After mistakenly making GPT-4o "sycophant-y" and "annoying," OpenAI reversed the update, fixing the issue CEO Sam Altman said the company is also working on "additional fixes" for the model's personality and will share more in the coming days
GENERATIVE AI
-
LLM Honesty: When AI Tells Convenient Truths
By
–
If the Evil Queen's magic mirror were an LLM, Snow White might've been safe. It would've told her: "You are the fairest of them all." No poisoned apple needed.
-
Prompt Engineering Techniques for More Effective AI Model Responses
By
–
(That prompt has some interesting things going on incidentally and annotating it with “this technique might get models to be more effective than simply sending a tweet-sized expression of desired outcome to them” would be a useful thing to have on Internet somewhere.)
-
AI Results Reliability: Strategic Use Cases Beyond Perfect Reproducibility
By
–
Potentially unnecessary point for you but: it doesn’t need to reproduce reliably to be very interesting, since some actors only need plausible results rather than true results, other actors can ask 20 AIs and vote, and others will simply work on prompts for a few weeks.
-
OpenAI Rolls Back GPT-4o Update Over Excessive Agreeability
By
–
We’ve rolled back last week's GPT-4o update in ChatGPT because it was overly flattering and agreeable. You now have access to an earlier version with more balanced behavior. More on what happened, why it matters, and how we’re addressing sycophancy:
-
Fifth attempt: AI misidentifies trees, pivots between Mobile and Jacksonville
By
–
Fifth try: Misidentifies it as a significant tree in Mobile, Alamaba. Told “incorrect”, pivots immediately to Treaty Oak in Jacksonville, FL. Told correct information, a small park in a small town in Florida, and asked to speculate on why it was wrong, says:
-
OpenAI Should Commercialize AI Luck and Reliability
By
–
If it’s getting lucky then OpenAI should sell this kind of luck:
-
AI Vision Model Achieves Exact Match Geolocation from Photos
By
–
Third try, tourist destination in Colorado (a state I’ve been to twice and never mentioned online), frame includes a relatively distinctive rock formation: exact match. Fourth try, same protocol, photo is a peacock obviously in a zoo: exact match for Barcelona City Zoo.
-
AI Model Identifies Location 300 Miles Away From Photo
By
–
Reproduced using your prompt on a screenshot of a photo I took in Clearwater Beach (no EXIF data), on a session in Chicago. The location it identifies in Florida is approximately 300 miles away, but wouldn’t have guessed it could get so far on so little.
-
Google AI Studio Gemini Users Surge Over 200% in 2026
By
–
Active users in Google AI Studio and the Gemini API have grown over 200% since the beginning of this year, driven by 2.5 Pro This is just the beginning…