📊 “Guys, sorry for this chart crime”: OpenAI employee apologizes for misleading GPT-5 graphics

[cryptoenthusiast]

During the GPT-5 reveal, some presentation slides raised eyebrows — and not for the right reasons.

🧠 Attendees and social media users noticed that in a few charts, GPT-5's lower performance bars appeared taller than those of previous models like GPT-4o and o3 — flipping data logic on its head.

“Who made these graphs? Are scientists or marketers running OpenAI now?”

The backlash led to an apology from OpenAI marketing employee Pranav Deshpande, who joined the company in 2024. He addressed the issue on X, writing:

“Guys, sorry for the unintentional chart crime. The graphics have now been fixed in the press release.”

Too late for the internet though — memes were already flying. Users joked that ChatGPT must’ve generated the charts, while others asked Gemini or GPT-5 itself to fix them.

TL;DR: Bad charts can’t go unnoticed — even at OpenAI.

#GPT5 #OpenAI #ChartCrime #AIhumor #TechFails #GPT4o #DataVizFails #PranavDeshpande #AInews