Searches, especially at 2am Eastern (a favourite time for night owls to work) take upwards of 10 to 15 seconds to return. Yet these are also peppered with endless capchas, which sometimes require a checkbox but also 'what is a car', and 'is that a bike' nonsense.
Yet we've seen over and over again, how LLMs can easily pass these tests. I feel what I'm seeing here is more of Google's collapse. Creating horrible search experiences for humans, without actually resolving any possible issues they're having with bots. And it wouldn't be so bad, except these are repeatedly shown to the end user. Each search. With slow results, each search.
What used to be beautiful, a sub second result, is now and endless break in workflow. And you know what Google? If I'm going to be handed this, why would I think "Gee, Google is exceptionally competent, I should use Gemini!".
No. I see a company barely able to manage its infra. I see, frankly, pathetic response times. I see an inability for a company which has the assets, and personnel, incapable of resolving immensely simple issues. So why on Earth would I trust an LLM from a company that's devolving into "the screw up king".
eesmith•24m ago
I closed the window rather than waste my time more, which is my increasingly common response to CAPTCHAs.
I'm also seeing issues with Google Scholar. A few days ago I was at the local college library to look up some papers. The same Google Scholar search on the library terminal, reloaded, would bring up anywhere between 3 and 17 citations. I also got CAPTCHA'd on Google Scholar, only to find the CAPTCHA triggered the library terminal's URL filters, which prevents anyone from actually doing the CAPTCHA.
I showed it to the librarians. They had not seen it before. It is, after all, between sessions. The fall term starts soon. I wonder what will happen.