A search result that leads to another search result is the web equivalent of being transferred to another department.
Sometimes that second page is useful.
Sometimes it exists mainly because the query itself generated an indexable URL.
Many websites expose internal search pages at addresses such as example.com/search?q=old+radios. If outside search engines crawl and index those pages, somebody searching the open web may land on a page whose main contribution is another list of results for the same words.
The user searched once and received instructions to keep searching.
Collections can be useful
There is nothing inherently thin about a page containing links to other pages.
A carefully maintained topic index, product category, bibliography, archive page, or curated collection can save readers enormous time. The page adds structure, selection, summaries, filters, or context that the individual items do not provide by themselves.
An automatically generated search-results page can also be useful inside a site.
The problem is whether it deserves to function as a standalone answer for people arriving from outside.
If the page merely repeats the query, displays a few weak matches, and offers no stable editorial value, indexing it can create thousands or millions of low-substance destinations as users try different searches.
Every query can become a page
That is the scaling problem.
A site’s internal search may accept nearly unlimited combinations of words, filters, sort orders, dates, and parameters. If those combinations become crawlable URLs, the theoretical page count can explode without the publisher creating any new underlying information.
Google gives site owners explicit controls for preventing pages from appearing in search results. Its current documentation explains that a noindex rule tells Google to drop a page from its index while leaving the page usable to visitors. See Google’s documentation on noindex.
That does not mean every internal search page should be hidden.
It means publishers can make a deliberate decision about whether those pages contribute enough value to be external destinations.
Search should terminate somewhere
For Dead Internet Theory, indexed internal searches are another way apparent web volume can exceed substantive information volume.
Ten thousand URLs may represent ten thousand documents.
They may also represent ten thousand ways of asking the same database what it already contains.
The useful test is simple:
When a person arrives from search, does the page answer, organize, or explain something?
Or did the internet just hand them another search box?