🤖 ChatGPT API released

PLUS: What is Kosmos-1?

GM. Welcome back to Sentient.

In Today’s Email:

  • ChatGPT API released 👋 

  • What is Kosmos-1? 💭 

  • Youtube hint’s at AI tools 👀 

ChatGPT API released 👋 

OpenAI just released APIs for ChatGPT and Whisper.

Developers will now be able to integrate ChatGPT and Whisper models into their own products.

Through system-wide optimizations, there’s been a 90% cost reduction for ChatGPT. The Whisper large-v2 model will also be much faster and cost-effective.

You’ve probably heard of ChatGPT but you might not be familiar with Whisper.

Whisper is a speech-to-text model open-sourced by OpenAI in September 2022. The Whisper API will available through transcriptions (transcribed in the source language) or translations (transcribes into English), and will accept a variety of formats (m4a, mp3, mp4, MPEG, mpga, wav, webm)

Here are some early users of ChatGPT and Whisper APIs:

  • Snap Inc.: recently released ‘My AI’ which is run using the ChatGPT API.

  • Quizlet: has been working with OpenAI for the past three years and has used GPT-3 in multiple use cases (vocabulary learning and practice tests).

  • Instacart: uses ChatGPT + their own AI to help customers discover ideas for open-ended shopping goals.

  • Shop: ChatGPT has been used on Shopify’s consumer app to power its new shopping assistant.

  • Speak: an AI-powered language learning app that uses Whisper API to power a new AI-speaking companion.

The amount of AI products that’ll be built off these APIs will be staggering. 🤯 

Check out the official release here.

What is Kosmos-1? 💭 

Microsoft recently released a research paper, called - Language Is Not All You Need: Aligning perception with Language Models.

The paper introduced a ‘multimodal large language model’ (MLLM) called Kosmos-1. 

What can Kosmos-1 do? A lot.

The model can understand images, text, images with text, OCR (Optical Character Recognition), image captioning, and visual QA (Question Answering).

The model was also used to perform IQ tests.

We’re excited to see the use cases for Kosmos-1 in the future.

Check out the full paper here.

Youtube hint’s at AI tools 👀 

Youtube is likely to join the AI boom soon.

Neal Mohan, CEO of Youtube, wrote his first blog post to the Youtube community yesterday.

The post addressed Youtube’s priorities for 2023 which include: ‘supporting the success of creators, building for the YouTube of the future, and protecting the YouTube community’.

Mohan also went on to mention AI, stating: “And the power of AI is just beginning to emerge in ways that will reinvent video and make the seemingly impossible possible. Creators will be able to expand their storytelling and raise their production value, from virtually swapping outfits to creating a fantastical film setting through AI’s generative capabilities. We’re taking the time to develop these features with thoughtful guardrails. Stay tuned in the coming months as we roll out tools for creators as well as the protections to embrace this technology responsibly.

Looks like the pressure is on. 😈

That’s it for today! We’ll see you back here tomorrow.