3 ms·
I don't know why I would want to query a ML algorithm to generate an answer as opposed to finding a webpage to read about the topic. Sure sometimes I am just lo
by ecshafer 4y ago
I don't know why I would want to query a ML algorithm to generate an answer as opposed to finding a webpage to read about the topic. Sure sometimes I am just looking for "Who was the actor in Alien?" and want a name to spit out. But most of the time I am using the query as a key to a larger amount of information. Search has gotten worse over time, but I don't think this solves that problem. Instead I think the corpus for these will end up being corrupted by SEO and we will have garbage there as well. Increasingly I want a search that is just Google 20 years ago, Kagi is the most promising company in this space imo.
- Always42 4y ago1 possibility, not to be plummeted by ads.
- shostack 4y agoWhich is ironic because that's how this will likely end... In MS serving ads based on your queries and embedded within. Imagine if you asked it what to take for a headache and it embedded a pharma brand with the most minimal ad disclosure they can get their lawyers to agree to. At least with display ads it is blindingly obvious it's an ad.
- genericone 4y agoMicrosoft's cost-per-click for ads embedded in a chatgpt query/response is going to be an order of magnitude higher, since the engine will be able to reason linguistically about what ads are best.
- NoPicklez 4y agoYou might not always want to, but sometimes you might. Sometimes you don't want to read about it, sometimes you just want an answer. The great thing about search is that is covered a wide range of needs.
- brundolf 4y agoI know people who have struggled to learn eg Python programming just from googling (because they lacked any programming background at all so they didn't even know what to search for to have the foundational stuff explained), and then tried using ChatGPT instead and made tons of progress because it could figure out what they meant to ask, and then digest and tune the answers to meet them where they were at Of course it's sometimes wrong, so it's good to verify, which is why it's good that the Bing version cites its sources