Hacker News new | past | comments | ask | show | jobs | submit login

Unfortunately this overhyped launch has started the LLM arms race. Consumers don't seem to care in general about factuality as long as they can get an authoritative sounding answer that is somewhat accurate...at least for now



Somewhat opposite - if LLMs continue to perform like that with made up information and such, their credibility would erode over time and a defacto expectation would be that they don't work or aren't accurate which would result in less being reliant on them.

Same like self driving cars didn't have a mainstream breakthrough yet.


Unfortunately? This is the best kind of arms race, the one where we race towards technology that is ultimately going to help all of humanity.


I'm trying to decide if this is a valid arms race or jumping the gun. Kind of feels like if someone came up with auto racing before the invention of the ICE and so they spend a bunch of time pushing race cars around the track only for them all to realize this isn't working and give upon the whole idea.


I think its more like Tesla Autopilot

In the beginning there was lots of hype because you couldn't use it

But now that its in consumer hands there's tons of videos of it messing up, doing weird stuff

To the point that its now common knowledge that autopilot is not actually magical AI driving


This hasn't really been put in front of consumers, has it? This is all very niche - how many even know that there is a Bing AI thing going on? I think it's far too early to make statements about what people think or want.


Right now, if you go to bing.com, there's a big "Introducing the new Bing" banner, which takes you to the page about their chatbot. You have to get on a waitlist to actually use it, though.


So it's limited to those who use bing and who opt in? Still fairly niche in that case.


OpenAI raced past 100 million users, that's hardly niche. All tech people i've talked to have played around with it. Some use it every day.


As someone who does SEO on a regular basis, I thought it would be brilliant to have this write content for you. Google already made updates to its algo to ferret out content that is created by AI and list it as spam.

I figure we're going to see a lot of guard rails being put up as this gains wider usage to try and cut off nefarious uses of it. I know right now, there are people who have already figured out how to bypass the filters and are selling services on the dark web that cater to people who want to use it for malware and other scams:

Hackers have found a simple way to bypass those restrictions and are using it to sell illicit services in an underground crime forum, researchers from security firm Check Point Research reported.

https://arstechnica.com/information-technology/2023/02/now-o...


But is it a product or a toy for the majority of those users?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: