- Farpoint Vision
- Posts
- 🧋 GPT-4 Explained & Predictions
🧋 GPT-4 Explained & Predictions
PLUS: How to optimize meetings with AI
Top of the mornin! This is GPTEA, your guide to everything AI. We're like that guide Dora the Explorer always whips out anytime she's lost.
Lo hicimos!
In Today’s Post:
GPT-4 Explained and Our Predictions
Tidbit News: Yamaha invests $100M, world's first vision processor unit, SNAPPLE uses AI for ads
Tools to Make Your Meetings Better
GPT-4 BREAKDOWN & PREDICTIONS
Microsoft Germany is set to announce GPT-4 this week and many are speculating this is going to be another huge milestone event for generative AI.
You'd think after unveiling one of the biggest technological breakthroughs since the iPhone, you'd wait awhile before your next big release. Well, you'd be wrong. If you thought GPT-3 was powerful, GPT-4's juggernaut 100 trillion parameter dataset will make it look like child's play.
GPT-4 Quick Highlights
Multi-modal inputs & outputs: text, images, audio & video
Instant real-time translation. Ask questions in English; answer in French.
Provide answers more quickly and in a more human-like way
MULTIMODAL
Not only is GPT-4 trained on significantly more data than GPT-3, it is also multimodal. This is a fancy way of saying it can take it multiple inputs such as text, images, audio and videos. Where it becomes interesting is if the output is also multimodal, meaning you can take an audio file from a design meeting and have the AI build out a mood board.
Today, ChatGPT only allows you to interact with it via text.
Now imagine asking it to analyze a company's financial situation from pictures of paper financial documents or showing it a picture of a burger you ate and asking it to find the restaurant because you can't remember where you ate it at.
We'll have to wait until the release to see what "modes" GPT-4 will support, but we suspect this'll break the internet again.
GPT-4 TEXT-TO-VIDEO MAY BE LARGEST BREAKTHROUGH
Although Meta and Google have already showcased prototypes, they have yet to be released to the public and first-movers advantage is incredibly important right now.
Imagine being a content creator in 2023 and never using photoshop or iMovie, just ChatGPT.
That's bonkers!!!
On the other hand, Sama is warning people that "OpenAI's newest model is overly hyped and people are begging to be disappointed."
Our Predictions:
Microsoft stock will rise as people become optimistic about Bing gaining more market share from Google
The need to learn new languages may decrease significantly as real-time translation becomes more prevalent
Humanoid robots will be integrated with GPT-4 by the end of 2023. This means they will be able to speak and understand humans in real-time
Ideas, data and people become a differentiator for companies and organizations
There will be a drastic explosion in content creators and content created
Winners in this era will be those who understand how to appeal and communicate to humans regardless of these tools
STEEPED MEMES
🤣🤣
TIDBIT NEWS
Quick scan of the top AI stories
Autobots Rollout Yamaha Motor infuses $100M into robotics and AI growth initiatives
Literally Computer Vision Hailo introduces first AI-centric vision processors for next-gen cameras
AI's Call of Duty Pentagon's historic R&D funding request for developing advanced technology such as AI
Fix Your Own Car General Motors is creating AI assistant for its cars using Open AI's
Red Light Green Light City in India to install 500 AI-based traffic signals
THE MEETING STACK: TOOLS TO MAKE YOUR MEETINGS BETTER
A daily tool stack to optimize a workflow.
1/ Magical - Schedule meetings and hop on calls faster from your browser.
2/ Yous AI - Remove language barriers with real-time AI translation of meetings and calls.
3/ Voice AIÂ - Real-time AI voice changer for team-building, streaming or presentations.
And that's the tea for today! Cheerio till tomorrow, Nick aka. the early bird, Ty aka. the night owl.
Here's your daily dose of cuteness from DALL-E.
Join the conversation