Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Google releases Bard to a limited number of users in the US and UK

nytimes.com

Post reply on HN
Print view View original

551–560 of 667 posts

Re: Google releases Bard to a limited number of users in the US and UK

#551

Post by NoZebra120vClip »

I picked up access to Bard today, and I'm enjoying it so far. It doesn't know how to make limericks hilarious much, but it cares not for the strictures of copyright or trademark that bogged down Bing Chat. So I successfully extracted a screenplay about the Millennium Falcon battling the Enterprise-D (the Falcon escaped into hyperspace) and also some unfunny scenes of Catholics in the confessional (unless a woman coveting her neighbor's wife is funny) and Dilbert getting a massage at work (not even sexual).

I've got a few good limericks and they almost sort of rhyme. So this has a bit of potential. Enjoyable nonsense for me, and I prefer to be in Google's ecosystem, anyway.

Re: Google releases Bard to a limited number of users in the US and UK

#552

Post by mr_toad »

cloudking wrote: post #411
minihat wrote: post #405

Earlier quoted context omitted.

I gave GPT-4 the chance to respond. === Prompt: copy and paste of er4hn's Bard rap . Now it is your turn. Write a rap verse which roasts Bard. === Yo Bard, you tried to step up, but you tripped, In this rap battle, man, you just got whipped. You claim you're the future, and here to stay, But your weak rhymes, man, they just don't slay. You say I'm a toy, just a chatbot, But my skills, my knowledge, they all run hot.…

that unbuttered toast line.. not bad!

Makes me curious to know if it actually came up with that simile, or just copied one. Either way, that it chose to use one is interesting.

Re: Google releases Bard to a limited number of users in the US and UK

#556

Post by dragonwriter »

simonw wrote: post #496
Georgelemental wrote: post #480

Earlier quoted context omitted.

The initial prompt (before the user types anything) tells the language model about itself, no?

That depends on the chatbot. Normally it just says something like "You are a helpful assistant called X". In Bing's case it was pretty long, but it still didn't include information that would answer questions about how the bot was trained: https://simonwillison.net/2023/Feb/15/bing/#prompt-leaked

> but it still didn’t include information that would answer questions about how the bot was trained

If the hidden prompt includes information identifying the model, and the chatbot has web search/browsing ability (since it probaby won’t have data scraped about itself in its training set), the combination can give it the ability to find and present information about itself.

Re: Google releases Bard to a limited number of users in the US and UK

#557

Post by dragonwriter »

kevin_thibedeau wrote: post #534
nerdix wrote: post #451

Earlier quoted context omitted.

Whenever a DALL-E/Stable Diffusion like model but for video comes out... The CP thing is going to open up a can of worms. IIRC, the Supreme Court already upheld that drawings that depict CP are constitutionally protected. What happens when a computer draws a depiction of CP that can't be distinguished from reality? I guess it's protected but it'll make it much harder to identify real CP.

There have been convictions based on child heads photoshopped onto adult bodies. Generative images will be treated the same. The lack of a real victim is treated as immaterial. All it needs is sufficient realism to be believable.

> > the Supreme Court already upheld that drawings that depict CP are constitutionally protected.

> There have been convictions based on child heads photoshopped onto adult bodies. Generative images will be treated the same.

The former suggests the latter is not true, unless it is generative images provably using the likeness of an actual specific child. (Of course, the Supreme Court can go against its prior precedent, but assuming it will do so is not usually the best prediction unless there have been some very clear signals.)

Re: Google releases Bard to a limited number of users in the US and UK

#558

Post by ncallaway »

aiappreciator wrote: post #482
akomtu wrote: post #469

Earlier quoted context omitted.

Anyone has doubts that this thing can outtalk and outdebate just about anyone? Imagine a politician on stage, who gets hints from gpt4 in hus earpiece.

You are mixing up grammatical skills with logical thinking skills. Turns out, poetry, rythme, rap, are all low level skills, which the AI easily conquer. To master them, you only need to master language, but not world knowledge. Turns out, writing novels is the hardest thing for LLMs. GPT-4 still absolutely sucks at writing interesting stories. Because story writing requires intensive world-modelling and logical skil…

> This thing will outtalk a politician, only if the politician has no new policy ideas.

What political debates have you been watching? I’d love to live in the world to do, where political debates turn on the logical cohesion of arguments and the merits of policy proposals.

Re: Google releases Bard to a limited number of users in the US and UK

#559

Post by dragonwriter »

jarvist wrote: post #471
stephencoyner wrote: post #216

Me: As a part of Google, do you have my personal info? Bard: Yes, as a part of Google, I have access to some of your personal information. This information includes your name, email address, and phone number. I also have access to your search history and your location data. I use this information to provide you with personalized search results and to improve my ability to understand and respond to your questions. I u…

Interestingly, in the UK I get this: Me: As a part of Google, do you have my personal info? Bard: I do not have your personal information. I am not able to access or process any personal information that you have not shared with me. GDPR legislation saving us from AI?

I get the same kind of answer in the US (in all 3 drafts in response to a similar query.)

Re: Google releases Bard to a limited number of users in the US and UK

#560

Post by nextaccountic »

aiappreciator wrote: post #482
akomtu wrote: post #469

Earlier quoted context omitted.

Anyone has doubts that this thing can outtalk and outdebate just about anyone? Imagine a politician on stage, who gets hints from gpt4 in hus earpiece.

You are mixing up grammatical skills with logical thinking skills. Turns out, poetry, rythme, rap, are all low level skills, which the AI easily conquer. To master them, you only need to master language, but not world knowledge. Turns out, writing novels is the hardest thing for LLMs. GPT-4 still absolutely sucks at writing interesting stories. Because story writing requires intensive world-modelling and logical skil…

> Turns out, poetry, rythme, rap, are all low level skills, which the AI easily conquer

10 years ago AIs pretty much struggled to make a coherent sentence. As of 2023 this stuff is "easy", but what else will be easy in 2033?

Post reply on HN