Hacker News Reader

Comment by jehaj | Hacker News Reader

Not anymore (for me at least). I am able to use it with Firefox. However I do not know if it has something to do with European regulation.

View on HN

Is that a recent development? I can use it on Firefox now as well, but the previous time I checked they were still telling me to use Edge instead.

It was definitely locked to Edge at launch and for quite a period of time. However it has been open to other browsers for at least a few months now.

Too late, they blew their one shot. Ain't no one who tried it when it was new and hyped and was told that they couldn't use it since they weren't using Edge setting reminders to go back occasionally and check whether Microsoft has finally pulled their head out of their ass, it just instantly ceased to exist for them.

I tried it a few times but was mostly disappointed with the results. I built a superior solution (for my usage anyway) in like 50 lines of Python.

Just uses duckduckgo for search, extracts text from top 8 results, shoves into gpt-3.

(I think I got a massive cost reduction by filtering paragraphs with embedding API, but that's just an optimization.)

Phind is better than what I built, though. I don't actually use it that much (I'm not sure why—habit? Impatience?), but every time I do I'm surprised with the accuracy.

I think a big part of it is which pages to select for sources. Google Bard failed hard here in my testing, gave nonsense answers because its sources were AI generated SEO spam!

Yup, I too use Phind sometimes, and it's amazing. Plus if I want to chat with approximately any model known to man, there's chat.lmsys.org and labs.perplexity.ai. Microsoft thoroughly shit the bed by taking a narrow opportunity to get people to move to Bing as an opportunity to instead go "hey, let's blow this on trying to force people to use Edge!"

I too can vouch for Phind for programming questions.

I hadn't had much luck with any of these. Phind is indeed better for programming but it still has the same trouble with the obscure stuff.

It seems that the main problem right now is that they only look at the first few results. Basically they all do this:

1. Ask model for a search query.

2. Run search and feed the results into the model.

3. Ask model for an answer.

The problem is that for very specific questions the search results won't contain the required information. So it has to say something irrelevant or hallucinate. Especially for negative queries like "a foo library that doesn't depend on std" or similar it really struggles because it can't effectively filter out libraries that fail the requirements and can't keep searching until it finds one.

Basically they can be fast if a decent query can have the answer within the first page of results. But fail otherwise. In many cases you are better off just reading through search results on your own.

Reply on news.ycombinator.com