My Research Into AI
For Idea Generation:
Google’s Gemini (replacing Bard)
The latest AI model by Google is designed to communicate like a human – understanding what you ask and crafting responses similar to how a knowledgeable person might. The more information and tasks Gemini does, the better it gets, constantly refining its language abilities.
You give Gemini a question or instruction (like "write a script about Karachi's water crisis"). Gemini processes this input using what's called a neural network, a super complicated web of calculations designed to mimic how our brains work.
Think of Gemini as having read millions of books and articles. It searches for patterns in all that knowledge, comparing your prompt to things it's "learned." Gemini doesn't just copy an existing text. It predicts, word by word, the most suitable response based on what it "knows" and what you ask for. The answer you get might not be perfect the first time. With feedback and more instructions, Gemini refines its output, aiming to keep getting better each time.
Gemini is an impressive tool, but it doesn't think or feel like a human. It's simply manipulating information. It is trained on data from the real world, which includes both good and bad information. It's crucial to recognize that what it generates may sometimes reflect that bias.
Gemini is constantly being improved by Google researchers, helping it become more helpful, safe, and unbiased.
The capabilities of multimodal AI | Gemini Demo
Shortly AI
ShortlyAI can definitely be a powerful tool to battle writer's block, but whether it's the absolute "best" depends on your needs and preferences.
Key Features
- Unstructured Text Editor: It's almost like working in a clean word processor, but with the AI writing assistant always waiting for your instructions. This provides the freedom to create anything.
- "/commands": This is where ShortlyAI shines. Slash commands let you directly tell the AI what to do (example: "/rewrite this as a haiku" or "/instruct write a conclusion"). It's conversational and powerful.
- Content Controls: Three basic sliders on the side let you set how much the AI should stick to your writing or get creative:
- "A Little": For minor rephrasing and ensuring factual consistency
- "Somewhere in Between": Gives it room for interesting tangents and variations.
- "A Lot": Unleash the AI's imaginative, slightly chaotic side for unexpected inspiration.
- Article Brief: A section on the left of the editor to capture essential keywords and your broad purpose for the piece. Helps provide some focus throughout the writing process.
- "/expand" and "/shorten": Quickly elaborate on or condense an existing section of text. Useful for editing and restructuring as you go.
- Unlimited Documents & WordCount: Unlike some services restricting projects, ShortlyAI's pricing is only based on how much you generate each month. Great for longer projects!
ShortlyAI Tutorial (Episode 1) - AI Powered Writing Assistant - (GPT-3)
For Image Generation:
Adobe Firefly
Firefly is Adobe's exciting new project focused on generative AI capabilities similar to what you see in DALL-E or Midjourney. Adobe Firefly is a standalone generative AI web application available at firefly.adobe.com
It offers new ways to ideate, create, and communicate while significantly improving creative workflows using generative AI. In addition to the Firefly web app, Adobe also has the broader Firefly family of creative generative AI models, along with features powered by Firefly in Adobe’s flagship apps and Adobe Stock.
Key Features
- Image Generation & Manipulation: Create images from text prompts, modify existing artwork, generate variations, and seamlessly expand composition space.
- Artistic Styling: Easily experiment with styles, emulating artistic movements or creating intricate graphic treatments with simple text instructions.
- 3D Material Generation: Produce photorealistic materials and textures for 3D applications and visual effects projects.
- Text Effects: Design eye-catching typography with complex effects based on descriptive prompts.
- Vector Designs: Quickly create precise vector assets (icons, graphics, patterns) from detailed instructions.
Adobe Firefly: Out of Beta & Ready For Your Imagination
Storyboarder.ai
Storyboarder.ai transforms a basic script (film, video, even presentation) into a simple visual storyboard, automatically creating images for each scene. Focuses on speed, not fancy art tools. You upload your script, it turns it into visual panels. Also generates a shot list, breakdown, and basic timeline to jumpstart pre-production planning.
Key Features
Script Upload: You can write a basic script directly, paste existing text, or upload formats like .fdx or .fountain.
AI Visualization: It analyses your script for action, location, characters, etc., and generates basic representative drawings for each scene.
Manual Review & Edits: Storyboards aren't perfect. You can adjust, re-generate specific blocks, and tweak descriptions for better results.
Storyboarder.ai - How-to-Use - Video
For Video and Audio Editing:
Descript
Speech is automatically transcribed with high accuracy, transforming words into your editing timeline. Manipulate content (cuts, reordering, word replacement) by modifying the transcript, streamlining edits without complex audio/video interfaces.
Key Features
- Text-Based Editing: Adjust timings, remove filler words, or rephrase spoken content entirely with written edits reflected instantly in your audio or video project.
- Overdub (AI-Powered Voice Editing): Descript's advanced Overdub function uses your voice profile to synthesise flawless spoken audio corrections or edits, minimising the need for re-recording sessions.
- Screen Recording & Editing: Effortlessly capture screen recordings and apply text-based edits. Ideal for tutorials, educational materials, or software demos.
- Collaboration Tools: Supports real-time comments, versioning, and seamless project sharing to empower team workflows The new Descript: A new way to make video. A better way.
Firecut AI
FireCut is an online video editor and a plug-in for Adobe Premiere Pro specifically designed to make editing simple and painless, relying on powerful AI tools to automate some of the most tedious tasks.
Key Features
- Silence Removal: One-click and its advanced AI detects pauses and filler words ("um","like") with surprising accuracy, offering options to automatically cut them or review individually. Saves hours of manual editing.
- J-Cuts: Automatically overlaps your audio and video in cuts to make speech edits flow smoother. A small detail, but saves you fiddly timeline work.
- Caption Generation: Create and customise captions in 50+ languages. Great for social media, subtitles, and accessibility.
- Auto-Zoom: Create dynamic "talking head" style zooms to keep things visually engaging. It analyses content to figure out where to focus.
- B-Roll Search (Beta): Suggests stock footage relevant to your content – helps spice up simple videos without going on a hunt for assets yourself.
FireCut: Your Lightning-Fast AI Video Editor (w/ @aliabdaal)
AutoPod
AutoPod works directly within Adobe Premiere Pro to streamline and automate many essential podcast editing tasks. It's primarily aimed at video podcasts or interviews, where clear audio and polished presentation are crucial.
Key Features
- Automatic Multi-Camera Editing: Analyses your audio to determine who is speaking, seamlessly switching between the corresponding camera angles during editing.
- Silence Removal: Utilises AI to locate and remove unnecessary pauses, "uhms," and background noise, improving the flow of your podcast.
- Jump Cut Creation: Quickly generates jump cuts (often used in social media clips) from your chosen footage to maintain engaging pacing.
- Social Clip Creation: Exports automatically resized versions (like 1:1, 16:9, and portrait) optimised for different social media platforms, saving creators time on manual re-formatting.
- Endpages and Watermarking: Offers optional automatic insertion of watermarks, custom intros/outros, and endpages to polish your content.