Google Bard’s New Visual Feature is a Game Changer
“Bard Can Recognize Images. 4 Business Use Cases 
|
|
|
| Bard is like a college dropout (& not the startup founder-type): bad at writing, coding, and overall reasoning. |
| But it does have one redeeming feature: image recognition (upload a photo → Bard describes the image). |
| This can be particularly helpful for sorting or categorizing any physical items. Here are a few neat applications: |
- Classifying scanned invoices.
- Evaluating images of products or real estate.
- Documenting meeting notes.
- Transcribing text from actual books.
|
| For now, Bard only supports image uploads from your phone. And much like a college dropout: Bard overpromises. It describes images wrong. A lot. |
| Check out how badly it does handling simple children’s images. |
| Why it matters: Bard’s image capabilities are impressive, but they appear to do nothing more than a reverse image search (locating similar visuals on the internet). We only recommend it for non-work-related tasks as of now.” |
“Meta Announced CM3leon, A ‘First-Of-Its Kind’ Image AI 
|
|
|
| Midjourney generates photorealistic images. |
| Bard describes what’s in an image. |
| CM3leon can do both. |
| This new foundational AI model from Meta can do both text-to-image and image-to-text. It’s like LeBron James: insane at offense and defense. |
| This means CM3leon (pronounced “chameleon”) has the capabilities to: |
- generate images (level sits between Midjourney & DALL-E).
- edit photos with text prompts (“alter the color”).
- place objects with precision (“put sink at coordinates 175, 47”).
- create captions and answer questions about its images.
|
| Why it matters: Meta keeps dishing out innovative AI tech. But so far, they haven’t been commercialized (you and I can’t use them). |
| But at the end of the day, Meta is a social media business, and we’re confident that many of its models will find their place in media creation. |
| CM3leon’s superpower is that it’s multimodal: it’s proficient in text and image. This could be a big unlock for advertisers, marketers, and/or journalists to generate images and content under one umbrella.” |
0 Responses
Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.