Three important announcements happened in the Generative AI space recently. Let’s take a look at all of them.
OpenAI’s Sora
The first, and the most widely discussed launch was Sora. Sora is a video generator launched by Open AI – the company that powers Chat GPT, Dall-e, and now, Sora. ChatGPT is a text based AI engine, Dall-e generates images based on user prompts, and Sora generates up to 60 second videos based on a text description given by the user.
Introducing Sora, our text-to-video model.
— OpenAI (@OpenAI) February 15, 2024
Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions. https://t.co/7j2JN27M3W
Prompt: “Beautiful, snowy… pic.twitter.com/ruTEWn87vf
In the launch post on February 15, Open AI wrote on X (earlier called Twitter):
Introducing Sora, our text-to-video model. Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions.
Users have, since then, posted very detailed, realistic videos.
You can access Sora here: https://openai.com/sora
NVIDIAs’ RTX
NVIDIA is an American company that is famous for designing and manufacturing GPS (Graphical Processing Units), high-capacity chips, and other components for computers.
On February 13, it announced the launch of RTX.
Create a personalized chatbot with the Chat with RTX tech demo. #AIonRTX
— NVIDIA Studio (@NVIDIAStudio) February 13, 2024
Quickly get tailored info from your own files and content by connecting your data to an LLM on RTX-Powered PCs for local, fast, generative AI.
Download demo here: https://t.co/jcWr2a3lCG pic.twitter.com/qgWbOlnoX5
Here is why RTX is different and merits a mention.
1. RTX is trained on Mistral or Llama 2 – both open source datasets.
2. It runs on the user’s own computer and searches the user’s own files.
3. Since all conversations are on the local computer, nothing generated by the user is shared with a larger dataset, and there is no violation of user privacy.
Think of it as a giant indexing engine on your own computer.
4. Because it runs on a single computer and loads LLMs on the machine, the download size is 35 GB!
5. Many early users have indicated that the software is clunky, hard to use, and crashes.
You can download the RTX here: https://www.nvidia.com/en-us/ai-on-rtx/chat-with-rtx-generative-ai/
Claude: The AI with a Constitution
While Sora and RTX are unique because of their capabilities, Claude is unique because of the way the engine is designed.
Claude is built by a company called Anthropic – a US-based startup founded by Dario Amodei and Daniela Amodei – who are siblings and worked with Open AI earlier.
Claude is based on Constitutional AI – A set of values that define what the AI engine should do in terms of political, violent, and explicit content.
Is a constitution like the conscience of an AI engine?
If we want to think of an AI engine as if it is a human mind, yes, that comes closest to describing the Constitution. But a more accurate description would be an elaborate set of Dos and Don’ts that the engine follows.
Claude has both a free and a paid version. The Paid version allows users to ask detailed questions and get relevant answers.
Claude can be accessed here: https://claude.ai/