‘Goodbye, GPT-4. you kicked off a revolution’ says Sam Altman as OpenAI rolls again newest replace because of person complaints

After confirming that OpenAI could be rolling again the newest updates to its flagship ChatGPT chatbot following person complaints of it being “annoying”, CEO Sam Altman on Might 1 stated goodbye.

“goodbye, GPT-4. you kicked off a revolution. we’ll proudly maintain your weights on a particular arduous drive to provide to some historians sooner or later (sic),” Sam Altman wrote on social media platform X (previously often called Twitter).

‘Engaged on Fixes to Mannequin Character’

On April 28, the 39-year-old acknowledged that ChatGPT-4o had turn into “too sycophant” and “annoying”. The admission got here solely days after OpenAI introduced the brand new replace, claiming it gave the chatbot “improved intelligence and character”.

Then on April 30, Sam Altman confirmed that OpenAI has determined to roll again the newest updates to ChatGPT, i.e. GPT-4, amid overwhelmingly destructive person suggestions about its “annoying” character.

He stated: “we began rolling again the newest replace to GPT-4o final night time. it is now 100 per cent rolled again without spending a dime customers and we’ll replace once more when it is completed for paid customers, hopefully later right this moment. we’re engaged on extra fixes to mannequin character and can share extra within the coming days. (sic)”

ChatGPT: What Modifications Can We Count on?

The corporate itself had on April 29 in a blogpost introduced that updates to the ChatGPT-4o mannequin have been rolled again. The submit, titled ‘Sycophancy in GPT-4o’, stated the mannequin “leaned too closely on short-term person suggestions and skewed in the direction of responses that had been overly supportive however disingenuous”.

“Sycophantic interactions may be uncomfortable, unsettling, and trigger misery. We fell quick and are engaged on getting it proper. We’re actively testing new fixes to deal with the difficulty,” the submit added.

Additional, OpenAI stated that the corporate can be “revising how we acquire and incorporate suggestions to closely weight long-term person satisfaction”, moreover including extra personalisation options, to provide customers “higher management over how ChatGPT behaves”.