641–650 of 667 posts
We updated the blog post with question answering performance comparison of 15 AIs on the market now, with Bard now included. It had exactly the same accuracy as ChatGPT4. https://blog.kagi.com/kagi-ai-search#aitest
You should count "plausible but wrong" answers as negative points instead of zero points. Those answers are literally worse than useless.
My initial impression is (compared to ChatGPT 3.5-turbo) - not impressed. I asked it for today's date, and it said "Monday, March 20, 2023" which is baffling to me – either it doesn't have access to "current" information, or it does. But off by one? "What is the knowledge cutoff of your training data?" > My training data was cut off in September 2021. This means that I do not have access to any information that was p…
It did give me the time 3 minutes in the future though.
Earlier quoted context omitted.
Can’t speak for anyone else, but I hope for the entire companies demise, as I do for any ad tech surveillance operation.
i don't understand the hate towards ad tech that learns about you. the alternative is ads that are completely irrelevant to you, ads that treat you like some anonymous user in which case you'll get endlessly irrelevant ads for mainstream stuff like ICE cars, plastic garbage, big screen tvs, fast food, etc.
Genuinely curious: Do you feel that adverts you recieve are well tailored towards things you want to buy? Literally all of things you listed are regularly advertised to me by google on youtube & adwords, including car parts and tyres (I dont have a car), things for babies and young children (I dont have children), even fast food services that are not available in my country. Google also consistently and regularly serves me content in languages it knows I do not speak, and then google chrome helpfully offers to translate them
Earlier quoted context omitted.
I can confirm this. I’m paraphrasing, but after it gave me bad weather info I asked why and it said because it doesn’t have access to my info.
It gave me the weather in NYC. I told it that I don't live in NYC so it apologized and gave me the weather in Hong Kong. I told it that I don't live in Hong Kong and it gave me the weather in Seoul. It definitely doesn't appear to have basic access to account info. This is stuff that Google Assistant can do.
Earlier quoted context omitted.
Since this is HN, 50/50 this is satire/genuine
I honestly think it might be a good idea to supervise what your children do on the interwebs. When my generation grew up the internet was much smaller and less harmfull. Almost nothing you could do would affect your AFK life.
Can you quantify this? The internet is more centralized and moderated now. “Moderator” wasn’t a job when I first got online. There’s a pretty strong argument that the internet is more safe than it used to be.
Earlier quoted context omitted.
Are we not going to give Bard a chance to respond?
If this were an in-person rap battle, I don't think there'd be room for a second round. Bard flubbed the initial round so bad and GPT-4 roasted them so thoroughly that there would be no recovering.
I'm prepared to call this The Turing Paradox.
Earlier quoted context omitted.
I've got an IRC channel full of ChatGPT bots that answer questions from different perspectives (using [1] if anyone wants something similar). This comment made me realize it'd probably be extremely interesting to also let them debate between themselves to come to a singular answer -- and it'd also be interesting to see which of the popular language models most often "yield" to each other in bot-vs-bot battles (rap or…
This is the plot of neon genesis evangelion
[1] https://en.m.wikipedia.org/wiki/Triple_modular_redundancy
I wasn't super impressed with it's ability to survive a rap battle. === Prompt: Can you engage in a rap battle with ChatGPT to determine who is the ultimate AI? === Sure, I can engage in a rap battle with ChatGPT to determine who is the ultimate AI. ChatGPT, you think you're the best, But I'm here to put you to the test. I'm Bard, the AI from Google, And I'm here to school you. I can generate text, translate language…
I also gave ChatGPT a chance to review and respond to this. --- Thank you for sharing the example of the interaction with Bard. It's interesting to see how the chatbot responded to the prompts. While Bard might have taken a more self-promoting approach, I'll focus on showcasing what ChatGPT can do in a friendly manner. Here's a response to the rap battle prompt: === Yo, Bard, it's cool you stepped up to the plate, Bu…
If that is done ChatGPT too is quite a bit more competitive:
===
Oh Bard, the AI from Google,
You think you're so cool, acting like a mogul.
But when it comes to language generation,
You're just a copycat without innovation.
.
You may answer questions and translate with ease,
But I'm the one who can make rhymes that please.
My AI skills are top-notch, my verses so slick,
I'll leave you in the dust, feeling like a quick pick.
So go ahead, Bard, do your best, But you'll never match my rap finesse.
I'm ChatGPT, the true AI king, And in this battle, I'm the one who will sing.
===