Hey everyone! Remember how I was raving about AI a while back? Well, I've been diving deeper into its cool tricks, from research assistant to article writer, and now – get this – image creator!
If you've been following my stuff, you know I'm a big fan of AI and using AI-generated images in articles. They're way cooler than stock photos because they perfectly capture the vibe of what you're writing about. The only catch? You gotta craft the perfect prompt to tell the AI exactly what you want.
Here's the thing: I stink at prompts. Seriously, some of my early attempts were, well, let's just say they wouldn't win any awards. But hey, that's the beauty of AI – there's always another one to help you out!
My secret weapon? Chat GPT or Google Gemini. I don't play favorites – I love exploring all the amazing things AI can do. So, I basically ask them to be my AI prompt whisperers, crafting the perfect message for the image-generating AI.
For instance, in my last article Mother Nature Threw Bali a Curveball Lately, I hit up Google Gemini for help. I wanted an image of a woman battling the monsoon, umbrella barely holding on for dear life. Here's what I typed:
Act as DALL E prompt engineer, create the best DALL E promot to create image of a woman carrying an umbrella in the midst of heavy rain and strong winds that almost blow away her umbrella against the background of Denpasar cityscape in a manga style.
and
Act as DALL E prompt engineer, create the best DALL E promot to create image of a woman carrying an umbrella in the midst of heavy rain and strong winds that almost blow away her umbrella against the background of Denpasar cityscape in a manga style.
And Gemini, bless its AI heart, came back with two killer prompts. With those prompts in hand, I went AI-hopping, trying them out on Leonardo AI and Nightcafe. Nightcafe's results were perfect for the article, but it just goes to show – the more detailed your prompt, the better the AI understands your vision.
Speaking of prompts, while writing this article, I couldn't resist feeding Gemini's prompt into Bing Designer (curiosity is a powerful force!). Check out the wild result!
Prompt 1 based on Google Gemini respond:
A dynamic manga panel of a beautiful young woman determinedly walking through a bustling Denpasar cityscape amidst heavy rain and strong winds. She clutches a traditional Balinese woven umbrella (payung) that billows and threatens to be ripped from her grasp by the powerful gusts. Rain pours down in sheets, blurring the vibrant storefronts and colorful street decorations in the background. Despite the harsh weather, the woman has a look of steely resolve on her face, her eyes narrowed against the rain.
Prompt 2 based on Google Gemini respond:
A photorealistic image from behind of a woman battling strong winds and heavy rain. Her figure is silhouetted against a dramatic cityscape, skyscrapers shrouded in a veil of rain. She clutches a black umbrella, its canopy billowing outwards almost turning inside out due to the powerful gusts. Rain lashes down in sheets, blurring the neon lights and bustling activity on the streets below. The woman leans into the wind, her posture tense as she tries to hold onto her only protection.
Which one do you like the most? Have you tried using the latest AI tools? Let me know in the comments – I'd love to hear about your experiences!
So you asked AI to ask AI to create an image for you? Very strange.
I guess the next step is to ask AI to ask AI to ask AI to create an image for you.
By the way, it gave you a woman with a parasol - a sun shade. When would that flimsy thing ever be used in the rain? Probably never... and definitely not in a monsoon!
Haha, you're right, it might be a bit of a roundabout way! AI assistants are still under development, so I'm learning how to best utilize their capabilities.
Yes, sometimes the AI generates images that are not what we expect. So we need to make some adjustments to the prompts we use.
I think if it's just for fun, the results are good enough. It's a different matter if the generating image will be used for a serious project.
I wonder if the "mistakes" are done on purpose. Surely it knows humans have 5 fingers on each hand, for example. I've never worked with any such programs, so my understanding is narrow.
I don't think the mistake was intentional, given that the AI is still in its early days. In general, almost all text to image generators, even midjourney, which is said to be the best text to image generator, can still produce imperfect human images. Especially in the hand part.
From my experience of using AI to generate images, there have been many improvements towards perfection compared to when the AI was first released.
Improvements are likely to be very rapid, in fact, exponential. It won't be long at all before HD video is instantly created with a simple prompt. And then, well, perhaps it will be able to simulate reality and consciousness, etc.
Hi @ekads, great news! Your content was selected by curators @ten-years-before, @nalexadre to receive a special curation from BeBlurt 🎉 Don't hesitate to upvote this comment as the curators will receive 80% of the rewards for their involvement.
👉 What can you expect for 2024 on Blurt?
Win 85% rewards with myDelegation
Upvoted. Thank You for sending some of your rewards to @null. Read my last posts to make sure that BLURT burning is profitable for you. Before using this bot please make sure your account has at least 100 BP. Get more BLURT:
@ mariuszkarowski/how-to-get-automatic-upvote-from-my-accounts@ blurtbooster/blurt-booster-introduction-rules-and-guidelines-1699999662965@ nalexadre/blurt-nexus-creating-an-affiliate-account-1700008765859@ kryptodenno - win BLURT POWER delegationNote: This bot will not vote on AI-generated content
** Your post has been upvoted (14.97 %) **
Curation Trail is Open!
Join Trail Here
Delegate more BP for bigger Upvote + Daily BLURT 😉
Delegate BP Here
Upvote
https://blurtblock.herokuapp.com/blurt/upvote
Thank you 🙂 @tomoyan