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

Bard and new AI features in Search

blog.google

Post reply on HN
Print view View original

601–610 of 1000 posts

Re: Bard and new AI features in Search

#601

Post by 1024core »

MuffinFlavored wrote: post #417
shawntan wrote: post #369

Earlier quoted context omitted.

Like the earlier responses said, if it is retrieving information from an external data source, that would be a much easier way to update things. REALM ( https://arxiv.org/abs/2002.08909 ) comes to mind as an example of active work in this area. Updating a monolithic LLM seems like a harder problem at the moment, and probably would fall under the umbrella of "continual learning", though that sub-field would have their…

> Like the earlier responses said, if it is retrieving information from an external data source I was under the impression it went through like a multi-month long ultra expensive training process involving many GPUs on terabytes of data "snapshotted" point in time. Is it possible to do this at a smaller scale once a day at the end of the day with "all new content scraped from the Internet nightly"?

The paper about PaLM mentioned that they used 6144 TPU chips to train it over several weeks. That's a metric shit ton of processing power (pardon my language). If Google could dedicate 2 racks of TPUs to just continuously training PaLM (note: I'm not sure if LAMDA is PaLM, but let's assume they're similar), they could do this retraining.

Catch is: consider these statements:

Donald Trump is the President.

Joe Biden is the President.

Barack Obama is the President.

George W Bush is the President.

George H. W. Bush is the President.

Bill Clinton is the President.

All of them were true at some point of time. How do you train a model to disambiguate these?

Re: Bard and new AI features in Search

#602

Post by zalyh »

It would be brilliant if Google can pull this off and take marketshare away from ChatGPT/OpenAI. Just like how people stuck with MS Teams and stopped subscribing to Slack and Zoom. Not because it was better, but it was well-integrated users who are already subscribed to MS Office 365 gets it for free.

Re: Bard and new AI features in Search

#603

Post by coder543 »

alsodumb wrote: post #579
yunwal wrote: post #550

Earlier quoted context omitted.

If you’re talking about physical hardware, Google pales in comparison to Amazon, and Gcloud is still smaller than Azure. It’s possible google’s private compute makes up for the Azure difference but it’s not like they’re in different leagues in terms of access to hardware.

Unlike Azure and Amazon, Google doesn't have to rely on Nvidia GPUs for training and inference, they achieved significant performance gains by using their custom TPUs.

Your comment is simply incorrect. Amazon has had their own TPU equivalents for training and inference for years:

https://aws.amazon.com/machine-learning/trainium/

https://aws.amazon.com/machine-learning/inferentia/

I really don't think this would be a limiting factor regardless, even if Amazon didn't already have multiple generations of these products. It's not as if an Amazon or Microsoft sized company is incapable of developing custom silicon to meet an objective, once an objective is identified. TPUs also aren't really that complicated to design, at least compared to GPUs.

I'm slightly surprised Microsoft hasn't bothered to release any custom ML chips for Azure yet, but I guess they've run the numbers and decided to focus on other priorities for now.

Re: Bard and new AI features in Search

#604

Post by captaincaveman »

nevir wrote: post #362
hn_throwaway_99 wrote: post #304

Question for AI experts out there: ChatGPT is awesome, but it is obviously hamstrung by the fact that it crawled the web at the end of 2021, so all of its data is essentially "frozen" and it doesn't "know" of any topics that occurred after 2021. Seems to me like the next "holy grail" in large language models is building a model that can be continually updated. If Google can achieve that, I think they could leapfrog O…

Google already as its Knowledge Graph - which is a pretty huge set of (semi) structured data about a bunch of things out there: (totally speculative) If Bard were to form queries against knowledge graph, and then summarize the results, it could be very up to date for a good amount of information

They should be able to use it to check for contradictions against what the Bard has generated and facts in the KG.

For example I asked chatGPT about birds that can't fly and it started bullshitting about all sorts of birds that clearly can, those facts would likely be in the KG.

Re: Bard and new AI features in Search

#605

Post by mrkramer »

Miraste wrote: post #168
chipgap98 wrote: post #105

Earlier quoted context omitted.

Microsoft feels like an outlier on this list. They are thriving right now. Also, there are still many categories of things that ChatGPT-like bots can't help with yet, such as shopping. I think we're still very early in this cycle. That doesn't mean Google will succeed, but it feels premature to be writing their obituary.

Windows, Office, and more recently Azure are big enough to make these failures survivable and give Microsoft more chances, but IE and Windows Mobile are textbook examples of complacency.

IE stood no chance against Google Chrome because just like Microsoft shoved down IE down the users' throats the same thing happened with Chrome and Google.

Re: Bard and new AI features in Search

#606

Post by ben_w »

adastra22 wrote: post #563
RangerScience wrote: post #553

Earlier quoted context omitted.

Ehh. From what I've seen in a lot of places - MIRI, AI + prisoner's dilemma experiments, moral philosophy, life - there do seem to be the categories of "clear good behavior" and "clear bad behavior" even if there are also really big categories of "unclear good/bad behavior". In other words, while sure, some "well behaved" is "passing on our biases", there does (IMO) seem to be a big chunk that's "universally well beh…

Nothing in morality is universal.

Sure, but the iterated prisoners dilemma has a nice result that looks like morality if you squint and don't ask difficult questions about if human "altruism" might just be a self-delusion.

Re: Bard and new AI features in Search

#607

Post by pphysch »

xtracto wrote: post #545

At this point, Microsoft should do whatever it can to make a deal with Apple to place a ChatGPT powered platform in the iPhone/i Devices. Most of the people using "computers" nowadays do it through mobile phones, and Google has a strong grip in that market with Android, with only Apple giving them some competition. Thus the only platform where Microsoft could really compete is in iOS. If Microsoft + Apple could get t…

What's less likely? MS+Apple alliance or Google shipping a solid product?

Re: Bard and new AI features in Search

#608

Post by inimino »

deltree7 wrote: post #402
alain94040 wrote: post #388

Earlier quoted context omitted.

Those new language models are "Google killers" because they reset all the assumptions that people have made about search for several decades. Imagine that people start using those chat bots massively as a replacement for Google search. Then the notion of keyword disappears. Google AdSense becomes mostly irrelevant. Of course, Google is a giant today with a history of machine learning innovation. So they have a good c…

Delusional to think Ads work only on Keywords. Where there is human attention, there will always be ads. The more context, the better ads.

"Attention is all you need"?

Joking aside, there's no reason AdWords can't become AdWordVectors and be even more effectively targeted.

Re: Bard and new AI features in Search

#609

Post by sshnuke »

rvz wrote: post #87

But but but 'Google is done', 'it is all over for Google Search', 'Google will be dead in 2 years'. /s Such recent headlines and early proclamations of Google's competitiveness in AI have been erroneously under-estimated here. As soon as OpenAI announces their AI advancement and model, someone else with more money (i.e Google) scales up the parameters, training data to surpass them. The only way to make both of Googl…

Well, I see a big problem with this release.

Let's imagine a user journey:

User makes a request and then there's text, no ads (for now) no links to get you out of search page. So where does Adwords get triggered if there are no linked events? Then advertisers will see a drop in traffic because Google is literally giving the answer instead of making people click to search. This will accelerate the cycle of reducing ad spending, which will make things more difficult for Search. Then SEO marketers and basically all the web industry around Google will see a massive impact as well, because they cannot "optimize" anything that makes people click. The main problem is the dependency of Google on ads (60% of revenue). I see Bard as a negative downwards trend that will decimate the company even further.

Re: Bard and new AI features in Search

#610

Post by kccqzy »

sethbannon wrote: post #161

I was able to see LaMDA in action about 1.5 years ago. It was just as awe-inducing as ChatGPT. Google is not behind in this race, they're simply slower to roll out new functionality. And maybe rightly so? Some of the things I saw ChatGPT say would have been multi-day international bad headlines for Google, if a Google AI had said them instead of an upstart startup.

There is a phrase called "all publicity is good publicity" in fact. It expresses the very real truth that a media outlet reporting a bad interaction by Google AI would probably make people even more curious to try out the AI. If they get the business model nailed down, they should not be afraid of bad press. (Despite our imagination of the ways Google could profit from LLMs, I don't think Google itself has the business model nailed down.)
Post reply on HN