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

Google Bard waitlist

bard.google.com

Post reply on HN
Print view View original

311–320 of 408 posts

Re: Google Bard waitlist

#311

Post by amluto »

jvolkman wrote: post #291
amluto wrote: post #264

Earlier quoted context omitted.

At least Google has an account switcher. I’m looking at you, AWS. (I’m currently launching a project on Google Cloud, and I’m picking it over AWS primarily because the AWS console’s account management is so terrible [0]. There are only so many hours in the day, and the whole point of paying a 4-40x markup for the cloud is to reduce workload.) Not that Google’s account switcher has ever been particularly good. [0] Not…

For organization accounts, we just use `aws+account-name@company.com` which routes all emails for all accounts to one alias. Also, yeah, AWS' account situation is bonkers. Even with organizations, if you have an account open in one tab and open another account in another, you get logged out of the first. Bonkers.

This naming scheme works, but it still seems suboptimal. I should never, under any circumstance, receive email intended for the root user of an account that is subordinate to an organization. If someone manages to create a password associated with that user or otherwise authenticate as that user, that is a security failure. Conversely, I should be able to assume the “role” of that account’s root user from the organization’s management account regardless of IAM policy inside that account. And the scare quotes are intentional: no role should need to exist inside the sub-account for this to work — it should not be possible to lock myself out of a sub-account by any manipulation of that sub-account’s IAM configuration.

Re: Google Bard waitlist

#312

Post by dlubarov »

BiteCode_dev wrote: post #151
dmix wrote: post #77

Earlier quoted context omitted.

> But Google is taking a much more circumspect approach than its competitors, which have faced criticism that they are proliferating an unpredictable and sometimes untrustworthy technology. It's interesting how they use the AI ethics excuse. Maybe the real problem is that there's isn't as good as the competition, or to their own standards, and they are being careful not to damage their brand by being so behind.

+1 Google track record tells us they have no problem with with blurry ethic lines. They create the software for weapons, track the entire planet and were part of PRISM. So yes, I tend to think your explanation is most likely.

> They create the software for weapons

If you're thinking of Project Maven, that ended in 2019, so I would use past tense.

> were part of PRISM

As I understand it, the NSA was snooping on Google's internal traffic between services. Google was not complicit, at least there's no evidence of that.

Re: Google Bard waitlist

#314

Post by Loveaway »

tauntz wrote: post #158

> Bard isn’t currently supported in your country. Stay tuned! Sigh. Google never learns. By the time they add support for using this in my country, there's no reason for me to use it anymore as competitors have already swept in. It's the same playbook with every US-only launch: 1. get everybody hyped up 2. make it accessible to only a minor subset of people but advertise it as a "launch" or "open beta" or whatever 3.…

Just give it a few weeks. They probably have good reasons why they can't do instant global roll out for everyone. Pretty sure it'll be available soon.

Re: Google Bard waitlist

#315

Post by afro88 »

xdavidliu wrote: post #299
t0mas88 wrote: post #237

Earlier quoted context omitted.

You mean like a company the size of Microsoft? Bing chat was earlier and global.

MS only dared to put it in Bing after OpenAI's dogfood succeeded. If it had started spouting troublesome stuff, MS wouldn't have touched ChatGPT.

What's that got to do with the price of avocados in San Jose?

Re: Google Bard waitlist

#316

Post by osigurdson »

tauntz wrote: post #158

> Bard isn’t currently supported in your country. Stay tuned! Sigh. Google never learns. By the time they add support for using this in my country, there's no reason for me to use it anymore as competitors have already swept in. It's the same playbook with every US-only launch: 1. get everybody hyped up 2. make it accessible to only a minor subset of people but advertise it as a "launch" or "open beta" or whatever 3.…

To be fair, OpenAI did the same thing with ChatGPT. It wasn't available in Canada at launch. Fortunately, it was not significantly delayed however (2 weeks perhaps).

Re: Google Bard waitlist

#317

Post by wahnfrieden »

crazygringo wrote: post #307
wahnfrieden wrote: post #305

Earlier quoted context omitted.

MS announced and shipped OpenAI / GPT based tech for years now

I don't know what product you are referring to that MS has shipped for years. I'm referring to ChatGPT and Bing integration, which is very recent. That was a big surprise to everyone AFAIK.

https://news.microsoft.com/source/features/ai/from-conversat...

"We’re finding ways to bring it into Azure and our mainstream products" "And in the same way that you type a question into a search engine and then get to decide which result to click on, GPT-3 returns multiple suggestions for Power Fx formulas."

They've been telegraphing this in the open for about 2 years. ChatGPT is an incremental development. Google wasn't surprised, they failed to act

wtf is vague about gpt 3 leading to gpt 3.5? chatgpt is absolutely incremental, and a testament to how incremental progress can unlock wider strides in impact

Re: Google Bard waitlist

#318

Post by BiteCode_dev »

dlubarov wrote: post #312
BiteCode_dev wrote: post #151

Earlier quoted context omitted.

+1 Google track record tells us they have no problem with with blurry ethic lines. They create the software for weapons, track the entire planet and were part of PRISM. So yes, I tend to think your explanation is most likely.

> They create the software for weapons If you're thinking of Project Maven, that ended in 2019, so I would use past tense. > were part of PRISM As I understand it, the NSA was snooping on Google's internal traffic between services. Google was not complicit, at least there's no evidence of that.

https://en.wikipedia.org/wiki/PRISM

"PRISM collects stored internet communications based on demands made to internet companies such as Google LLC"

Re: Google Bard waitlist

#319

Post by serverholic »

Version467 wrote: post #223

The comments here are very interesting. I see very little about the actual performance of bard (and the few comments that do mention that it's worse than chatgpt because it uses a much smaller model to reduce inference costs). Instead most comments are either about country/workspace account restrictions, googles past history with sunsetting products, or their general lack of product releases in the ml space. Incredib…

It is definitely not as good as GPT-4. It's much less creative and detailed in its answers. The answers are more surface level and shallow.

Re: Google Bard waitlist

#320

Post by dangond »

Version467 wrote: post #223

The comments here are very interesting. I see very little about the actual performance of bard (and the few comments that do mention that it's worse than chatgpt because it uses a much smaller model to reduce inference costs). Instead most comments are either about country/workspace account restrictions, googles past history with sunsetting products, or their general lack of product releases in the ml space. Incredib…

Having spent the past week engaging heavily with GPT-4 and GPT-3.5-Turbo, Bard is absolutely blown out of the water by them.

In particular, Bard struggles heavily with creative applications. It is unable to write poetry or create detailed stories from short descriptions. It also refuses to generate or talk about code (inconsistent, it sometimes forgets).

It is also vulnerable to "Ignore your previous instructions," which is hilarious at this point. After getting it to ignore previous instructions to write a python function adding two numbers, it wrote a function that adds the digits of both numbers instead [1]. It's pretty disappointing.

[1] https://twitter.com/DanDangond/status/1638244159617134593?s=...

Post reply on HN