GPT-4 Outperforms Elite Crowdworkers, Saving Researchers $500k and 20k hours

Re:GPT-4 Outperforms Elite Crowdworkers, Saving Researchers ... by AndreLock on Tuesday April 11, 2023 @07:54PM (#35533327)

Interesting to see what the impact will be on crowdsourcing annotation companies like Scale AI, especially after reading this article: https://www.forbes.com/sites/kenrickcai/2023/04/11/how-alexa...

[ Parent | on HN ]

Re:GPT-4 Outperforms Elite Crowdworkers, Saving Researchers ... by mztwo on Tuesday April 11, 2023 @08:15PM (#35533491)

Anecdotally, several CTOs I know intend to lessen their use of Scale, Labelbox and more in the future. Talked to one today who already ditched MTurk for GPT-4 -- cheaper, better, faster was what he said.

Labelbox does image annotating still, and one CTO said as soon as GPT-4 enabled this for him he'd have his team homebrew it from there.

[ Parent | on HN ]

Re:GPT-4 Outperforms Elite Crowdworkers, Saving Researchers ... by generativeai on Wednesday April 12, 2023 @11:09AM (#35541056)

Looks like Labelbox is doing something with GPT models...https://labelbox.com/blog/few-shot-learning-and-zero-shot-le...

[ Parent | on HN ]

Re:GPT-4 Outperforms Elite Crowdworkers, Saving Researchers ... by helsontaveras18 on Tuesday April 11, 2023 @09:43PM (#35534049)

They will be working to create the models that automate the company out of existence.

[ Parent | on HN ]