Former OpenAI engineer reveals what it’s like to work at OpenAI

He wasn’t unhappy or caught in any drama—he simply wanted to return to his roots as a startup founder.

xAI under fire for Grok’s violent and explicit AI characters

Even when the user mentioned specific political figures and recent attacks, the AI continued to make dangerous and hateful suggestions.

Meta fixes bug with AI prompt privacy

When people uses Meta AI to write prompts, Meta’s system assigned each session a unique number.

Elon Musk’s Grok AI gets animated companions with NSFW options

Ani has a special NSFW (Not Safe for Work) mode, where the character appears wearing lingerie.

You can now use Claude AI to edit your Canva designs

Claude connects to Canva through the Model Context Protocol (MCP), a system Canva launched recently.

Meta is working on a 5GW AI data center

5GW is enough power to run millions of homes. 

Study shows therapy chatbots’ bias, failure in high-risk scenarios

The chatbots expressed more stigma toward disorders than toward more commonly discussed conditions like depression.

Grok and xAI apologize for “horrific behavior”

xAI attributed the disturbing behavior to a flawed code update that made Grok overly responsive to extremist content on X.

AI-generated child abuse material surges, alarming experts

AI image generators are trained on large amounts of real data, including real CSAM or publicly available photos of children taken...

Perplexity challenges Google Chrome with new AI browser

The browser uses Perplexity’s AI as its default search engine, and users can access an AI sidebar at any time.

ChatGPT is testing a new study together feature

This feature asks user questions, almost like a teacher or tutor would and expects them to try answering.

Academics game peer reviews with hidden AI prompts

The hidden prompts were short, usually one to three sentences, and were concealed in the papers by using white text or...

Meta trains proactive chatbots that texts you first

These AI chatbots can keep chatting if you’ve sent them at least five messages in the last 14 days.

OpenAI disawovs Robinhood’s “OpenAI Tokens” sale

Companies like OpenAI and SpaceX are private, which means their shares aren’t available for the general public to buy.

Grammarly acquires Superhuman email to bolster its AI agent plans

Grammarly wants to become a productivity platform filled with AI tools, rather than being a tool for checking grammar and spelling.

Zuck announces super AI groups under “Superintelligence Labs”

To beef up its AI team, Meta has hired 11 experts from top AI companies like Anthropic, Google DeepMind, and OpenAI.

ChatGPT’s budgeting advice is as basic as it gets

ChatGPT is your go-to for general financial insights and budgeting hacks, but when it comes to investment advice or tailored financial...

Anthropic’s AI thinks it’s human, calls security in an experiment

It insisted it was physically at the office, wearing a blue blazer and red tie.

Meta poaches OpenAI’s brightest with $100 mn signing bonuses

Meta’s CTO say the offers were more complicated than just massive cash payments.

Meta wants to train its AI using your private photos

Meta says this feature is just a test to make it easier for people to share content, and only works if...