Cursor 2.0 and Sora-2 Update | Weekly digest

PLUS HOT AI Tools & Tutorials

Oct 31, 2025

Hi! Welcome to the latest Creators’ AI Edition.

This week took the foot off the gas with big releases, but still brought a lot of updates. Cursor 2.0 is here, and everyone seems very satisfied with the update. Sora 2 is picking up steam, and Musk delivered on his promise.

But let’s get everything in order.

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

Meet Averi: The AI Marketing Workspace

Tired of copying content between ChatGPT, Google Docs, Figma, and email threads?

Averi is where your entire marketing flow happens… so you can Think, Create, Execute & Scale, all in one place.

Brand Core stores your voice and guidelines

Averi AI strategizes with marketing-trained AI built to scale your business

/create mode discuss, draft, and edit all in one canvas - turning your guidance into deployment-ready blogs, emails, ads, strategies, and social posts

Libraries organize everything and make each project smarter than the last.

Want to build your team? Get matched by AI to vetted marketing specialists and collaborate directly in the workspace.

It’s your marketing workspace. Your way.

Enter Your Workspace

Featured Materials 🎟️

Cursor has been updated to version 2.0

Cursor’s latest platform introduces some updates. The headline is the new Composer model, a frontier model 4 times faster than similar models. It is capable of completing most conversational coding turns in under 30 seconds.

The platform is turning into an agent on steroids, running many AI agents side-by-side at the same time to crush tough problems better. The interface got a glow-up, ditching files for agents so you just focus on what matters while AI handles the gritty details, but you can still jump into the classic IDE if you want.

Everyone can use the cursor, and everyone can ease their routine with it. To do this, we are preparing an overview and REAL-LIFE CASES of how you can use cursor together with Claude code in everyday life. Stay tuned!

Sora 2 without invitation?

OpenAI has officially unveiled an update to Sora, which is now available without the need to enter an invitation code, BUT there is a catch: it works only in the US, Canada, Japan, and Korea, and only for a limited time.

With this change, you can turn almost any object, such as pets, favorite objects, or personal items, into an AI video. It also brings small improvements to video editing — you can now stitch multiple clips together to create multi-scene videos. The app introduces a kind of moodboard view, showcasing the most remixed videos and the most featured users and characters.

As we have an opportunity to give a name to the character and tag it whenever we want it to appear in a video, this opens up several scenarios for entrepreneurs and creators: creating branded storytelling featuring company mascots and products, and going deeper into personalized marketing for different audiences.

News of the week 🌍

ChatGPT Creates a “Group Chat”

ChatGPT released a Shared Project for all users. Now you create a project, invite your teammates or friends, and everyone works in the same space. Noticed? The timing aligns with OpenAI’s broader push into workplace tools; they are really into building the system.

Any participant can take someone else’s message and take the conversation to a separate thread without breaking the main line.

Grokipedia

Elon Musk presented Grokipedia, an online encyclopedia where Grok is doing the whole fact-checking. Elon Musk is preparing the ground to make the internet free of censorship and with true freedom of speech, confident that Grokipedia is better than the usual Wikipedia.

There are currently 885,000 materials in English on Grokipedia (the original Wikipedia contains more than 7 million English-language pages). You also can’t create new articles. Probably, administrators are still deciding for themselves which topics should be covered on the site.

Adobe Goes All-In on AI

Adobe has released AI assistants for two of its Creative Cloud products, Express and Photoshop. They are designed to help users create and edit images.

First of all, compared to other sidebar-style interfaces, Adobe built a separate AI mode inside Express. In Photoshop, you can now use not only Adobe’s own Firefly model but also third-party AI models like Google’s Gemini 2.5 Flash and Black Forest Labs’ Flux.1 Kontext. Among other updates is the Firefly Image 5 model that now generates 4MP images and handles human details with much more precision.

New Open Safety Reasoning Models

With all public scrutiny, OpenAI’s been catching for scaling too fast and running a red light with safety, they just dropped open safety models, gpt-oss-safeguard (120b and 20b) under Apache 2.0.

Instead of baked-in rules you can’t touch, the model reads your safety policy on the fly. This means AI developers using OpenAI’s new model can set up their own specific safety framework, not OpenAI.

Two big wins:

Transparency - model shows its reasoning through chain-of-thought.
Flexibility - tweak your rules whenever, without retraining the whole thing.

Bottom line: Devs can finally customize their safety setup for their actual use case instead of wrestling with one-size-fits-all platform restrictions.

New Hailuo 2.3

MiniMax has come with a new version Hailuo AI 2.3. The update brings improvements in realistic movement, facial expressions, and physics simulation. The main feature is emotions. The model’s actually trying to capture facial expressions, mood shifts, and how characters react to different things.

Two modes:

Hailuo 2.3 Mode - cinematic quality with pro-level detail
Hailuo 2.3 Fast Mode - optimized for speed, giving you a complete 1080p video in seconds.

In general, complex body movements are way more fluid and natural now, even with moving cameras. Better stylization support for anime, illustration, ink wash painting, and game CG.