⏺ It's happening! GPT-4 is officially launched.

As it was stated earlier, the new version is now multimodal. Now AI understands not only text queries, but also defines what is drawn on pictures (oh, everyone will start to throw memes at it).

GPT-4 can solve more complicated tasks than GPT-3.5, as it is more creative and is able to process more subtle instructions.

The developers have done a lot of tests, and GPT-4 is much better than its predecessor.

Now you can tell the AI in what style it should respond to you. For example, as Socrates, GPT-4 will answer your question with an answering question, encouraging you to think for yourself.

For now, GPT-4 is only available to ChatGPT Plus subscribers, and even then with restrictions.

By the way, there will be a live presentation on YouTube in an hour at this link.

What do you think about the new iteration of GPT?

🟢GPT4 is able to explain memes

Here it understands that the image is made of Nuggets representing a map of the world🤯

It may make you smile but it's pretty damn impressive that an AI can make the connection in this way🫣

Notes from yesterday's presentation of the new GPT-4.

The main things from the presentation:

⏺You can build a bot for Discord in 2 minutes. Neuronka writes its own working code and fixes bugs.

⏺Calculating a married couple's complicated taxes is no longer a problem. It's enough to describe the situation, ask qualifying questions and in some cases you won't need an accountant. GPT-4 also knows how to do math and logic.

⏺Make a website out of a prototype drawn on a napkin. Literally: you draw a design sketch on paper, and the AI will write the code for your site. Then you edit by hand or ask the AI to fill in the page.

⏺The new version easily handles large amounts of information, up to 32,000 characters. It also understands what's going on in the photo, reads documentation in any form, and even turns all the answers into poems.

🟢WebDesign within the reach of anyone?

GPT4 will be able to code a website for you from an ALMOST IGNOBLE drawing made on a piece of paper.🫠

Its ability to analyze images and create code will revolutionize the digital world.✅

(And not only..🫡)

Just imagine what it will look like with a cleaner, more detailed plan.

When you see the MONSTEROUS evolution between GPT3 and GPT4, just imagine what we will be able to do in the coming months and years with AI..🫣

Those who don't admit the power of AI today are simply in DENIAL.

The opportunity of the decade is revealed every day but some will refuse all their life to admit what is happening in front of their eyes since the beginning.🙃

💠Everyone has so much eyes on GPT4...

That no one has seen what Google has announced in terms of AI!

🔻 PaLM API and MakerSuite

🔻 AI in Gmail, Google Docs and workspace🤯

🔻 Generative AI support in Vertex AI

🔻 Generative AI application builder

🔻 Partnerships, programs and resources for each segment of the ecosystem

💠6 prompts for ChatGPT to save time in your email writing within a company:

- Write an email to [recipient] discussing [topic] focusing on [specific points].

- Write an email to [team] addressing [problem or concern] and offering [solution or action].

- Write a follow-up email to [client or partner] after our [recent meeting or event], discussing [key findings].

- Write a monthly report on [department or team] activities, focusing on [goals, accomplishments, and challenges].

- Write an email to [employee] providing feedback on [recent performance or work] and suggesting [areas for improvement].

- Create a sales report for [product or service] in [month or quarter], analyzing [sales performance, trends and recommendations].

🙂It's been a crazy week for AI news!🙂

🔸 Release of GPT4 (More creative, more precise and more scalable than GPT3)

🔸 Google releases PaLM API (their own LLM)

🔸 Google adds generative AI to its Gmail/Google Docs/Sheets suite

🔸 Google announces Med-PaLM 2, a new medical LLM

🔸 Microsoft365 Copilot adds generative AI to emails/docs/teams

🔸 MidjourneyV5 (impressive photorealistic image generation)

🔸 Anthropic launches Claude (their version of chatGPT)

🔸 Stanford announces Alpaca 7B

🔸 Adept raises $350 million

🔸 PyTorch releases version 2.0

🟢Use AI to scan what is around you and transpose it into 3D!✅

That's the promise of poly.cam

Perfect for architects, interior designers, 3D artists and anything else you can imagine!🫣

Just take a video and the app will transpose it into a perfect 3D model

In the video above it's an entire apartment that has been scanned.🫠

But nothing prevents you from using it for small objects for example.🤩