Google Bard waitlist
bard.google.com
bard.google.com
Google releases Bard to a limited number of users in the US and UK - https://news.ycombinator.com/item?id=35246253
How about we use that thread for comments about the product itself, and this thread for comments about the waitlist/country/account restrictions.
The discussion is partitioned almost evenly between these two distinct topics, though, so I think it makes sense to keep the threads separate.
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. tell the rest of the world that "lol, by "avialable" we mean US-only" or "Open means "open for people with US IP addresses only"" etc.
4. Wait till competitors have launched global-from-the-start alternatives
5. Open up the service globally
6. ...
7. Shut down the service because there's no uptake (surprised_pikachu.jpg)
Rinse and repeat.
Self inflicted wounds and all that..
After getting access: You do need to have a US/UK IP to use it but they don't seem to use any VPN detection heuristics - which would be trivial for them since I'm signed in to them using multiple private/work accounts on multiple machines without VPN.
Google tries hard to make their paying consumers hate them.
And not to mention Pixel devices, which you can't even order.
if you sell something, then sure goods might have laws. but just text ? lol
And on top of that, there are plenty of novel things that may or may not be illegal. If it can be tricked into making a pro-Nazi statement, is it violating German law? What if it offers medical or legal advice without a license?
It's a pretty big legal minefield, and different jurisdictions have different laws and standards, so it makes sense to limit your exposure.
This attitude is what is leading to the regulation and probable over regulation due to backlash of ‘big tech’.
Those days are over.
Just like any other sector doesn’t get to ignore laws.
Unregulated US social media hit our democracies like a hurricane.
The body politik doesn’t seem to be of a mind to let them do the same with AI.
If that's the problem then they're not supplying the right fix. Surely restricting this to users located in the US doesn't prevent people from writing in a language that isn't English. Google should be capable of detecting the language of a query, and potentially rejecting it based on that with an apology.
> PR,
Obviously there must be _some_ reason why they're doing this. Doesn't mean it's a good reason.
> scaling, not enough capacity for the new users.
That would be rather embarrassing for Google.
Why? They have no idea how many users they will get. If they buy A100s on the assumption they will get 50M daily active users they run the risk of wasting an enormous amount of money if they get 1M users instead. And its not like these GPUs grow on trees. Clearly MSFT is struggling to set up compute fast enough, see the decreasing rate limits on GPT-4.
> Clearly MSFT is struggling to set up compute fast enough, see the decreasing rate limits on GPT-4.
Well, even so, this is how you'd deal with capacity problems, rather than by arbitrarily shutting out parts of the world.
In order to not geo-block, Google would need to be able to scale infinitely? What kind of straw man is this?
> there was a shortage of memory chips, which required a lot of creativity.
Geo-blocking doesn't strike me as particularly creative as far as solutions go. Whatever problem they're trying to address, the excuses being made here seem weak and don't make Google look any less incompetent.
Not sure about incompetence, it's just a very hard problem. Sam Altman's estimations for ChatGPT were off by 10x, for reference. Apparently some employees thought it would be a total flop that would barely make the last page of major newspapers.
> You're also saying Google, _a cloud computing provider selling access to A100s_, can't scale this dynamically?
I'm saying that Google can not acquire enormous quantities of A100s overnight, yes.
We don't want that. Every time Google launches something it takes months for it to understand that just because I'm currently located in country X does not mean I want localization or personalization for that country. Just release the service for those that can use it without the localization. After all we know that the vast majority of the models are trained on English, it's useless to give me a subpar undertrained model on local language.
This isn't "just text"
There could be a hundred reasons why they released it like this. ChatGPT was also not available world wide when it launched.
But you're on a website with founders and tech workers. Part of what we do here is talk about why things work, why they don't, and ways to improve. There's nothing wrong with criticizing our industry or pointing out when one of the big companies regularly shoots itself in the foot.
Actually, in some cases, there is: exposing your endpoints to nefarious actors from countries where there is no jurisdiction, functioning tech laws, etcetera. It is better & safer to launch small with a somewhat "trusted crowd" first. You live in a dream world or have not launched anything really on your own before. It is wrong & dangerous to follow your advice or criticism because you don't know what you don't know.
> Bard does not currently support Google Workspace accounts or when our systems indicate you may be under 18. Learn more
Google: 15 years of punishment for being a loyal Dasher/ Google Apps/GSuite/Google Workspace user.
Agree, very frustrating.
Not that Google’s account switcher has ever been particularly good.
[0] Not just the console. AWS organizations are sort of okay, but you still need to give every account an email address. Seriously, Amazon? I bet you could invent a new subtype of account called, say, “project”, where a project has a name, does not have an email address, does not have a root non-IAM user, and is only accessed via IAM roles. And I bet you could, with more effort, teach the console to accept a URL prefix that encodes the name of an account and a role and gives a console view with that role assumed without globally assuming that role. And then I could administer a VPC according to current best practices, from the console, without cursing at multi-account containers and needing to correctly type the account and role every time I want to do anything.
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.
Also better to have no browser extensions installed when doing anything sensitive.
Bad experience Google.
What gives.
Meanwhile ChatGPT is used everywhere in France, and it's now synonymous with AI for most people in Europe.
Game over.
But GPT is just miles away better than everything else, and it keeps getting better at a scary rate, especially because the feedback loop from so many users is making it better.
- heavy administration
- schools have a strong emphasis on essays
- getting a job often implies writing a cover letter
And since French is not easy, so many people find it to be a barrier to entry.
While I love GPT for quickly giving me code snippets of things I would take minutes to come up with, my friends have this problem for the regular day-to-day language.
I never noticed it before, because it's so natural to write to me, that I never have to think to hard to express anything. And because I've been a freelance for so long, doing formal writing is fluid. I know all the canned phrases by heart. The polite words to end the message.
But I came to realize: they don't.
As soon as I showed them chatgpt, their eyes light up. The killer feature for them was not the ability to answer complicated questions or save on search, nor was it the possibility of getting a summary. It was producing boring letter-like messages, with no typo.
It's just that 99% of US big tech companies are very anti-consumer and have gotten away with it in the US in the past. But that also seems to be changing over time.
When Clubhouse finally opened up to a wider audience of people who weren't being incentivized to use it, new users churned out almost immediately because it wasn't actually very cool at all.
Basically, the story of Clubhouse was the story of Google Glass.
You're talking about maybe ~1% of search that Bing Chat is practically available to (in Europe).
It's a lot less risky for Microsoft to do what they did than for Google to just open the floodgates in Europe.
If Google does something on Chrome OS - the EU isn't going to scrutinize it the same way if MS does something on Windows.
Same, but opposite for Bing vs Google Search.
Better than have the whole product fail because competitors won the market with a global launch.
That's just how it works. That stuff takes months, so do you prefer to withhold it from the US until that's all completed? In a competitive environment that moves quickly, that's not an option.
While Google was caught flat-footed and has had to rush this in response.
I'm basing this just on what I've read, of course.
And if that's the case, I don't think my explanation is invalidated at all, nor "trivially".
I'm referring to ChatGPT and Bing integration, which is very recent. That was a big surprise to everyone AFAIK.
"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
Not a general vague sense of AI investment. You can't do legal review on a generalized "direction", no matter how telegraphed you think it is.
Which lawyer you get seems to be a sort of dice roll that perhaps one could understand if one were a lawyer.
I find it funny when massive corporations say they can’t legally figure stuff out and the other massive corporations do it.
It must feel back to be the lawyer who can’t work this stuff out when other lawyers can. Perhaps similar to when I think some programming is impossible and then see it done in the wild in production.
That just shows how far ahead Microsoft is on that one.
It didn't know the difference between a month and a week, and when I asked it a coding question, it said "I'm just a language model, so I can't help you with that.".
Lol, lmao even.
- Google+
Correct me if I'm wrong, but aren't ChatGPT and Bing Chat already worldwide releases? So it's not as though anyone is sweeping in after Bard. They were out before Bard and global from Day 1.
Besides that, you're absolutely right, of course.
Can't wait to ask "Bard" what's the rationale to "not support" a service like this based on my country..
These "global companies" are so global regarding their revenue and tax schemes, but when something is different than the "business as usual" they turn into hyper local shops. It's just an observation, I'm not really that mad and desperate to try their "chat-bot" since I'm sure it will be as exciting as eating cardboard.
Not being in the EU doesn't help either.
Even at FAANG jobs, we hired contractors to translate things out of en-us, so that could be a factor for Google to get this out.
Also everyone else’s point about limiting legal discussions. They’re probably using this data for training purposes and don’t want to deal with European laws. If they’re hosting the data in limited data centers, they may also have to deal with data locality laws.
Google+ was mediocre and not better than the competition. Young people didn't care about it. The exclusivity was generated by a company - it didn't appeal to the user's inherent bias. Facebook initially appealed to narcissism - the early users "knew" they were "better" than myspace users.
What was Google+ appealing to? Being a nerd?
Reader and Google+ were services I really enjoyed using for content.
Now that I think about it, my whatsapp groups intersect in a similar manner, and people post (manually) the same thing to multiple groups, but not every group we have in common.
> Conflicts regarding Google+ began in July 2011 when the social networking site began enforcing its real name only policy by suspending the accounts of users it felt were not following the policy.
That's nothing like the Google+ launch - making that a closed beta was death from the start because you could only interact with the few other people in the closed beta.
The point is, invite-only isn't inherently counter-productive to hype. But you need to bring the goods.
Bard, on the other hand...
[1] https://open.spotify.com/episode/0oRAHcP9g41jcEzpDtOa4b?si=1...
It didn't work out so well when they tried to reproduce that success by making Google Buzz invite-only. My feed consisted of a few people posting "buzz!" and then never posting again. Google+ was similar.
My opinion is that they are hostage of the image/ego they built around themselves, to the point where they say "incredible" things like: "ChatGPT? oh yeah we have those things internally and ChatGPT it's child's play comparing to what we have. We just don't release it to the public because we are responsible and the good guys.. We also have a girlfriend but she's in Canada."
My take is, regardless of how "good" Bard is, they will make a fool of themselves because their past, present and honestly, people are kind of fed-up with their "Googleness" bs..
The waiting list to the waiting list to the waiting list…
Of course I agree that this is high-risk in that their location thingy is unstable and trigger happy enough as is, and good luck reaching a human if some system does decide to block you.
Or as the old saying goes, “if a google product is felled in a forest, and there’s no users to notice, did google really fail?”
If we consider the desired outcome of a particular thing, then failure to achieve that outcome cannot be considered acceptable. However, we must allow for a sufficiency of time wherein the thing will progress and improve before coming to a conclusion.
On the other hand, if we contemplate the impact of a thing on its intended audience, then it must meet their expectations and desires to be deemed as effective. It is important to note that the relevance of a thing to its observers may depend on various factors such as timing and perspective.
Ultimately, effectiveness is determined by the desired outcome and the audience it serves. We must assess the purpose of the thing and its surroundings to determine its success or failure. It is similar to crafting a powerful message - we must consider the intended recipients and create a message that resonates with them.
I'm guessing it's purely regulatory?
I really get the feeling that Google is terrified of success. They haven't launched a new sucessful product in God knows how long.
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.
Incredible how much they've fallen from grace. Their image has suffered so much that people just don't care anymore.
Between corporate users going all-in on microsoft software + apple hardware, people in private only really have one thing thats flocking them towards some google, and that is... android. This and the longform tiktoks called "youtube".
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=...
Moreover, even if there is a reason to georestrict it, there’s no reason to georestrict a waitlist - that’s just bizarre.
I'm sure the majority of people search for actual questions and not just keywords like previously - and take the first hit, so the process of selecting a link and reading past the initial noise / ads (like long intros in youtube videos) can become a thing of the past
These are not for venting.
In pg’s seminal essay “On Complaining”, he identifies five levels of venting from simple criticism to full on meltdowns.
Bard is completely beta and a raw one at that, not something I'd want confused for a part of a paid offering at this stage.
I get your point as well - I think there's a tension between the two options.
Any connection you draw between native language and region is tenuous at best.
Similarly, it is increasingly common to get two Spanish tracks… one for “Latin American” Spanish (which is what you almost certainly learned if you took Spanish in school) and European (aka Iberian or Castilian) Spanish.
One of the most obvious is that ce/ci, and some z’s are pronounced “th” and not “s”. So a Spaniard would say “cinco” as something like “thinko”.
This why it’s “wor-es” (Juarez) Mexico, but “her-eth” (Jerez) Spain.
Is that hyperbole? Because that would be very sad if true, culturally speaking. There are 5 different official languages in my country (Spain) and there are several laws to prevent them from becoming extinct for this very reason.
I guess I should say I'm counting adults. Everyone who went through mandatory schooling was taught English. Everyone who didn't (immigrated) is more likely to know English than Swedish.
So the only people who know Swedish but not English would be those who flunked out half way through mandatory schooling, or immigrants who knew neither English nor Swedish, and only learned Swedish.
Swedish is not disappearing (c.f. your example in Spain), but basically everyone speaking Swedish is (at least) bilingual.
Is that how you call trying to protect consumers over there? Really, there are genuinely good reasons for those "stances".
A startup is free to release globally, because nobody will care if they spit out something false, there's no money to extract for penalties and fines.
If Google's Bard says one thing wrong in the wrong country, it has the potential to cost the company hundreds of millions or even billions of dollars.
In places like the EU, governments see Google as a giant piggybank of funding. They know Google will not pull out of any region as large as the EU, so they will use any excuse to introduce fines and penalties. They also know Google can, and will pay. Hence they've hit Google with nearly €10B in fines over the last few years.
Microsoft could completely kill Bing in the EU with very little effect to their bottom line.
But you do bring up a good point. Why be so conservative if Bard isn't even tied in with Google search?
My only guess is Google turning into a global whipping boy for fines has made them trigger shy.
Those fines were clearly not high enough.
Guess I won't try it then ¯\_(ツ)_/¯
You can use workspace data with open ai because there is no agreement in place restricting the use of that data.
"Bard isn’t currently supported in your country. Stay tuned!"
Back to GPT-4 then!
For bard specifically, you can do ?authuser=x where x is the account number as defined by other apps like Gmail or Drive.
ChatGPT Plus is very bad value vs ChatGPT API (particularly now that 3rd party chat clients exist)... I've spent $20 on ChatGPT Plus in the same month where I've spent < $1 on the API and used it significantly more via my github bot.
My subscription is due for renewal this week and I was going to cancel it, but for me it's worth it short-term for GPT-4 access. The increased context size and improvements in reasoning & general output quality are significant for my use cases.
GPT-4 access in ChatGPT Plus is quite limited (was originally 100 every few hours, currently 25 and going lower since they're struggling to keep up with demand) but that's still enough to get useful results out of it currently, especially if you have built up some intuition on how to prompt using 3.5.
all recent google product launches — are disappointment. Google is slowly dying.
It needs to change its virgin management for new chad one
You would prefer companies know your citizenship and immigration status, somehow verify that, and tie their services to it? And that would somehow preserve your privacy better than geolocation?
And then, when you go to change the language, it shows all the language options in the language you are trying to translate from. So useful for people travelling to countries where they dont speak the language, or cant distinguish the characters...
Geolocation is a sucky way of deciding anything about a user.
Look, the problem here is that internet providers are trying to follow the law ("avoid fines" if you prefer the cynical framing). The law is different in EU countries than it is in the USA, in some very important ways. And they need to be seen making a good faith effort to enforce those laws. A six month grace period for an account "primarily used" in one regime seems not terribly unreasonable to me.
But that's not what your original comment was asking for anyway. You wanted to be sitting in Barcelona or wherever and e.g. not be subject to the GDPR and EU-specific data retention requirements (which is presumably the issue here: Bard likely isn't yet compliant being probably all hosted in the USA).
And if all you have is $10mm on your checking account but no income, you don't pay any taxes anywhere.
Wealth taxes are a thing too, e.g., in Switzerland.
https://www.irs.gov/individuals/international-taxpayers/freq...
<quote> 1. I’m a U.S. citizen living and working outside of the United States for many years. Do I still need to file a U.S. tax return?
Yes, if you are a U.S. citizen or a resident alien living outside the United States, your worldwide income is subject to U.S. income tax, regardless of where you live. However, you may qualify for certain foreign earned income exclusions and/or foreign income tax credits. Visit Publication 54, Tax Guide for U.S. Citizens and Resident Aliens Abroad, for additional information. </quote>
Typical case is when you work for a multi-national company and they send you to work in a different country for less than 183 days, you will keep paying taxes in your own country despite your legal employer is now the legal entity in the new country.
Not available in your country. Sigh.
Cool, bye!
> Bard does not currently support Google Workspace accounts or when our systems indicate you may be under 18
OpenAI takes my Google Workspace Account just fine. Nice job Google. Locking out a user who is on a >$25 enterprise plan with you.
Why is every interaction with Google so stupid? Why can that multi-billion dollar company not deliver any good software?
An early release is not worth the time and effort for professional support.
OpenAI doesn't have those constraints with a google workspace account.
LOL, yeah sure, the world doesn't wait. This is dead on arrival.
Now the question is if I'll actually be able to use it...
Because you’re a Google One member, we’d like to offer you the opportunity to
be among the first to sign up for the new Bard experience and provide feedback.
Think of Bard as your creative and helpful collaborator, here to bring your
ideas to life using generative AI.I'm not impressed.
Yes, I understand the pressure is huge. But all those "soon" are not helping.
¯\_(ツ)_/¯
Whats the point of this big announcement then? Google is such a caricature of corporatism and blunders, it’s just sad.
This is their lifeline in a potential Bing + ChatGPT world, this product right here is supposed to save them from actual competition, and yet they act like complete and utter idiots.
Also, you have to have a Google account.
Thats a bit weird ?
> Bard isn’t currently supported in your country. Stay tuned!
Why does silicon valley hate Canada?
So no, no they didn't release it.
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.
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.
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.
"PRISM collects stored internet communications based on demands made to internet companies such as Google LLC"
If you're thinking of PRISM itself, I'm not sure in what sense Google is "part of" it? Google sometimes provides data in response to court orders, but there's not much they can do about that.
Google was sitting on this since 2017.
Ethics and possible business interruption are both very good reasons for them to be safe.
And since they released the paper, i think ethics was a much bigger concern than I'd normally think. Hell, even they don't even fully understand the math behind it since it was found by AI as far as I know.
Microsoft and OpenAI benefiting from it feels a bit off.
Why do they even limit access depending on the country? It’s a chatbot, not a financial service.
I can't imagine why even their waitlist itself needs to be geo capped completely, just let us join, release it to some US users first then sign me up when it's ready globally.
Basically, different jurisdictions recognize different individual data and privacy rights. Sometimes the requirements even conflict. It's not obvious that there is some universal way to treat data that is acceptable by all, nor what it would be if it did exist. Cookie banners are an example of this going horribly wrong. EU wants to "protect people" online so they require websites to ask consent to use cookies when every browser already allows users to configure cookie policy defaults and on a per-site basis in the settings. Everyone interpreted GDPR as some big movement towards user privacy and adopted cookie consent banners for all users as the "right thing" and the world is worse for it. EU is backpedaling on the banners. Multiple browsers offer features that block them outright.
You get the point.
Here it is just a site with nothing related to region.
Bard is probably not optimized and would completely overload the hardware Google has available for it should it be released worldwide.
That just shows how far ahead Microsoft is on that one.
It’s one of the costs to better privacy controls. You’re not going to get the latest hotness because teams have to think through the risks.
>Stay tuned!
No, I don't think I will.
nice /s