Live data from Hacker News

Bard and new AI features in Search

blog.google

351–360 of 1000 posts

Re: Bard and new AI features in Search

#351

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…

It seems like google is using search or some knowledge base to throw relevant information into the context when it's generating to get up to date results. It relies on knowing what to retrieve, but google is already pretty good at that.

Re: Bard and new AI features in Search

#352

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…

Disclaimer: I am not an AI expert. I think it depends on how reliable you want this AI to be. Opening search indexing to realtime data is what led to the endless mess we call SEO, and I fear that doing the same for AI invites the same problems.

Bingo - the reason ChatGPT looks so impressive now is the same reason Google was in 2000 - it was a breath of fresh air with a simple interface that was unencumbered by the sales and marketing and advertising barrage that targeted users of its entrenched competition.

Re: Bard and new AI features in Search

#353

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…

You do not need to constantly update the model. You just need to feed it with fresh data as inpput.

Example to try that in action:

https://labs.kagi.com/ai/context

Re: Bard and new AI features in Search

#354

Time will tell (a few weeks, according to the article) if it's any good, as that's when it will be released to the public. However, an interesting piece of info was > It draws on information from the web to provide fresh, high-quality responses. which seems like it will possibly be able to use some kind of external repository of information, instead of needing to be retrained? Very curious about that.

It may works similarly to https://labs.kagi.com/ai/sum, just based on the first couple of Google hits instead of a single page.

Re: Bard and new AI features in Search

#355

For anyone that is interested. This is the ChatGPT reply to the exact same question asked in the preview in the article. Here are a few things that the James Webb Space Telescope is expected to discover that you could share with your 9-year-old: The formation of stars and planets: The James Webb Space Telescope will be able to observe the birth of stars and the formation of planetary systems, helping us understand ho…

ChatGPT is a pure conversational model, while Google can integrate this into a vast knowledge model to work with actual facts. (Just speculation; Googler, but no inside info on this.) Of course, as Microsoft adds ChatGPT into Bing they should also integrate it into the rest of their search stack. This will be an interesting horse-race to watch since it won't be just the AI models but also a massive amount of secondary infrastructure for desired outcomes with content safety (no erotica or nazi pamphlets on demand), IP protection (no plagiarism), political and cultural sensibilities, disinfo, and a ton of other problems.

Re: Bard and new AI features in Search

#356
post #224

People forget that Larry Page's ultimate goal was to have Google predict what you wanted before you even asked for it. This isn't about hype. This isn't about beating the competition. This is about delivering on a 3 decade dream that's now coming to fruition and Sundar just happens to be the one leading the org as it manifests in the real world. We're going to end up with 10+ vertically integrated AI solutions like t…

I'd love for this to be true. I bought into the Google ecosystem in a huge way, but it seems like they have been more focused on the data collection, and advertising end of things, rather than the innovation, and productivity enablement one. Their ecosystem has consitently been a let down in many regards, and they have lost a lot of forward momentum. I hope they can super charge this stuff, but I feel their corporate…

They just brought this extremely innovative thing to market in very little time. The problem may have been that they haven't really had any competition in their core business to push them in a long time.

Re: Bard and new AI features in Search

#357

Lotus 123, Borland and Wordperfect were at the top of the software world. They failed to make the transition to GUIs/Windows. Microsoft missed the arrival of the web. Microsoft also missed the mobile revolution and ended up with zero mobile operating system market share. Internet Explorer, having vanquished Netscape, declared "job done" and stopped developing the web browser further, only to be cast to irrelevance fo…

Plenty of cases where you just want to type something in a search box and see a bunch of different options - like shopping. Let’s say you’re in the market for a keyboard. Would you ask ChatGPT what’s the best keyboard on the market right now, and just buy that? I wouldn’t. I want to see a few different options and I‘m not even someone who enjoys shopping. Where there’s a list of options to display, google will have p…

I would definitely ask ChatGPT what the best keyboards are, summarizing their strengths/weaknesses, and then use that as the starting point of my search.

Comparison shopping is a major PITA with current sites, clicking back-and-forth between various product pages and review sites, and trying to distill it down to a few relevant choices -- things that ChatGPT's excellent abilities at summarization could really help with.

Re: Bard and new AI features in Search

#358

Earlier quoted context omitted.

I guess a couple of reasons: 1) Because it is disruptive. Things may get shook up, and Google may not end up in as exclusive position as they are currently in. There's risk. 2) Because it's not obvious how advertising would fit in with a conversational interface. Google may stay as #1 search/answer engine, but would revenue be adversely affected?

Banner, sidebar, in-line, interstitial, modal… what advertising modes wouldn’t work on conversational interface?

Conversational interfaces seem ideally suited to voice assistants. When you ask "hey google, what's the default password for my router?" does it make you sit through a 30 second ad before saying "the default password is password".

Re: Bard and new AI features in Search

#359
post #307

Earlier quoted context omitted.

Imagine a government controlled AI that partners with mega corporations to "assist" AI in getting you the "correct" results you desire. Call me a pessimist but we may already be staring down the gun of a dystopian future

[flagged]

>Ye, u got a brain no? Go learn some hıstory.

artificial intelligence as a concept is at least as old as 300-400BC. Government corruption as a meme goes much further back.

I wouldn't be so quick to consider someone 'without brain' for having inadequate movie trivia knowledge; nor would I consider encyclopedic knowledge of movie trivia to be evidence of knowing history.

I don't know where the 'bureaucratic to a fault' concept of Brazil was first used, but similarly metaphors and complaints can be found in lots of Roman historical anecdote, many times older.

These concepts have all been explored in fiction to a huge degree, the novelty of what's going on right now is that it's been the first time where any of it has actually approached reality.

The concept of a government 'on autopilot' is now actually being realistically considered, rather than humorously/dramatically fantasized.

It's now becoming apparent to the public at large that such concepts are stepping out of fiction; that's huge.

tldr; don't fault other people for not knowing about the recent revision of ancient ideas and concepts that YOU prefer to discuss.

Re: Bard and new AI features in Search

#360
post #338

Earlier quoted context omitted.

You don't need to retrain the model to access fresh data. You can just tune the model to generate queries to various sources (search engine, etc) to retrieve data.

> You can just tune the model to generate queries to various sources (search engine, etc) to retrieve data. That sounds like a task runner instead of a LLM

Perhaps. I'm not sure that "LLM" is the best description of ChatGPT (see my recent comments), but certainly it's strength is language and weakness is sticking to factual answers (as expected in a search engine context, if not so much in an info-tainment chatbot).

Using ChatGPT as a front end for search, maybe also as a summarizing/presentation interface, and able to maintain context for a conversational interface would all seem to be playing to it's strengths.

There was apparently a brief appearance of a ChatGPT enhanced Bing (search engine) yesterday which was noted to be able to cite sources, which certainly suggest it being used more as a front end than as itself the source of content.

Post reply on HN