[ad_1]
OpenAI’s LLM is already insanely packed with knowledge, but the company is not stopping the pursuit of knowledge. It’s still looking for new ways that it can train its AI to bring it closer to human sentience. According to a new report, OpenAI is working with different organizations to help create new datasets to train its AI.
This is something that all AI companies are trying to do; stuff as much knowledge into their LLMs to keep up with the competition and protect their revenue (oh yeah, and maybe push AI forward). There’s a massive amount of human data floating around in these LLMs from Meta, Google, OpenAI, Anthropic, X Corp, and others. However, there’s not enough for them to be satisfied.
The dream for most of these companies is to develop AGI (artificial general intelligence). When that’s achieved, an AI will basically be as smart as a human being. That’s still several years away, however.
OpenAI is looking to create new datasets to train its LLM
OpenAI is on the prowl for more information to train ChatGPT and other products. It recently began reaching out to private and public organizations for them to submit datasets. These datasets could really be about anything, according to the company. They just need to represent human knowledge. That’s pretty broad, and it shows that the company isn’t being picky about the data.
However, it is looking mostly for data in the Icelandic language. So, it looks like the company is addressing a gap in its knowledge. Representatives from public and private organizations can submit their datasets to OpenAI today.
This is all a step forward to help the company improve ChatGPT and other chatbots. It’s also a way for the company to obtain data with the consent of others. Right now, companies acquire their data by scraping websites for their data. This is something that these companies have gotten in trouble for doing.
[ad_2]
Source link