maybe you run a restaurant. Create a polished restaurant menu for a fictional modern taco restaurant in San Diego called El Lobo Loco Tacos. It created a nice looking menu that you could print out and laminate if you actually have a storefront.
12:39Now I told it to just sort of make up the food and the prices, but obviously, if you have a restaurant, you'd I would guess plug in that stuff. I mean, it could be an interesting experiment to let AI make a menu for you, and then you have the people in the kitchen go and figure out how to make that menu. Logistically, it sounds like a nightmare, but you could. But come on. If you saw this menu laminated
13:01at a taco shop, you'd think it was fairly high end, I would think. Create a premium event poster for a creator meetup, location San Diego. Audience, creators, entrepreneurs, and people interested in AI tools. Here's our flyer. It puts some placeholder text because I didn't give it any sort of date or venue, but apparently, our creator meetup is called Signal and Spark. That was some creative license that it took. It's got a nice San Diego skyline here. Join an evening of demos, conversations, and practical ideas around AI for content, business, and productivity. Meet like minded people, discover new workflows, and get inspired by what's possible right now. I'd go to that. I tested it, creating a full schedule graphic for an event called the Creator AI Summit, and it created this, a whole sort of schedule of events for a Creator AI Summit. 8AM registration and coffee. 9AM opening keynote. 10 sessions, ten forty five sessions, 11:30 fireside,
13:4912:15 lunch break, etcetera. You get the idea. I'd imagine this would be kinda something you print out a bunch of copies of, and as people walk into the summit, you hand them out so they know the exact flow of the day. Okay. So now let's try this real estate one. I thought this would be an interesting idea to take an actual listing of a real house that's for sale right now from Zillow. Now this is a house that's $5,300,000
14:10in La Jolla in San Diego. I wanted to see if it would make a flyer and pull the information from the actual Zillow page to make the flyer. So I gave it this prompt, create a real estate listing flyer for this modern coastal home in San Diego, and then I linked it straight up to that Zillow page. I told it to basically figure everything else out, and it made this flyer here. Modern coastal living in La Jolla. It's got the actual address, the actual price, and all of these details are accurate from the actual website, but it also pulled in the images. If I look at the Zillow listing, this is the actual house. These are the actual images from this listing here, and it incorporated them into this flyer. Like, this was one of the things that probably blew me away the most when I tested it. I didn't know if it was gonna pull in actual images from the flyer and then incorporate them into this image, but it did. So if you're a real estate agent and you need to make one of those flyers that, you know, you you put in the little box outside the house. Your job just got way easier. I want to test it creating an itinerary for me. Create a beautiful three day travel itinerary visual for a first time trip to Tokyo. Decide the best scheduled neighborhoods, activities, and food stops. Do it over three days, morning, afternoon, and evening. Here's what it created for me. Tokyo in three days, a beautiful first time itinerary with iconic neighborhoods, food stops, and easy to follow plans. It actually shows, like, which towns I should visit and which days to do them, and it's pretty visually appealing as well. Now I've never actually been to Tokyo,
15:33so I don't know how accurate this is as far as, like, these are good places to visit or not, but this is something I print out and take with me and could actually follow along to. Like, this is a nice little itinerary to use. I don't know. Just a fun idea if you're planning on going on a trip and you don't wanna make a bunch of plans, you can let AI not only plan it for you, but make something that looks pretty as well for that plan. Here's something I've always tested with Nano Banana, and it never quite worked very well. It was actually placing landmarks
16:02where they're supposed to go. So I told it, create a simple illustrated map of San Diego for a visitor guide, include and label these areas, gave it some parts of town, and it made this map. Is it super accurate? No. But it's not horrible. I mean, Coronado seems to be in the right spot. They did give it an extra bridge that doesn't exist, but it's kind of in the right spot. Downtown's kind of in the right spot. Mission Beach, La Jolla Del Mar, roughly
16:27in the right spot ish. North Park, definitely not in the right spot. Balboa Park, yeah, not really in the right spot. Also takes up way too much of San Diego. So that's an area these AI models still leave a little bit to be desired. These are not quite very accurately located landmarks, but it looks nice. If you're going on vacation, maybe you need a packing list and you want it to be visual. Could a visual packing checklist for a five day family beach vacation, decide the categories and items, and here's the stuff to include. And it made a pretty good checklist, clothing, swimsuits, cover ups, t shirts, tops, shorts, underwear, socks, pajamas, blah blah blah, beach gear, toiletries,
17:06kids items, tech phones, phone chargers, power bank, etcetera, snacks, documents, IDs, wallets, cards, health insurance, reservation, confirmations,
17:16HOOKeasy to forget items, wet bags for swimsuits, zip top bags, laundry bags, stain wipes, aloe vera, nightlight, etcetera. So, I mean, a pretty handy checklist that you could print out and use every time you're about to go on a beach vacation. What about for Disneyland? Create a visual one day Disneyland plan for a family with kids. Basically, design my strategy for me when I go to Disneyland. One day, Disneyland plan, 07:30 to eight, arrive early, get to security before park opening, have tickets ready, eight to 10:30, head first to kid favorite rides with the shortest waits, focus on fantasy land style attractions, one must do ride and a snack break before lines build, 11:30 to 12:30 lunch break, 12:30 to three afternoon, great window for slower attractions, indoor shows, shaded walk throughs, three to 04:30, rest time 04:30 to seven, evening fun 07:30 and later fireworks, ways to avoid meltdowns,
18:08HOOKand reminders for packing. Again, another handy little checklist if you're planning on going to Disneyland. I told you, I'm testing, like, everything. Create a visual wedding weekend schedule for guests. Here's a wedding weekend schedule for guests that if they're coming out to attend your wedding, you can hand them one of these so they know the series of events and when they're supposed to be where. And it looks nice and fancy for a nice fancy wedding. Maybe you have kids. Create a visual weekly chore chart for a family with two kids. Here's our weekly family chore chart. Morning task, make bed, get dressed, brushed teeth, backpack, backpack, evening task, tidy room, homework slash reading, clothes and hamper, brushed teeth. Then a reward tracker to make sure they did it all, and you can check it off, and then they get their allowance or their screen time or whatever you give them for doing chores. Create a printable thirty day habit tracker. Audience, a busy adult trying to build healthier routines. Here's the habit tracker. Drink water, move body, read for ten minutes, sleep seven plus hours, no late snacks. Week one, two, three, and four, wins, challenges, focuses, monthly takeaways. You could print this out and do this on a daily basis. I'd personally probably vibe code like a digital version of this that I can access from my phone and computer, but if you're more of an analog, check things off on paper type person, this could be a good way to stay on top of stuff. Create a visual planning board for a garage organization project. Here you go. Here's your garage organization project planner broken down into five different phases. Assess and declutter, measure and plan zones, clean and prep, install storage, final organize and maintain. You got all your supplies you might need, your estimated timeline. Damn. People are gonna take four weeks to organize their garage. Budget categories,
19:41so storage bench shelving, etcetera, before and after, little, like, mini picture of what it should look like, a master checklist, and some smart tips. Alright. Now I wanna talk about another area I find AI image models to be really helpful, and Nano Banana was getting really, really good at this, but I feel like this is an area where ChatGPT images too is even better, and that is in creating infographics.
20:03Not necessarily because I think the infographics are much more visually pleasing. I actually think Nano Banana is maybe slightly more visually appealing, but the ChatGPT images are way more accurate. Like, they're just getting a lot of the details more right than what Nano Banana was. So for example, create an infographic that explains how AI agents work to nontechnical
20:22audience. We open this up and we have a breakdown of how AI agents work. You give it a tool, it figures out what matters, it makes a step by step plan, it uses tools, it checks that it works, and it gives you the result. And each step, we have a little bit of extra details here. Now, I wasn't super happy with this. I'm like, this is kind of boring. It's just like blue and white. So I literally went back and said, now make it even more colorful and infographic y. And it made it even prettier and more infographic y. I mean, it's all the same details.
20:52It just looks a lot better. But now here's where this gets even cooler. You can actually give it a website and tell it to make an infographic based on what's on that website. So I gave it this void website, which is a GitHub page that explains this video object and interaction deletion. Essentially, it's a tool where you can easily edit out things from like a video here. And I said, turn the information below into a clear visual explainer. Gave us some details, but then I gave it the source,
21:18which is that voidmodel.github.io we were just looking at. It actually looked at the website to figure out what that page was about and then created an infographic to better explain what was on that page. Void removing objects and their effects from video. So the problem, most video object removal methods can fill in a content behind an object and fix appearance artifacts like shadows reflections, but they often fail when the removed object affected other objects such as in collisions. Then it explains what void does, how it works, some extra refinement, the training and results, and then why it matters. So it built an infographic here. And again, all I gave it was turn this information into a clean visual explainer, make it easy for a general audience to understand it under thirty seconds, clear title, five takeaways, simple icons, visual hierarchy, short why it matters box, some style details, and then a link to the URL. So it can just go pull the information from the URL and then create an infographic.
22:11You can do this for anything. You can give it a news article and have it create a visual explainer of the news. You could link it to a tutorial that somebody created and have it give you a visual step by step explainer of that tutorial. All you have to do is feed it the URL and say, make an infographic based on what's at this URL. Pretty slick. Maybe you're making a video or a website and you wanna break down the timeline of something. Create a horizontal timeline graphic titled the evolution of AI image generation. Include these milestones, and I gave it a year by year milestone for AI image generation. It created a nice I mean, fairly simple, but nice timeline here. Or how about create a mind map titled ways creators can use AI? Here's the various branches. Under each branch, add two small sub points. And here's the mind map it created. AI for creators, research, script writing, thumbnails, editing, shorts, sponsorships, newsletters, automation, analytics. I mean, one of the lines kinda broke here, but then under each of these, it goes off and gives two more branches. And that is one of the areas this is so powerful. It's just that extra thinking. You can give it a fairly basic concept, and then it flushes out the idea and makes an even more detailed image with more information in it. Create a process diagram for a small business using AI to handle customer inquiries. Here's the diagram flow and some extra requirements,
23:24and here's our AI customer inquiry workflow. One thing I have been noticing, you might have been noticing this as well, is that it does like the detail to this like blue and white aesthetic if you don't specifically steer it on other colors. But here's our AI customer inquiry workflow. Customer message comes in, AI categorizes inquiries, is it billing, support, sales, booking, AI drafts a response, is there a human review needed? Yes. Human review slash edits? No. The response is automatically sent. Common questions are logged. Business improves FAQ.
23:54We get better answers over time, and eventually, the human review will probably go away because there's enough stuff in that FAQ. Here's something people might find helpful. Create a visual comparison chart for choosing between ChatGPT, Claude, Gemini, and Perplexity. Audience, normal people who are confused about which AI tool to use. And here we go, which AI tool should you use? A simple guide for choosing between ChatGPT, Claude, Gemini, and Perplexity. Chat GPT, best for all around everyday AI help. Claude, best for writing long documents and careful thinking. Gemini, Google users and everyday productivity, Perplexity, fast answers, and web backed research. I'd say for the most part, this is pretty accurate. When it comes to long documents, I'd probably lean more on Gemini than Claude, but I'd say the strengths and weaknesses that it found here, if you wanna pause and read, are pretty good. I I would say it did a pretty good job here. It's also kind of funny that this was done in ChatGPT, and ChatGPT
24:44sort of points out its own limitation, can sound confident when it's wrong. Create a printable cheat sheet for writing better AI image prompts. Here's our AI image prompt cheat sheet. Here's the simple formula, some useful prompt ingredients, common mistakes, being too vague, stuffing in too much, etcetera. Some example prompts, revision tips, quick check before you generate. Now is this something that really needs an image generated for you to create? Not really. But, hey, if you wanna print it out and have something pretty to look at, it does a good job at that. It can actually make, like, mini slide decks for you as well. Create a five slide visual presentation about the shift from AI chatbots to AI agents. We scroll down here, we can see it created five different slides for us. And if I open this up, we can click through the slides from chatbots to agents. What changed? Chatbots react to prompts, return answers, usually stop after one reply. Agents pursue a goal, use tools and memory, take multiple steps. Why it matters saves time, cuts busy work, etcetera. Real world examples, support agents, research agents, ops agents, creator agents. Final takeaway, chatbots talk. Agents do. So we have, like, a little mini slide presentation that we can use if we needed to. I just love now that you can give it one prompt, and it will make multiple slides or carousel images like we looked at earlier all from that initial one prompt. Create a one page sales pitch visual for an automation agency called Mattomation.
26:00Audience, small business owners. Here's my one pager. Automate the busy work, grow the business. Mattomation. Again, defaulted to that white and blue aesthetic. I mean, I don't really mind it. I actually think it looks kinda clean. But if you really want a specific color scheme, you do have to steer that it seems. Maybe you have an ecommerce business. This is actually a really, really great tool for ecommerce businesses. Create a product packaging concept sheet for a high protein snack brand called Wolf Power. We can see it created product
26:29packaging for me, even created the logo for Wolf Power, created our color scheme, the front of pack examples. And if I was building a new CPG brand from scratch, this might be a good place to start brainstorming what your products look like. Maybe you're using, a Shopify or something, and you wanna see what your product will look like in different environments so you can have your little, you know, slideshow at the top of the products in different environments. Create an ecommerce product image set for a futuristic desk lamp called the Lampinator.
26:56Create a grid of six product images, hero product image, desk lifestyle scene, feature callout size slash context image, mood lighting scene, packaging image. And there we go. We've got the sort of hero image, futuristic smart desk lamp. We've got one on a desk with a computer, one with some various details about the product, one with the specs, one with, like, darker mood, one with the actual packaging.
27:20Maybe you wanna sell some merch with, like, your favorite logo on it. Create a merch mock up sheet for a tech creator brand. Brand phrase, thanks for nerding out. Show the phrase on a black shirt, a white hoodie, a sticker, a coffee mug, and a dad hat. And there you go. We've got a black shirt, a white hoodie, the sticker, the mug, and the hat. Maybe I should make these. Those are kinda cool looking. Create a set of five app store screenshot mock ups for productivity app called Taskflow.
27:43So if you're putting an app in the app store, you know how you can kind of scroll through different image screens on the app? Well, you can mock all those up as well. There's one image. There's another. Plan smarter in seconds. Protect your focus time. See progress at a glance, and we can see exactly what this app could look like in the app store. Now you can feed it your own actual screenshots, and it will actually replace this with the real apps image as well. But you can mock it up here first and then dial it in how you want later. Create a website hero section mock up for an AI writing assistant called Draftly. Here's our hero section mockup for a website for Draftly. It put the menu up here. It put our call to action button. It put some screenshots from the image trusted by these brands.
28:26Some more credibility and trust markers here. We've got our headline, another call to action. It looks legit. Right? How about AI course curriculum? Create a visual course curriculum graphic for a course called AI for everyday productivity. Audience beginners who wanna use AI in daily life and work, you decide the module names and structure. Here we go. AI for everyday productivity. We got that color scheme again. A beginner friendly course for using AI in daily life and work. Module one, AI basics made simple. Module two, prompting for better results. Module three, everyday personal tasks. Module four, work and communication. Five, create your AI workflows. Six, trust privacy and good judgment. And then a final project down here. So if you're thinking about designing an online course, well, maybe AI could help you map it all out from the beginning. Create a new client onboarding checklist for a freelance video editor. Here's our client onboarding checklist, project kickoff, the brand assets we need, footage delivery, style reference, timeline, feedback process, and final delivery. Here's one that I kinda like because I tend to write a lot of messy notes. Turn these messy notes into a polished visual action plan. Audience, a creator or entrepreneur who feels overwhelmed, create a one page visual plan that could be printed and kept on a desk. You decide the best title sections, hierarchy, and wording. And then down here, here's the messy notes. I have too many ideas and keep getting stuck deciding what to make. I wanna make more videos, but I overthink the title and thumbnail. I need a simpler system. I wanna keep up with AI news, but also make more evergreen videos. Wanna spend less time researching endlessly and more time recording. I need a daily process that helps me pick one idea, make the packaging,
29:51record, and move on. That's not about anybody in particular for sure. And here's what it created from overwhelmed output, a simple creator action plan for making more videos with less friction. Our main goal, the biggest bottleneck, the priorities, what to stop doing, what to do next, a daily process, and a seven day action plan that I I mean, whoever needs to use this can follow. Here's a fun one. Create a beautiful printable recipe card for high protein breakfast tacos. Include this exact text, high protein breakfast tacos, and then here's all the details about the recipe. And look at that. We've got a nice pretty card that you could print out and save in your recipe book with high protein breakfast tacos, an image of the tacos, the details, the ingredients, and the instructions. And the last thing I wanna show just because it's kind of fun and while this model's really good at it is creating comic pages. Create a one page comic with six panels. The day my AI assistant took over my inbox, And then it's got each panel here, and here's what it looks like. I really like the aesthetic that it does with these comic panels, and we can see how do I have 437
30:50emails. AI, please help me clean up this inbox. And then we can see the AI talking sorted by priority sender and action needed. I also drafted replies for the important ones. Wait. These are actually good. I can breathe again. Okay. That was actually useful. I don't know. The comics look cool to me. They seem to have good color scheme and look like a nice comic. Okay.
31:08CTAThat's what I got for you. I know it was a lot. I wanted to just test, like, everything I can think of, so I threw, like, I don't know, 40 prompts. I kind of lost count. I'll probably count and mention it in the title or on the thumbnail or something. I I don't know how many prompts I did, but it was a lot. And I know there's a lot of gaps and a lot of things I missed. I talked about a few in my AI news video last week where people were doing, like, three sixty images that looked like they were shot with an Insta three sixty camera. Riley Brown did one where he was generating images of books with the barcode, and the barcode was actually scanning and taking you to that book. Like, mind blowing stuff with this. But I wanted to dig a little bit deeper and try a bunch of other random stuff that I didn't do in that video. So if you haven't watched that news video, check it out because I tested even more prompts in that one and talked about some other cool stuff that people were doing with it. But this one, I wanted to go a little bit more into, like, the research and showing that it looks at the web and reads websites for you and then makes images based on what it finds. To me, that's where this gets really cool, really useful, and in my opinion, does slightly better at it than Nano Banana. Now I actually think Nano Banana still generates more, like, realistic images, but this one is definitely the new leader in actually getting information and text into the image. But, again, that's what I got for you. This video was long as my videos tend to be when I start nerding out. So I'm gonna go ahead and wrap this one up. If this is your first time discovering this channel, what I typically do is I break down all the latest AI news, but sometimes in between, I like to make tutorials and break down all the cool use cases of these tools, give you some behind the scenes of my creator business. But my main bread and butter content are the videos I put out every single Friday where I break down all of the news that came out in the AI world that I think most people will wanna know about. There's, like, 300 AI news stories that come out almost every single week. I try to pick the, like, 10 or 15 that I think you most need to know about and that are the most useful to the most amount of people, and I share just those. I try to through all the noise and the hype and show you just the most useful stuff. Those videos come out every single Friday. If that's something that sounds interesting to you, maybe consider liking and subscribing to this channel, and I will make sure those videos show up as well as more breakdowns and tutorials and cool tools and things like that in between the news videos. But, again, my goal is to help and show you what's actually useful that's coming out in the AI world. If that's useful, you know what to do. Thanks again for hanging out with me. I really, really appreciate you, and, uh, hopefully, I'll see you in the next one. Bye bye.
§ · For Joe
Steal the format.
Mod Producer / Killing Excuses playbook
One screen-recording session, one tool, forty narrated prompts — and you've got a 33-minute video that doubles as forty short clips.
- Pick one tool that just shipped, then carpet-bomb it with 30–50 prompts in one sitting. You're filming the test, not preparing it.
- Wolfe's hook is 'you think you know this model — you don't.' Use the same loop on any tool MCN+ ships: 'You've seen Reels Editor demos. You haven't seen these 30.'
- His URL-aware demo (paste Zillow link → get a real flyer) is one 60-second short on its own. Mine your longform for the single moment that made YOU say 'whoa' and ship it as a standalone reel before the longform.
- The video has zero b-roll. The picture-in-picture cam plus the screen-share is the entire production. This is the lowest-friction tutorial format on YouTube — and it's the right one for shipping fast.
- He admits failures on the same axes he praises wins (aspect ratios, maps, content-policy hiccup). Honest negatives build trust faster than positive-only demos.
- End on the CTA template he uses: 'I cut [the noise] down to [the signal] so you don't have to.' That's the LFB pitch — own the filter so the audience doesn't have to.
§ · For You
What this means if you actually want to use ChatGPT to make stuff.
If you've been ignoring the image model
ChatGPT Image 2.0 is no longer for 'pretty pictures' — it's for real, printable documents that used to need a designer.
- Paste a URL in your prompt — your own site, a Zillow listing, a research paper — and ask for a flyer / ad / infographic. ChatGPT will read the page and use the actual content.
- One prompt can produce a 7-slide carousel, a 30-day social calendar, or a 6-panel comic. Stop asking for one image; ask for the whole deliverable.
- If you run a small business: try the menu, the flyer, the packaging concept, and the brand mood board prompts shown in this video. They give you a presentable starting draft in seconds.
- Don't trust it for maps or geographic accuracy. Do trust it for layouts, infographics, comparison charts, and printable plans.
- If the first generation looks generic and blue/white, just tell it to make it more colorful or follow a specific style — iteration works and is fast.
- Brain-dump your messy notes, then ask for 'a polished one-page visual action plan.' This single prompt is worth the price of the subscription.
