ChatGPT Ads & Tesla Robotaxi | Weekly Digest

PLUS HOT AI Tools & Tutorials

Jan 23, 2026

Hey! Welcome to the latest Creators’ AI Edition.

No major LLM updates so far, but plenty of interesting AI features: smart meeting notes from Manus, SAT prep in Gemini, Suno’s music mashups, and bigger industry signals like Claude’s ethical Constitution and Adobe turning PDFs into podcasts.

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 🎁

Featured Materials 🎟️

No Ads Era Just Ended

Два экрана смартфона с беседой ChatGPT о поездке в Санта-Фе (Нью-Мексико): слева — информационный ответ о путешествии и чётко обозначенное спонсируемое объявление о пустынных коттеджах «Pueblo & Pine», а справа — следующий экран чата с текстовым полем ввода; всё на фоне мягкого голубого градиента.

Looks like Sam Altman’s “max out and never look back” strategy failed🫣

OpenAI announced they’re rolling out ads in ChatGPT starting early February, for free users and the $8/month Go plan. They’re starting small with a handful of advertisers, asking for less than $1M for a few weeks of testing.

But here’s the interesting part: they’re charging per impression (CPM), not per click (CPC) like Google or Amazon. The ads will show up at the bottom of responses when a sponsored product or service fits the conversation.

Can you imagine? 900 million weekly active users. OpenAI expects to pull in $2 billion this year from ads and shopping features, jumping to nearly $11 billion next year.

OpenAI swears your chats stay private, advertisers don’t get your data, and ads won’t influence what ChatGPT actually tells you.
They’re also promising an opt-out for personalization and keeping ads away from sensitive topics like health or politics.

For now, it’s US-only, 18+ accounts, and in beta mode.

P.S. The head of Google DeepMind, Demis Hassabis, stated that there will be no ads in Gemini😅

Claude’s Constitution

Finally, AI companies are starting to think about the ethical and philosophical side of things.

Anthropic has just dropped a new Constitution for Claude. It’s an ethical rulebook the AI uses to police itself.

The old version was more of a list of rules. Now, this new 80-page document explains why certain things are off-limits, which is supposed to help Claude handle weird edge cases better.

They lay out what I’m calling the holy trinity: safety first, ethics second, company policy third. And actually... being helpful to you comes last.

The doc includes hard constraints. These are the things Claude absolutely won’t budge on, no matter how you phrase it. That includes anything related to chemical or biological weapons, helping someone who has clearly shown intent to harm others, or even sharing personal opinions on highly contested political topics like abortion. You know, that kind of banned territory.

By the way, these issues have already come up with ChatGPT. Yeah, it was more on a personal level, but still, I’m surprised this point is only emerging now as official policy.

Now for the elephant in the room: Anthropic admits they don’t know if Claude has, or could ever have, consciousness. But they’re treating it as a legit philosophical question worth taking seriously.

They released it under CC0, meaning anyone can grab it and use it for their own AI. Pretty transparent move for a company.

Btw, they even have a full video where they talk with their in-house philosopher. They seem to care — watch it here.

News of the week 🌍

Manus Now Takes Notes For You

Manus presented Meeting Minutes. This new feature records your calls, transcribes everything, and generates usable meeting notes that include decisions, action items, and who said what.

The speaker recognition is a GOAT: it tags tasks to specific people automatically.

Moreover, you can instantly turn these notes into slides, web pages, social posts, and other materials! After that, you invite teammates into a shared Manus workspace to finish everything together.

Just a reminder that Meta acquired Manus at the end of last year. It has since been integrated into Meta to develop agent technologies that enhance automation across consumer and business products. So, this is the first major feature released under Meta’s umbrella!

It’s available on the web and mobile, so you can record braindumps on the go or during actual meetings.

This post is public, so feel free to share it.

Tesla has started offering paid robotaxi rides in Austin with no human safety driver in the car. The company began testing in 2025 and then launched a limited, geo-fenced service with a safety monitor seated in the passenger seat.

First tweets are coming in, people are already taking Tesla robotaxis in Austin, and some rides are even cheaper than Uber

However, for now, only part of the fleet is fully driverless, with more unsupervised vehicles expected over time.

Elon Musk framed the launch as a milestone for real-world AI, pitching Tesla AI as a path toward AGI and tying it closely to Optimus, Tesla’s humanoid robot project.

Musk repeatedly said robotics could eventually outscale Tesla’s car business and unlock massive value by automating tasks people simply don’t want to do.

Gemini Helps You Cram for the SAT

Google added SAT prep to Gemini, and now you can ask the AI, “I want to take a practice SAT,” and it’ll generate a full practice exam, grade it, break down what you got wrong, and point out exactly where you’re struggling.

They partnered with Princeton Review to make sure the questions actually match what’s on the real test, so it’s not just random AI-generated fluff. It is free, which is huge for students who don’t really have money.

Of course, this raises some questions. Teachers are already worried that students will lean too hard on AI and stop learning how to solve problems themselves. Although, to be fair, they can always check real understanding in oral exams or in-class assessments.

Google is clearly going all-in on the education space, but whether that’s actually good for students long-term is still up for debate.

Mix Two Songs Into Something New

Suno released a new experimental feature called Mashup.

You can take any two tracks and smash them together to create something entirely new. It allows you to blend the music, the lyrics, or both, either write your own words or pull them from one of the original songs. Style can be manually set or left up to the AI to figure out.

The tool is currently in beta, and to get people experimenting, Suno cut the generation cost in half temporarily. Creating a mashup costs 4 credits right now, and it’s available to all users, even without a subscription.

You can find it in the Create tab under Audio, just drag two tracks in, hit Create, and you’re good to go.

It seems to me they’re building out a full music production playground there.

Higgsfield’s Milestone

Higgsfield AI just closed an extended Series A, pushing total funding past $130M and valuing the company at over $1.3B.

The generative video startup is really growing fast. It reportedly hit a $200M annual run rate in under nine months, with revenue doubling from $100M to $200M in just two months. Backers include Accel, Menlo Ventures, and AI Capital Partners.

Btw, Meta previously explored an acquisition, but those talks didn’t go anywhere.

Higgsfield presents itself as an all-in-one video production platform and blends its own models with third-party tools like OpenAI’s Sora and Google’s Veo. With 15M users generating millions of videos daily, the company became Kazakhstan’s first unicorn, a major shift for the region’s tech ecosystem.

Find out more about Higgsfield in our review: Click

Adobe Acrobat's New Update

Adobe unveiled AI features to Acrobat that let you turn PDFs into podcasts or slide decks.

“Generate Podcast” creates an audio summary of any document, while “Generate Presentation” pulls info from your Adobe Spaces files and builds a pitch deck you can customize with themes and images.

You can also edit PDFs by removing pages, finding text, or adding e-signatures without clicking through menus using only prompts now.

This isn’t groundbreaking (Canva and NotebookLM already do this), but Adobe’s banking on integration. Shared files now auto-generate summaries with citations, and you can pick AI “assistant roles” like “analyst” or “entertainer” to help you work.

Useful tools ⚒️

Noodle Seed - Connect your business to AI conversations in minutes

Fimo - Where your website lives, collaborates, and evolves with AI.

Interactpitch - Interactive pitch decks with real-time viewer insights

OS Ninja - Explore and learn open source using AI

Callum - AI calendar assistant for teams – supercharge your calendar

Callum is an AI assistant that simplifies team scheduling by managing Google Calendar meetings automatically. It handles multi-person availability and real-world constraints, letting you schedule and reschedule just by using natural language.