AI news from Ars Technica

• news
• newsfeed
• artificial intelligence

Tweets by arstechnica

1. OpenAI’s new ChatGPT image generator makes faking photos easy

   New GPT Image 1.5 allows more detailed conversational image editing, for better or worse.

2. Browser extensions with 8 million users collect extended AI conversations

   The extensions, available for Chromium browsers, harvest full AI conversations over months.

3. Merriam-Webster’s word of the year delivers a dismissive verdict on junk AI content

   Dictionary codifies the term that took hold in 2024 for low-quality AI-generated content.

4. Microsoft will finally kill obsolete cipher that has wreaked decades of havoc

   The weak RC4 for administrative authentication has been a hacker holy grail for decades.

5. Roomba maker iRobot swept into bankruptcy

   Shenzhen-based Picea Robotics, its lender and primary supplier, will acquire all of iRobot’s shares.

6. OpenAI built an AI coding agent and uses it to improve the agent itself

   "The vast majority of Codex is built by Codex," OpenAI told us about its new AI coding agent writing code.

7. OpenAI releases GPT-5.2 after “code red” Google threat alert

   Company claims new AI model tops Gemini and matches humans on 70% of work tasks.

8. Disney invests $1 billion in OpenAI, licenses 200 characters for AI video app Sora

   Three-year deal lets users create AI videos of Mickey Mouse, Darth Vader, and more.

9. Oracle shares slide on $15B increase in data center spending

   Company raises its capital expenditure forecast as it doubles down on AI infrastructure bet.

10. A new open-weights AI coding model is closing in on proprietary options

    Devstral 2 model scores 72% on industry benchmark, nearing proprietary rivals.