Bard’s latest update: more features, languages and countries • Hacker News Board

Re: Bard’s latest update: more features, languages and countries

#41

Post by llm_nerd » Thu, Jul 13, 2023, 3:18 PM UTC

canadiantim wrote: post #37

llm_nerd wrote: post #23

Earlier quoted context omitted.

Google is trying to "punish" Canada for bill C-18 [1]. It is absolutely gross, bully-like corporate behaviour and ultimately should earn Google much more scrutiny. For those who don't know, bill C-18 required Google and others to pay Canadian news orgs if they scrape and present their content. This is exactly the battle Australia had with Google (Australia won, btw), but now that Canada has moved forward Google has d…

As a Canadian I am glad that Google is taking this stand. If Canada succeeds in its current implementation of Bill C-18 it will absolutely destroy the revenue for media organizations in Canada rapidly accelerating the decline in journalism here. Not to mention the morales of taxing the free flow of information on the internet. So as a Canadian, THANK YOU GOOGLE!

>If Canada succeeds in its current implementation of Bill C-18 it will absolutely destroy the revenue for media organizations in Canada rapidly accelerating the decline in journalism here.

Every media org is behind C-18. Even PostMedia, about as "right wing" as it gets here, is behind C-18.

C-18 is complex. I have paid little attention to it as it isn't my field and I honestly don't care that much about it. But your take is quite humorous in the context.

Re: Bard’s latest update: more features, languages and countries

#42

Post by dunefox » Thu, Jul 13, 2023, 3:18 PM UTC

canadiantim wrote: post #37

llm_nerd wrote: post #23

Earlier quoted context omitted.

Google is trying to "punish" Canada for bill C-18 [1]. It is absolutely gross, bully-like corporate behaviour and ultimately should earn Google much more scrutiny. For those who don't know, bill C-18 required Google and others to pay Canadian news orgs if they scrape and present their content. This is exactly the battle Australia had with Google (Australia won, btw), but now that Canada has moved forward Google has d…

As a Canadian I am glad that Google is taking this stand. If Canada succeeds in its current implementation of Bill C-18 it will absolutely destroy the revenue for media organizations in Canada rapidly accelerating the decline in journalism here. Not to mention the morales of taxing the free flow of information on the internet. So as a Canadian, THANK YOU GOOGLE!

Nice try, Bard!

Re: Bard’s latest update: more features, languages and countries

#43

Post by stainablesteel » Thu, Jul 13, 2023, 3:19 PM UTC

nothing's changed still based on my tests

GPT can code - and its gotten better at it, and Bard can summarize things from the internet very well and do research for you in less time than it takes to write a prompt

Re: Bard’s latest update: more features, languages and countries

#44

Post by llm_nerd » Thu, Jul 13, 2023, 3:21 PM UTC

crazygringo wrote: post #39

llm_nerd wrote: post #23

Earlier quoted context omitted.

Google is trying to "punish" Canada for bill C-18 [1]. It is absolutely gross, bully-like corporate behaviour and ultimately should earn Google much more scrutiny. For those who don't know, bill C-18 required Google and others to pay Canadian news orgs if they scrape and present their content. This is exactly the battle Australia had with Google (Australia won, btw), but now that Canada has moved forward Google has d…

Do you have any evidence of this? Or are these just assumptions?

Google is in a very loud, very public fight with Canada about C-18. Canada has zero unique laws about LLMs, and actually has some of the weakest privacy laws among its peers. Canada is extremely pro corporate in this regards compared to the EU, for instance.

Google opened BardGPT to everyone on the planet except for a few despotic governments....and Canada. Their stated reasoning was "regulatory uncertainty". That is farce.

Re: Bard’s latest update: more features, languages and countries

#45

Post by ajross » Thu, Jul 13, 2023, 3:23 PM UTC

swyx wrote: post #21

fun story[0] about Bard's June update - implicit code execution (remember? https://news.ycombinator.com/item?id=36229782 ) - the example that they advertised on their blogpost/video demo now no longer works. Here's what I got as of today: https://g.co/bard/share/2580222b6fb2 (screenshot from a couple days ago: https://twitter.com/swyx/status/1678495067663925248?s=20 ) > What are the prime factors of 15683615? > (afte…

> now lets all acknowledge that maintaining stable performance of LLMs is an unimaginably hard problem, but how do we trust any Bard update when there's no regression testing on advertised outputs?

This framing ("how do we trust [target product] when there's [a problem]") is basically a fallacy. Advocates and evangelists and shitposters deploy it every time they want to take an absolutist position but only have one argument, and it only works if you assume through absence of evidence that all competitors are flawless. Holding new products to impossible standards isn't how the tech industry grows.

Google[1] Bard is an early access ("beta" essentially, though IIRC they don't use the term) product under very active development. You should expect bugs. In particular you should expect regressions, which are the inevitable result of development speed.

I mean, good grief, there's a giant "Use code with caution. Learn more" warning right on the conversation you posted! So why come here to nitpick and pretend that someone told you it was perfect?

[1] I work there, but on firmware and wouldn't know a ML transformer if you hit me with one.

Re: Bard’s latest update: more features, languages and countries

#46

Post by Xplune13 » Thu, Jul 13, 2023, 3:25 PM UTC

jug wrote: post #8

visarga wrote: post #6

I tried Bard and it feels worse than GPT-4, it doesn't follow the prompt very well. Uploading an invoice it failed to extract data, instead it responded with hallucinations.

Yes, I think it should be compared to GPT 3.5. Google doesn't really have a public contender to GPT 4 yet but this is what the Gemini project is intended to be, by the newly formed/reformed Google DeepMind team.

I don't think Bard can even perform at GPT 3.5 level, at least, not in my experience. Bard hallucinates pretty bad and you can't really be sure whether it's correct or not. Now, that's a given with any LLM, but in my experience, Bard usually gives you at least 1 wrong sentence in any of the results it produces whereas GPT 3.5 is mostly correct at surface level questions, but tends to hallucinate if you try to dig in deeper.

But, Bard can give you images as an output as well with the links to those images which may or may not be the correct links and GPT 3.5 can't do that. Still, in any general case, I'd say GPT 3.5 is way more reliable than Bard.

Re: Bard’s latest update: more features, languages and countries

#47

Post by canadiantim » Thu, Jul 13, 2023, 3:27 PM UTC

llm_nerd wrote: post #41

canadiantim wrote: post #37

Earlier quoted context omitted.

As a Canadian I am glad that Google is taking this stand. If Canada succeeds in its current implementation of Bill C-18 it will absolutely destroy the revenue for media organizations in Canada rapidly accelerating the decline in journalism here. Not to mention the morales of taxing the free flow of information on the internet. So as a Canadian, THANK YOU GOOGLE!

>If Canada succeeds in its current implementation of Bill C-18 it will absolutely destroy the revenue for media organizations in Canada rapidly accelerating the decline in journalism here. Every media org is behind C-18. Even PostMedia, about as "right wing" as it gets here, is behind C-18. C-18 is complex. I have paid little attention to it as it isn't my field and I honestly don't care that much about it. But your…

Every LARGE media organization is behind C-18, because they will get some money from it (not even that much, most of it goes to Rogers, Bell and the CBC; the PBO suggests only $80 million split between newspapers in Canada from the bill, and that's only if Meta and Google don't withdraw from the market which they are currently intent on doing).

Almost every Independent media organization is against C-18, because they will receive no benefit whatsoever from the Bill. Instead they receive irreperable harm in having their revenue and traffic sources from Meta and Google shut off, which for an industry already desperately hanging on by a thread is a significant blow and will result in the bankruptcies of most independent media organizations in Canada.

Re: Bard’s latest update: more features, languages and countries

#48

Post by aaomidi » Thu, Jul 13, 2023, 3:30 PM UTC

ajross wrote: post #45

swyx wrote: post #21

fun story[0] about Bard's June update - implicit code execution (remember? https://news.ycombinator.com/item?id=36229782 ) - the example that they advertised on their blogpost/video demo now no longer works. Here's what I got as of today: https://g.co/bard/share/2580222b6fb2 (screenshot from a couple days ago: https://twitter.com/swyx/status/1678495067663925248?s=20 ) > What are the prime factors of 15683615? > (afte…

> now lets all acknowledge that maintaining stable performance of LLMs is an unimaginably hard problem, but how do we trust any Bard update when there's no regression testing on advertised outputs? This framing ("how do we trust [target product] when there's [a problem]") is basically a fallacy. Advocates and evangelists and shitposters deploy it every time they want to take an absolutist position but only have one a…

> So why come here to nitpick and pretend that someone told you it was perfect?

They’re asking why is their own demos failing with new releases and asking for regression testing. That’s a completely fair criticism and I’m surprised you think this is nitpicking?

Re: Bard’s latest update: more features, languages and countries

#49

Post by ipaddr » Thu, Jul 13, 2023, 3:30 PM UTC

ajross wrote: post #45

swyx wrote: post #21

fun story[0] about Bard's June update - implicit code execution (remember? https://news.ycombinator.com/item?id=36229782 ) - the example that they advertised on their blogpost/video demo now no longer works. Here's what I got as of today: https://g.co/bard/share/2580222b6fb2 (screenshot from a couple days ago: https://twitter.com/swyx/status/1678495067663925248?s=20 ) > What are the prime factors of 15683615? > (afte…

> now lets all acknowledge that maintaining stable performance of LLMs is an unimaginably hard problem, but how do we trust any Bard update when there's no regression testing on advertised outputs? This framing ("how do we trust [target product] when there's [a problem]") is basically a fallacy. Advocates and evangelists and shitposters deploy it every time they want to take an absolutist position but only have one a…

Everything google offers is in beta and remains that way for a very longtime. Expect bugs but not regression of core product offerings.

Re: Bard’s latest update: more features, languages and countries

#50

Post by holoduke » Thu, Jul 13, 2023, 3:30 PM UTC

I asked a question and it gave a wrong answer. So i responded with "thats the wrong answer ditshit". It then made an apology for being rude. I was rude not Bard. Then every follow up question ended with an appology and explanation about being rude. Quite weird behavior.