The race to build a search engine based on ChatGPT

12 months ago
tgadmintechgreat
112

Another problem with a system like ChatGPT is that its responses are only based on the data it was trained on. Complete retraining of a model can cost millions of dollars due to its size and the scale of the data. You Chat is confused when asked about the latest sports scores, but knows what the weather is like in New York right now. Socher does not want to disclose how the current information is used, considering it a competitive advantage.

“I think right now a lot of these chat interfaces are far superior in some respects to search capabilities, but in others they are clearly much worse,” Socher says. “We are working to reduce all of these problems.”

Aravind Srinivas, founder and CEO of search startup Perplexity AI, who previously worked at OpenAI, says the challenge of updating a system like ChatGPT with the latest information means they need to be merged with something else. “On their own, they can never be good search engines,” he says.

Saam Motamedi, a venture capitalist at Greylock Partners who has invested in AI-powered search company Neeva, says it’s also unclear how compatible chat interfaces are with search engines’ main revenue model, advertising. Google and Bing use search queries to select ads that appear at the top of the list of links displayed in response. Motamedi suspects that new forms of advertising may be required for chat-style search interfaces to be viable, but it’s not entirely clear what those will be. Neeva charges a subscription fee for unlimited ad-free search.

The cost of running a model like ChatGPT on a Google scale can also be problematic. Luis Cese, co-founder and CEO of OctoML, a company that helps companies reduce the cost of deploying machine learning algorithms, believes that running a ChatGPT search can be 10 times more expensive than a Google search because each response requires running a large and complex AI model .

The scale of the ChatGPT mania has taken some AI programmers and researchers familiar with the underlying technology by surprise. The algorithm behind the bot, called GPT, was first developed by OpenAI in 2018, and a more powerful version, GPT-2, was introduced in 2019. This is a machine learning model for processing text and then predicting what will happen next, which OpenAI has shown can perform impressively when training with huge amounts of text. The first commercial version of the technology, GPT-3, has been available to developers for use since June 2020 and can perform many of the tasks that ChatGPT has recently become famous for.

ChatGPT uses an improved version of the underlying algorithm, but the biggest leap in its capabilities comes from the fact that OpenAI allows people to provide feedback to the system about what gives a satisfactory answer. But, like previous text generation systems, ChatGPT is still prone to reproducing biases from its training data, as well as “hallucinating” plausible but incorrect results.

Gary Marcus, professor emeritus at New York University and an outspoken critic of the AI ​​hype, believes ChatGPT is not suitable for search because he has no true understanding of what he is saying. He adds that tools like ChatGPT can create other problems for search companies by flooding the web with AI-generated, search-engine-optimized text. “All search engines are about to run into a problem,” he says.

Alex Ratnerassociate professor at the University of Washington and co-founder Snorkel AI, which works to train AI models more efficiently, calls ChatGPT a “legitimate overkill” in what the software can do. But he also says that it might take some time to figure out how to prevent the creation of language models like GPT. He believes that finding a way to keep them up to date with new information to keep searches up to date will likely require new approaches to training basic AI models.

How long it will take for these fixes to be invented and tested is not clear. It may be some time before technology can radically change how people search for answers, even if other use cases emerge, such as come up with new recipes or working as a study or programming buddy. “It’s amazing, and I told my team that people will see the years before and after ChatGPT,” says Chen of Moveworks. “But whether that will replace search is another question.”

Leave a Reply