Some researchers say GPT-4o’s Chinese token-training data is polluted by spam and porn websites, likely due to inadequate data cleaning (Zeyi Yang/MIT Technology Review)

Zeyi Yang / MIT Technology Review:
Some researchers say GPT-4o’s Chinese token-training data is polluted by spam and porn websites, likely due to inadequate data cleaning  —  Soon after OpenAI released GPT-4o on Monday, May 13, some Chinese speakers started to notice something seemed off about this newest version of the chatbot …

OpenAI has an unusual, extremely restrictive off-boarding agreement with a lifelong nondisparagement commitment; those who don’t sign it lose all vested equity (Kelsey Piper/Vox)

Kelsey Piper / Vox:
OpenAI has an unusual, extremely restrictive off-boarding agreement with a lifelong nondisparagement commitment; those who don’t sign it lose all vested equity  —  Why is OpenAI’s superalignment team imploding?  —  On Monday, OpenAI announced exciting new product news: ChatGPT can now talk like a human.

Apple limits the development and testing of third-party browser engines to devices physically located in the EU, forcing browser makers to have staff in the EU (Thomas Claburn/The Register)

Thomas Claburn / The Register:
Apple limits the development and testing of third-party browser engines to devices physically located in the EU, forcing browser makers to have staff in the EU  —  Rival coders must have EU staff to build and test non-WebKit surfing  —  Apple’s grudging accommodation of European law …

Source: the Superalignment team was promised 20% of OpenAI’s compute resources but requests for a fraction of that were often denied (Kyle Wiggers/TechCrunch)

Kyle Wiggers / TechCrunch:
Source: the Superalignment team was promised 20% of OpenAI’s compute resources but requests for a fraction of that were often denied  —  OpenAI’s Superalignment team, responsible for developing ways to govern and steer “superintelligent” AI systems was promised 20% of the company’s compute resources …