Claude can now search the web

Post by matwood » Thu, Mar 20, 2025, 9:53 PM UTC

johntb86 wrote: post #215

tcdent wrote: post #102

Searching the web is a great feature in theory, but every implementation I've used so far looks at the top X hits and then interprets it to be the correct answer. When you're talking to an LLM about popular topics or common errors, the top results are often just blogspam or unresolved forum posts, so the you never get an answer to your problem. More of an indicator that web search is more unusable than ever, but inte…

I've found that OpenAI's Deep Research seems to be much better at this, including finding an obscure StackOverflow post that solved a problem I had, or finding travel wiki sites that actually answered questions I had around traveling around Poland. However it finds its pages, they're much better than just the top N Google results.

Post by wongarsu » Thu, Mar 20, 2025, 9:55 PM UTC

eli wrote: post #146

HankWozHere wrote: post #124

Earlier quoted context omitted.

Kagi Assistant allows you to do search with LLM queries. So far I feel it bears reliable results. For instance - I tried couple of queries for product suggestions and came back with some good results. Whilst it’s a premium service , I find the offering to be of good value.

It's neat but I've found the value kinda variable. It seems heavily influenced by whatever the first few hits are for a query based on your question, so if it's the kind of question that can be answered with a simple search it works well. But of course those are the kinds of questions where you need it the least. I find myself much more often using their "Quick Answer" feature, which shows a brief LLM answer above th…

Post by jimbokun » Thu, Mar 20, 2025, 9:58 PM UTC

exe34 wrote: post #262

jimbokun wrote: post #254

Earlier quoted context omitted.

Unfettered capitalism is pretty good at figuring out which is which. It’s pretty core to Elon Musk’s animating philosophy: get as many jobs as possible then see if there’s any negative impact. Not as appropriate in a government setting where the impact goes far beyond personal profit and loss.

if a CEO can run 7 companies and still play monopoly with the ship of government, then maybe CEOs aren't really that useful.

Post by DrammBA » Thu, Mar 20, 2025, 10:00 PM UTC

deivid wrote: post #129

joshstrange wrote: post #119

Earlier quoted context omitted.

> Web search is available now in feature preview for all paid Claude users in the United States. It is for all paid users, something OpenAI is slow on. I pay for both and I often forget to try OpenAI's new things because they roll out so slow. Sometimes it's same-day but they are all over the map in how long it takes to roll out.

For all paid users _in America_. It's not available for me in Europe.

Post by jsbg » Thu, Mar 20, 2025, 10:02 PM UTC

birken wrote: post #166

_2d30 wrote: post #39

Earlier quoted context omitted.

I don't think it should. If a user asks the AI to read the web for them, it should read the web for them. This isn't a vacuum charged with crawling the web, it's an adhoc GET request.

The AI isn't "reading the web" though, they are reading the top hits on the search results, and are free-riding on the access that Google/Bing gets in order to provide actual user traffic to their sites. Many webmasters specifically opt their pages out of being in the search results (via robots.txt and/or "noindex" directives) when they believe the cost/benefit of the bot traffic isn't worth the user traffic they may…