241–250 of 280 posts
Earlier quoted context omitted.
I mean, they’re already using LLMs. Namely, BERT.
There is no exact definition but BERT would probably just be a LM, not a "large" LM.
Earlier quoted context omitted.
There is no exact definition but BERT would probably just be a LM, not a "large" LM.
Is the exact definition important? The point is, they developed a thing and integrated into their core product. BERT allows them to handle conversational queries much better than before.
Earlier quoted context omitted.
Yep, as AI starts to get trained on AI-generated data the output may well become unstable, you can't build an infinite motion machine (or an infinite gain machine/infinite SNR amplifier) and the system may degrade to essentially white noise. Sort of a cyber-kessler syndrome basically. You really don't want AI-generated content in your AI training material, that's actually probably not generating signal for building f…
Thirteen years ago I met a traveller who paid their way with travel writing, which was basically blog spam. They soon ran out of authentic material so they started writing about places they'd never been using some light googling for inspiration. For a long time now people have been making advertising money by creating bullshit on a large scale. How are you going to prove that any content is organic?
The good news is the internet is relatively good at routing around the shit, for now. And I guess de-facto that is something you could apply to your content inputs: what's the pagerank for this content? actual pagerank, not the advertising/engagement bullshit that the search model has turned into. If the AI generated stuff is correct enough that it has a high pagerank, maybe it's correct enough to be used as an input.
but the thing is honestly there's already been an uptick in ML or AI-generated content that is already surfacing in searches and other places and it's not always correct... and honestly the relevance of google's search results has been noticeably decaying for 10+ years now. Things I know are out there and are relevant are not being surfaced anymore. Is AI generation contributing to that problem? Maybe. Probably not helping, at least.
Since we are posting ChatGPT Twitter links, I like this one: https://twitter.com/goodside/status/1598129631609380864 > explain the worst-case time complexity of the bubble sort algorithm, with Python code examples, in the style of a fast-talkin' wise guy from a 1940's gangster movie
Earlier quoted context omitted.
I am already willing to pay 15-20 dollars a month for this. Google will fall behind as others give this out as a paid service. Don’t need ads for something this useful.
From my knowledge, the cost of large language model search engine will be closer to $150~200 subscription per month than $15~20 in the status quo if the implementation is done naively. The cost will go down rapidly, but it's just not there yet.
Earlier quoted context omitted.
I just tried Googling "when did the moon explode?" to see if it still gave authoritative answers to bogus questions: > About an hour after sunset on June 18, 1178, the Moon exploded. "when did lincoln shoot booth" > April 14, 1865 Mostly they seem to catch and stop this now, but there was a fun brief period where it was popping up the fact-box for whatever seemed closest to the search terms, so "when did neil armstro…
You are providing only a small part of the result. If you provide the full result it make prefect sense why Google would suggest it. When you search "when did the moon explode?". The full result is actually > About an hour after sunset on June 18, 1178, the Moon exploded. That's what it looked like to five terrified, awestruck monks watching the skies over the abbey at Canterbury, in southeastern England, anyway. Whi…
And maybe I'm specifically looking for something which might be wrong? Like, maybe I'm looking for fictional story told as if Lincoln and Booth were in reversed roles?
Earlier quoted context omitted.
As everyone starts to adopt AI, are we going to get to a point where the AI is eating itself. I could imagine AI failing similarly to incestuous genetic lines creating mutations.
Yep, as AI starts to get trained on AI-generated data the output may well become unstable, you can't build an infinite motion machine (or an infinite gain machine/infinite SNR amplifier) and the system may degrade to essentially white noise. Sort of a cyber-kessler syndrome basically. You really don't want AI-generated content in your AI training material, that's actually probably not generating signal for building f…
Here's an example on how misleading ChatGPT is: Prompt: Can you explain what happens when i enter "ping 16843009" in a linux shell? Answer: When you enter the command ping 16843009 in a Linux shell, you are instructing the ping utility to send an Internet Control Message Protocol (ICMP) echo request to the IP address 168.43.9.0. [...] The correct answer is that it will ping 1.1.1.1 Also ChatGPT missed that fact that…
Earlier quoted context omitted.
Exactly, i talked to ChatGPT and it gave me a lot of wrong information in an authorative tone. I consider it dangerous as-is.
What wrong information did it give you?