Hacker News Reader

Comment by pavlov | Hacker News Reader

I guess it hinges on the definition of “looking”.

Receiving files, copying them within a file system, and transmitting the contents? Clearly not looking.

Computing a hash for each file? Doesn’t seem like looking either.

Parsing the contents, or tokenizing them for a ML model? That’s looking IMO.

I think the defining factor is the size and nature of the derived state update on the server. Copying and transmitting retains only metadata about the file. Computing a hash produces only a handful of bytes of state about the file, and there’s nothing that can be inferred about the file’s contents from the hash. But if you start storing something like word frequencies, that’s not neutral anymore, it’s a “look” into the contents.

View on HN

What about parsing and tokenizing the code for code search?

Yes, that’s clearly GitHub looking at the contents. At that point you just have to trust they don’t do anything else with the internal representation they’ve built.

When Gmail was introduced two decades ago, there was some short-lived furore about the fact that Google admitted they were reading your emails for ad targeting purposes. To a lot of people like myself, this seemed a bridge too far when you’re used to servers that don’t “look” at your private stuff. But it certainly didn’t hurt Gmail’s adoption. (I still don’t use Gmail because of this original lingering distaste.)

They don't "read" emails for ad targeting anymore (according to them). But they certainly do for search and spam filtering, which are features that users expect.

Once things hinge on the definition of commonly-used verbs and nouns, you've got a problem.

And if you don't agree, just remember that my claim hinges on the definition of the words "you", "definition", "got" and "a".

The problem is GitHub’s if they’re the ones accessing copyrighted content under poorly defined terms.

Go to github.com, open your repository and open a file. It shows up on the screen. Is that not "looking"?

No, because the server doesn’t update any state based on the transmitted file contents except some metadata like “last viewed”.

The state update based on the file contents happens in my brain, and that’s the “look” action.

They do code highlighting, search indexing and symbol resolution. I don't know if those can be disabled but they need them to look at the code, update state (indexes and such, maybe caching on highlighting).

code highlighting is done in the client, the infrastructure just serves the content for that particular feature.

Reply on news.ycombinator.com