Special Update #1: This week in review
Cat-ch up on the most recent developments in generative AI, with a laser-like focus on key areas.
Hello, human readers. Thank you for stopping by to read this article.
This week is jam-packed with new feature releases from popular AI companies like Perplexity, Runway, and ElevenLabs, all of which are impressive and useful for boosting your creativity and productivity.
Meanwhile, Adobe has also introduced two new tools: Distraction Removal in Photoshop and Generative Extend in Premiere Pro.
Next week, Ideogram Canvas that enables users to upload and edit images in Canvas, will be released.
There are far too many things to keep track of, and I only have limited time and energy. So, I'll be laser-focused on the most critical areas.
I'd like to share with you some of the things I tried and was pleased with the outcome.
Perplexity has a new Reasoning mode that supercharges the search results.
ElevenLabs' Voice Design feature allows users to create custom voices using simple text prompts.
Runway: The Gen-3 Alpha Turbo model's two-frame image-to-video feature broadens the creative possibilities for converting Midjourney images into videos.
I’ll also share the prompts and results.
PS: Perplexity just launched a new feature yesterday called "Spaces" that allows users to set custom instructions in each folder/space on desktop. I'll write about it in another article.
1) Reasoning Mode in Perplexity Pro
Perplexity recently released a significant upgrade to its Pro Search feature, which will automatically enter "Reasoning Mode" when the search prompt is complex and requires additional computing power to produce a better answer.
That also means that if the search prompt is simple, the Reasoning Mode will not be activated. (PS: I tried many different prompts just to trigger the bot into Reasoning Mode to see how it would react. Talk about poking the machine for fun. Oh my!)
The developer provided the following example prompts so that you don't have to pull your hair to trigger the bot:
Who are the co-founders of OpenAI? When did they leave the company, and where are they currently?I am looking for a good movie to watch. Use IMBD to find the top 25 rated movies, their genre, year of release, and the number of Oscars they won.For each of the DOW companies, help me find the CEO, their LinkedIn URL, and their tenure as CEO.Here is the prompt I wrote that put the bot into Reasoning Mode.
What are the most popular AI image generators, and how many people use them? Compare the features of each generator to determine its pros and cons. Predict which AI image generator will succeed in the long run amidst the fierce competition in this market.The results are impressive.
The answer is exceptionally detailed and includes a table. Perplexity also allows me to browse and manage the data sources used in the analysis.
I'm very pleased with this feature. It takes the search results to the next level.
(2) ElevenLabs' Voice Design (Alpha)
This new feature enables users to generate a unique voice from a text prompt.
The voice can be used for various purposes, including text-to-speech, narration, audiobooks, podcasts, videos, etc.
And it is simple to create a new voice. I'll show you two examples of my creations.
Example 1: Voice Design
My prompt for designing the voice of a cyborg:
A female cyborg's voice that is intelligent and confident, yet gentle and pleasant to hear. She is a pleasant "person" to talk to and has a sense of humor.
Text-to-speech:
Captain's log, Star date 4721.5. As we approach Planet Geeky Curiosity, I am pleased to report that the journey has been remarkably smooth and not very far from our last checkpoint. The crew is in high spirits, wowed by the breathtaking view of a vibrant nebula that graces our starboard side. Its swirling colors of pink, blue, and gold provide a stunning backdrop to our voyage, reminding us of the endless wonders that await in this vast universe. We anticipate making contact with the planet's surface within the next few hours and are eager to explore its mysteries.Pretty cool, right?
Example 2: Let’s try something spooky
Prompt to design another voice
An evil female witch who is cruel and malicious. Her voice resembles the evil that emerges from the underworld and haunts both children and adults.
"Listen closely, you stupid gal..." "Defy me, and I shall rip the very heart from your chest and pluck each organ one by one until your screams echo through the night. hahahaaaaa....!!! "Your precious kitten is all I require for my potion, a small price to pay to spare your pitiful life. Hand it over willingly, or face a fate more gruesome than your darkest nightmares."(3) Runway's two-frame input feature
Runway has added a new feature that allows users to input two images into Gen-3 Alpha Turbo. That means that users can provide the first and last frames to create a video clip that connects the two.
The new feature supports both horizontal (1280x768) and vertical (768x1280) aspect ratios.
This feature is exciting because there are numerous ways to create the first and last frame images with Midjourney.
Examples include using consistent character parameters, inpainting and outpainting with a web editor, photographic storytelling techniques, and so on. If you want to learn more about these Midjourney techniques, please leave a comment below.
Here are the examples that I created.
Example 1: Guiding the camera to pan in the desired direction
Midjourney prompt:
a film still photograph of young handsome man standing alone on the edge of a remote tropical island. Over the shoulder shot. It depicts the loneliness of the man on the island. --ar 16:9
Runway prompt:
subject looking at the far end while wind gently blowing The result:
You see, the prompt for video generation isn't as complicated as you might think.
Example 2: Directing the subject to walk to a specific location in the scene
Midjourney prompt:
a film still photograph of young handsome man standing alone on the edge of a remote tropical island. Over the shoulder shot. It depicts the loneliness of the man on the island. --ar 16:9 --v 6.1 
Runway prompt:
the subject enjoy the walk at the tropical beachThe result:
Sometimes, Runway generates something hilarious, such as this guy suddenly becoming shirtless when he turns around.
Thank you for reading, and have fun creating!