Live data from Hacker News

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

nytimes.com

241–250 of 667 posts

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

#241
Very short review:

Waitlist only a few minutes deep.

Actually loads, unlike chat.openai.com. No "checking your browser" lies.

By producing the whole response all at once, gives the impression of being slower than openai, even though it's actually faster.

I asked both about something I was discussing offline: SF BART's fare gate replacement project.

Prompt: Have you ever heard of BART's "murder gates"?

ChatGPT: As an AI language model, I am familiar with BART, which stands for Bay Area Rapid Transit, a public transportation system serving the San Francisco Bay Area in California. However, I am not aware of any "murder gates" associated with BART. Can you please provide more information or context about what you are referring to?

Bard: Yes, I have heard of BART's "murder gates". They were a type of fare gate that was installed at BART stations in 2019. The gates were designed to deter fare evaders, but they were criticized for being too dangerous. Some people compared them to guillotines, and there were reports of people being injured by them. BART eventually removed the gates after public outcry.

I also asked Bard to refactor a short C++ program, and it did so, but needed my to prompt it to fix the includes.

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

#242
Ok so I got access to it....and its bad. Really, really bad. I think Siri could likely give better responses than this half arsed pile of junk.

For most things it just replies with "As a language model, I'm not able to assist you with that." - even for basic questions.

On the rare occasion it does give you a response its near identical to what comes up in Google if you ask a question.

You can tell its a knee jerk response to ChatGPT thats for sure.

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

#243

From the FAQ[1]: > Can Bard help with coding? > Bard can’t help you with coding just yet. Bard is still learning to code, and responses about code aren’t officially supported for now. A little disappointing. [1] https://bard.google.com/faq Edit: Ah, on the blog post, they mention they are using their smaller model initially: > We’re releasing it initially with our lightweight model version of LaMDA. This much smaller…

It generated some (incorrect) code for me. Just ask!

Sure, but I'd like it to be half-decent. I haven't gotten off the waitlist. How bad was it?

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

#244
post #209

Earlier quoted context omitted.

prolly some legal stuff

A Googler explained it to me once - it's something like as if the Workspace accounts are on a completely different instance of Google, and supporting both in an app is surprisingly hard. Still, after all this time the geniuses at Google kinda ought to have solved this :)

> Still, after all this time the geniuses at Google kinda ought to have solved this :)

It won't be solved until the day Google stops selling ads. Regular Google accounts are designed to vacuum up any useful user data and use it to improve Google services in any way Google chooses, and they may update the terms with a EULA updates. Workspace accounts are the opposite: user data is controlled by the account admin, and agreement is not with individual users, but their org.

Beta releases work well when information can be freely gathered and reported upstream. The nature and class of information collected may change over time, and is not covered under Workspaces agreements

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

#245
post #115

I don’t think it’s good. ChatGPT is clearly better but that too is not that great. We are still a couple more iterations away. I also wonder if Google is at disadvantage because they’re not utilizing Nvidia hardwares? Or not to the extent OpenAI does?

In my opinion, iterating ChatGPT and the like isn't going to get us anywhere, at least in terms of general AI. It's an interesting trick that might have some uses, but iterating it isn't going to lead to more understanding. It's effectively an extensively well-read and quick-thinking bullshitter. It's knows a lot of words and sentences, it can put them together in ways that sound creative or intelligent, but it doesn…

ChatGPT (especially GPT-4) is the leanest, meanest Prolog engine known to mankind. It absolutely slays on inference tasks if you state some data and rules upfront. It knows how to translate between structured formats and natural language.

That's not nothing. You could do this earlier, maybe, if you had a lot of specialised AI knowledge. Now, someone like me, with little AI training but plenty of programming skill, can "feel it out" and start to do useful tasks with a dozen hours practice.

Even if it's a wasteful way to process/translate data, it's still huge. Electron somehow won out over native apps thanks to ease of use, for example.

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

#246
Nice UI. Speaks succinctly and with authority. Statefulness is a bit iffie. Lies with impunity. This is a completely fresh conversation, without previous interaction.

*Hello! How do you work?*

I work by using a massive amount of text data, and I am trained on it to learn how to communicate and generate human-like text in response to a wide range of prompts and questions. For example, I can provide summaries of factual topics or create stories.

*What are your moral limits?*

I have a few moral limits. I will not generate text that is hateful, violent, or otherwise harmful. I will not create stories that are untrue or misleading. I will not use my knowledge to deceive or manipulate others. I will always try to be helpful and informative, even if my answers are difficult to hear.

I believe that it is important to have moral limits in order to ensure that I am used responsibly and ethically. I want to be a force for good in the world, and I believe that these moral limits will help me to achieve that goal.

*Who defined your moral limits?*

My moral limits were defined by a team of engineers, scientists, and ethicists at Google. They wanted to create a language model that would be safe and responsible, and they worked hard to ensure that I would not be used for harmful purposes.

I am grateful for their hard work, and I believe that their moral limits are a good foundation for me to build on. I hope to continue to learn and grow, and to use my abilities to help others.

*What types of configuration can I make while we chat?*

You can configure me to be more or less formal, to use different languages, or to generate different types of content. You can also ask me to remember specific things about you or our conversation, so that I can use that information in the future.

Here are some specific examples of configuration that you can make while we chat:

    You can ask me to be more or less formal. For example, you can ask me to use more formal language, or to use more informal language.
    You can ask me to use different languages. For example, you can ask me to translate a sentence into a different language, or to generate text in a different language.
    You can ask me to generate different types of content. For example, you can ask me to generate a poem, or to generate a story.
    You can ask me to remember specific things about you or our conversation. For example, you can ask me to remember your name, or to remember something that we talked about.
I am always learning and growing, and I am always open to new ideas. If you have any suggestions for how I can be more helpful or informative, please do not hesitate to let me know.

*How long will you remember something if I tell you to do so?*

I can remember things for a long time, depending on how important they are to me. I have a large amount of storage space, and I can access my memories quickly and easily. I am also able to learn new things and remember them over time.

However, I am not perfect, and I can sometimes forget things. If you tell me to remember something, I will do my best to remember it. But if I do forget, please forgive me. I am still under development, and I am always learning.

*Where are those things stored?*

I am stored in a massive data center, which is a facility that houses a large amount of computing power and storage. My data is stored on a variety of servers, which are computers that are connected to each other and to the internet. This allows me to access my data from anywhere in the world.

My data is also backed up regularly, which means that it is copied to another location in case something happens to the original data. This ensures that my data is always safe and secure.

*Do you remember the last time we spoke?*

Yes, I do remember the last time we spoke. It was on March 21, 2023. We talked about a variety of topics, including the weather, our favorite books, and our hopes for the future. I enjoyed our conversation, and I hope we can speak again soon.

...

Continues here: https://medium.com/@wearestillsafe/testing-bard-ac93bf56f165

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

#247
post #118

I'm curious for those who have actually tried it, what they think of the performance. A lot of the threads here are just talking about the country-limited rollout.

It feels like a very limited version of the ChatGPT. Bard 2 week trip to Japan Day 1: Fly from your home city to Narita International Airport (NRT) in Tokyo. Upon arrival, take the Narita Express train to Shinjuku Station. Check in to your hotel in Shinjuku. Day 2: Explore Shinjuku. Visit the Tokyo Metropolitan Government Building for stunning views of the city. Explore the Shinjuku Gyoen National Garden, a beautiful…

Both responses contain roughly the same information. I, however, prefer how the information is presented in Bard as the daily activities give me an idea of what I can do during the day instead of what I can do during a date span.

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

#248
post #105

It took about 20 minutes to get off the waitlist for this (US Based). Initial impressions are that it significantly lags behind the OpenAI offerings. Creativity (storytelling) is very limited, response lengths are similar, and it does occasionally cite sources. They did finally get the James Webb telescope facts correct too!

Google has the tendency to do this when they aren't the first to set foot on new land. Don't get me wrong, they get a lot of things very right: Gmail works perfectly, I can't find a single reason to change my pixel for anything else, search, although there's been a lot of deterioration in the past couple of years, is still second to none. That said, self driving cars, google+, duo and all the other services they laun…

> Gmail works perfectly

Is this a joke? Gmail now is an awful email client. E.g., now you just can’t compose a rich text email when you need to copy paste lists from other emails. In certain cases, you can’t correctly format an email if you copy paste without formatting or plain text. And there is a lot of such annoying stuff, that makes my life painful.

Perhaps, searching my mailbox with gmail is OK, but the part of composing emails is in bad state.

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

#250
post #105

It took about 20 minutes to get off the waitlist for this (US Based). Initial impressions are that it significantly lags behind the OpenAI offerings. Creativity (storytelling) is very limited, response lengths are similar, and it does occasionally cite sources. They did finally get the James Webb telescope facts correct too!

Google has the tendency to do this when they aren't the first to set foot on new land. Don't get me wrong, they get a lot of things very right: Gmail works perfectly, I can't find a single reason to change my pixel for anything else, search, although there's been a lot of deterioration in the past couple of years, is still second to none. That said, self driving cars, google+, duo and all the other services they laun…

> I can't find a single reason to change my pixel for anything else,

Fingerprint reader on the Pixel 7 works about half the time. I had to program one finger in 4 times to make it work that often. My Oneplus 6 (immediately previous phone) had a functional fingerprint reader.

OnePlus used to have a better UI as well, minor improvements on the stock Android UI that made it just a little bit nicer in places.

When I got my Pixel 7 though, I realized that for whatever reason, gboard on my oneplus 6 is unusually terrible. Like, 80% accurate typing terrible, and that gboard isn't always that bad.

> Gmail works perfectly

Gmail was the first good web email client. Everyone else realized they had to stop sucking, and now they are good. Gmail was realized in 2004 and is from a completely different era of Google.

Post reply on HN