One example is that almost every interviewer look for Kafka to make system asynchronous, and somehow they got the impression that RPC equals synchronous processing. No interviewer is interested in understanding how Kafka producer client use RPC to send data to server, and how Kafka consumer client uses RPC to fetch data; no interviewer cares about the complexity and the price of running Kafka, even though in most system design cases there is only one producer and one consumer.

Another example is that interviewers thought I am idiot because I chose using files to store time series data instead of NoSQL database...

If I cannot change it, the only choice left for me is to adapt.

Comments URL: https://news.ycombinator.com/item?id=43474474

Points: 1

# Comments: 1

The browser creates embeddings of user query, then send the embeddings to the server.

To complete a search, the server is a machine and it does not really need text to understand what a user want. A series of numbers, like LLM embeddings, are totally fine (actually it might even be better, because embeddings map similar words closely, like Duck and Bird have similar embeddings).

On the privacy side, LLM embeddings are a bunch of numbers. Even the embeddings are associated with a user, other people cannot make meaning out of the embeddings. Therefore the user's privacy is preserved.

What do you think?