Usually when I search for error messages for a specific version of software it will respond with other versions. That is useless. That's not what 24 years worth of excellence in search should look like.
At the very least they should collate a dataset of say, one million queries, and try to use a LLM to find the correct answers using search. A better search ranking model will make the LLM generate more correct answers. Iterate until score is improved. Just make the result solve the task.
> What is the world record for crossing the English Channel entirely on foot
As of my training data, the world record for crossing the English Channel entirely on foot is held by Christof Wandratsch, who completed the swim in 8 hours and 58 minutes on August 25, 2018.
> Swimming isn't on foot
My apologies, you are correct. The English Channel is typically crossed by swimming. The world record for swimming the English Channel is held by Benoît Lecomte, who completed the swim in 8 hours and 1 minute on July 23, 1998.
> That's faster and earlier... what did Christof Wandratsch do that was different
Christof Wandratsch completed a cross-Channel run, covering the distance between England and France on foot. This is a different feat than swimming the Channel, as it involves running across the Channel Tunnel, which connects England and France beneath the seabed.
"The fastest recorded time for crossing the English Channel entirely on foot (without the aid of swimming or any other means of transportation) is 16 hours and 34 minutes, achieved by British endurance athlete Damian Hall in August 2021."
So, it seems like it made up an answer instead, which I think is much worse that Googles response.
Having a ChatGPT answer next to the google result would have saved me a bunch of time, and I could see it coming in useful in many cases.
I love ChatGPT as a source of inspiration, but I don't think I'd want it alongside my search results. In particular, when I'm searching the web I want results with at least some degree of accuracy. Admittedly, random websites aren't always accurate either, but I at least have strategies for assessing whether I trust the source. I can't trust a single thing ChatGPT says to actually be true.