You know how Google’s new feature called AI Overviews is prone to spitting out wildly incorrect answers to search queries? In one instance, AI Overviews told a user to use glue on pizza to make sure the cheese won’t slide off (pssst…please don’t do this.)
Well, according to an interview at The Vergewith Google CEO Sundar Pichai published earlier this week, just before criticism of the outputs really took off, these “hallucinations” are an “inherent feature” of AI large language models (LLM), which is what drives AI Overviews, and this feature “is still an unsolved problem.”
I’m agreeing with most of what you said, but Google has been working on AI for a long time. Google’s purchased DeepMind in 2014 and kept it as a separate subsidiary, and started their own AI division inside Google itself in 2017.
They also developed a machine learning processor called the TPU, which has been used in their data centers since 2015.
So to Google, AI really means All In. Which is particularly concerning since they don’t even have the best performing AI after a decade of research with a bottomless pit of money.
That are some good points, i didnt really hear about deepmind for a long time and forgot about it. But replacing google websearch with “AI” really sounds like a decision made by marketing department, where they dont understand their own product, their customers or the techs limitations.
Unless of course they want to remove/hide all outgoing links from google search, so the user will spend more time there and google has more opportunities to show them ads from their own ad network, instead of losing the visitors to another website…