ChatGPT for $200 | Weekly Edition

PLUS HOT AI Tools & Tutorials

Dec 06, 2024

Hello and welcome to our weekly roundup!

What a week! I know I say this quite often, but in this case, it feels like Christmas has arrived much earlier. And we already know this is just the beginning.

Let's go to discuss.

This week we wrote two posts that summarize this year and pave the way for 2025:
AI Creators & Founders to Follow in 2025
Stepan Ikaev
·
Dec 5
Read full story
Biggest AI Failures of 2024
Stepan Ikaev
·
Dec 3
Read full story

This Creators’ AI Edition:

Featured Materials 🎟️
News of the week 🌍
Useful tools ⚒️
Weekly Guides 📕
AI Meme of the Week 🤡
AI Tweet of the Week 🐦
(Bonus) Materials 🎁

Your advertisement could be featured here!

Sponsor a spot in our newsletter to connect with subscribers interested in AI, technology, and startups. For inquiries, click the button below.

Featured Material 🎟️

12 Days of OpenAI

Sam Altman has announced that starting December 5, OpenAI will introduce new products or demos every weekday. These will range from minor updates to full-fledged products that can be accessed on the announcement day. There are no details, but that's the trick. OpenAI intentionally creates an expectation of Christmas presents. This event will last for 12 days.

🎄🎅 starting tomorrow at 10 am pacific, we are doing 12 days of openai. each weekday, we will have a livestream with a launch or demo, some big ones and some stocking stuffers. we’ve got some great stuff to share, hope you enjoy! merry christmas.

— Sam Altman

And the first present is already under the Christmas tree (well, for $200/Mo).

New Plan for ChatGPT

ChatGPT Pro interface showing a notification about resolving a C++ deadlock issue with mutexes, featuring code context in the background and a progress bar with a ‘Details’ button.

OpenAI announced ChatGPT Pro. This new $200/mo subscription provides unlimited access to all models, including a full version of its o1 “reasoning” model. This is the highest-performing and highest-quality AI the company offers right now.

According to Altman, this subscription will suit the most advanced users, while simpler tiers will fulfill the needs of the rest. To that end, OpenAI has made the standard o1 version available as part of a paid subscription to ChatGPT.

In case you forgot or missed it, here's a reminder: OpenAI unveiled a preview of o1 in September. It was the first model, which endeavors to check its own work. This makes it one of the most robust on the market (but also quite slow).

Keep your mailbox updated with key knowledge & news from the AI industry

So, o1, included in ChatGPT Pro, is not a preview but a full version. It's better in every way. The model generates responses faster and is more concise in its reasoning. It also supports image processing and is 34% faster at solving complex real-world questions. Users must select “o1 pro mode” in the model selection tool to access the functionality and ask their questions.

If you decide to subscribe to ChatGPT Pro, you should know that you get a nice bonus. This tier includes unlimited access to GPT-4o and Advanced Voice Mode.

Every time it seems like OpenAI is about to cede leadership to Anthropic, Google, xAI, or whoever, Sama comes on the scene and shows who's boss.
There's some Steve Jobs vibe in that.

News Of The Week 🌍

Amazon Unveils New Family of AI Models

Amazon has made a series of equally important announcements. The company introduced a family of multimodal AI models called Nova. They are available on Amazon Bedrock and include several options for different use cases. There are four text generation models in total: Micro, Lite, Pro, and Premier. The first three are already available to AWS customers, with Premier coming in early 2025.

Micro is a model for receiving and outputting text only. It generates responses faster than other members of the family.
Lite is a platform that can handle image, video, and text inputs.
Pro is the most productive model for now. It supports all content and is enterprise-oriented.

The company also showed the Nova Canvas image generation and Nova Reel video generation models. These can also be accessed now.

Amazon picked a lousy week to wow us. It's a bit of a shame. Anyway, the company is in the race to be the best LLM developer.

DeepMind Shows AI Model That Generates Interactive Worlds

DeepMind, Google's AI research organization, also unveiled its model this week. But unlike Amazon, its platform, called Genie 2, is designed to generate interactive three-dimensional worlds. It is the successor to a model released earlier this year and can create interactive scenes from a single image and text description. We've seen something similar recently from startup Decart and its Minecraft simulation AI.

DeepMind says that Genie 2 can generate a huge variety of 3D worlds, including worlds where users can run, jump, and swim using a mouse or keyboard. The video-trained model can simulate object interaction, animation, lighting, physics, reflections, and “NPC” behavior.

DeepMind is my personal favorite this week. AI for creating full-fledged video games seems like the kind of force that could turn an entire industry upside down.

Elon Musk’s xAI Secures $6B

There's never much money for AI startups. Elon Musk's company, xAI, has raised another $6B in equity financing, bringing its total to $12B (the last round occurred this spring). According to the Financial Times, only investors who backed xAI in the previous fundraising round were allowed to participate in this round. Investors who helped fund Musk's acquisition of Twitter reportedly gained access to 25% of xAI's shares. Apparently, the capital raised will be used to develop more advanced Grok models.

I used Grok for a while and concluded that it now performs at the same level as the GPT-4o. Whether Musk will be able to catch up with the newer and better OpenAI models, time will tell.

Sharing is caring! Refer someone who recently started a learning journey in AI. Make them more productive and earn rewards!

Refer a friend

Runway Introduces AI Model Frames

Runway has unveiled “Frames,” its latest base model for image generation. The company says this AI marks a big step forward in stylistic control and visual fidelity. The main draw of the update is the multitude of styles and their consistency. In particular, Frames allows you to create 80's style photos of people almost indistinguishable from reality.

You can see a few examples below and on the company's website.

The Runway team said it is “gradually” rolling out access to Frames inside Gen-3 Alpha and the Runway API. In the coming weeks, everyone will be able to try it out.

Google’s AI Models Now Identify Emotions

And some more models. Google has announced the PaliGemma 2 models, which specialize in image analysis. Specifically, the company says the AI goes beyond object identification and can describe actions, the overall narrative of a scene, and even people's emotions. These models are available at Hugging Face and Kaggle.

TechCrunch talked with experts about PaliGemma 2 and got different opinions. Some are alarmed at the prospect of AI that reads emotions and believe such technology could threaten society. Others are pretty skeptical and think that emotions are arranged too complexly to calculate the mood of a random person.

Microsoft Launches Limited Testing for Copilot Vision

Microsoft has begun testing its new Copilot Vision feature. First introduced in October, Copilot Vision allows Microsoft's AI assistant to see what users see on a web page running in Edge. The user can then ask questions about the text, images, and content being viewed or use it for other tasks. The feature is currently being tested among a limited number of Copilot Pro subscribers in the US.

If you can't wait to experience AI controlling your PC, you can try a similar solution from Anthropic. Here's a tutorial on how to do it:
Claude Can Control Your PC
Stepan Ikaev
·
Oct 24
Read full story

Useful Tools ⚒️

Athina – Build, test, and monitor AI apps and agents

Coval – Simulation & evals to ship delightful voice & chat AI agents

Hypelist – Discover personalized recommendations of everything you love

Plot – Unlock AI-powered consumer insights from social media videos

Pointer – AI editing co-pilot for Google Docs

Conversational AI by ElevenLabs – Build AI agents That Speak

A few days ago, ElevenLabs, which develops the most popular voice AI platform, introduced a new Conversational AI feature. It allows you to create custom voice agents for various applications. You can use this feature to quickly build a voice chatbot for customer service or an interactive game character. Conversational AI currently supports 31 languages and integrates with the most popular models, including Gemini, Claude, and GPT.