Rattibha

I wouldn't blame you for missing all these AI releases since Google I/O 🤯
AirTable AI
Meta AI Ads
Coca-Cola AI Ad
Stable Animation
ScaleAI War AI Platform
Anthropic 100K Input LLM
HuggingFace Transformers Agent
ChatGPT Plugins & Web: No Waitlist 🥳
Let's get you caught up:

1. Airtable AI
AirTable, the no-code leader last valued at $11B, announced it's entering the AI race.
It's going to allow companies to build in AI to their apps easily.
In this demo, a Marketer creates a campaign brief in seconds:

Out this summer, @airtable AI is going to enable AI not just for marketers, but also:
Product Managers
Sales Reps
HR

2. Meta: Generative AI for Ads
Meta made >$100B in advertising revenue last year.
It has finally *truly* entered the generative AI race with its main product, ads. It has 3 features to start:
AI variation on text
Image outcropping
Background generation

Meta already had one of the top AI research divisions.
It's great to see Meta apply it to this real problem: creating great ads.
It's a better direction than the Metaverse, as @maxwellfinn points out.

3. Coca-Cola released an AI Ad
The first major brand to use AI tools to create a video ad is... Coca-Cola?
They created this amazing video, which is one of the absolute best AI videos we have seen yet.

The ad uses Stable Diffusion created by StabilityAI.
Stable Diffusion can be used to create amazing deepfakes like this one that came out this week De-Aging Harrison Ford
(by Jim Derks on Reddit):

4. ScaleAI: Donovan War AI
ScaleAI, last valued over $7B, is working with the White House, US Army, and Air Force.
They deployed an LLM on a classified network to help warfighters.

ScaleAI has been in the industry for 7 years.
The difference between ScaleAI and Palantir or other competitors is the product is actually live.

5. StabilityAI's: Text-to-Animation
StabilityAI, the company powering the Coca-Cola Ad (number 3), ALSO released Stable Animation.
It's a ground-breaking new text to video tool with unlimited length (unlike Runway).

StabilityAI is reportedly seeking fundraising at a $4B valuation.
Artists like @o_o_ftw are already pushing the limits with the technology to create amazing videos like this:

6. Anthropic's 100K ChatGPT Killer
"That input is too long" ring a bell? The largest version of GPT-4 has a 32K input window.
The new Claude from Anthropic, which has a $4.1B valuation, can take 100K tokens.

Claude could analyze the entirety of 1984 by George Orwell in a few seconds.
In this demo by @preangelleo, it creates a transcript of, and summarizes, a 3 hour long Lex Fridman podcast almost instantly.

7. HuggingFace Transformers Agent
HuggingFace, which has a >$2B valuation, released an API to accept natural language on top of multimodal transformers.
These agents enable coding LLMs to compose other models on the fly to solve multimodal tasks. It's AutoGPT on steroids.

Here we can see @1littlecoder use it it to easily handle text to image
(just one of its many transformative potential modalities):

8. No Waitlist for ChatGPT Plugins & Browsing
And the most important news of the week: OpenAI is releasing Plugins & Browsing to all ChatGPT Plus users (in phases).
You go to Settings > Beta Features to enable them.

This is OpenAI's response to Google's numerous AI launches at I/O:
Bard Tools → ChatGPT Plugins
Ability to Code → Code Interpreter
Free Access to the Web → Browsing
The AI wars between giants like OpenAI/Microsoft and Google are just heating up 🔥