I wish the folks who clearly do not like Google would just not use their products instead of spamming every thread about how they will kill the product, true or not.
——
Anyway,
It’s not clear which model they’re using for this. I assume whatever Bard is using, but who knows. This is relevant because depending on the intended experience the latency will matter.
Overall it’s not a bad idea, but I do wonder what the monetization path will be for Google. I imagine this will be part of workspace. Perhaps they will add more tiers to include these offerings.
I wish they shared a bit about how this will be differentiated from Bard. Is this simply a new front end to Bard? It’s really an open question. I haven’t seen many products that use LLMs that are better than the prompt response UX.
The most interesting thing about this blog post is the “source grounding.” I’m curious if there’s actual engineering behind it, or is it prompt tweaking contextualized behind the scenes on a given doc.
People are wary about Google product launches because of a well documented[1]
pattern of corporate behaviour[2] when it comes to either the project's planning and execution, resourcing, or long-term support.
You don't have to agree with those comments, but they are legitimate concerns being expressed by people who have been burned enough times by the company that it has developed a degree[3] of notoriety[4] for it as a result.[5]
To be fair, there are a lot of low-effort comments here to this effect. I agree with you and will happily upvote the first well-written comment pointing out Google's self-inflicted reputational damage, but the myriad one-line jabs aren't adding anything.
I'd normally feel the same as you about the "one-line jabs" but in this case, IMHO they help to underline the fact that Google really did screw a lot of people over.
Like I mentioned, don’t use google products then. What’s the value of spamming this all of the time? The validity of the concern isn’t the problem, it’s the fact that it’s irrelevant to the content being posted.
How is it irrelevant? If a new project/product is being hyped by a company with a well documented history of hyping then abandoning projects and products, it's objectively a relevant comment.
For example, if there's a post about the new Threads app, and someone comments about Meta's record on, say... security, data privacy, moderation or support of Instagram, WhatsApp, and Facebook, or something about their failed projects like Beacon, Sponsored Stories, Creative Labs, and Facebook Credits, those are likely going to be relevant comments because the corporate culture and past record on projects and products reveals something about what we might be able to expect from Threads.
It's not irrelevant, it's just incredibly tiring reading the same comment over and over again on every thread that remotely mentions Google launching a new product.
> For example, if there's a post about the new Threads app, and someone comments about Meta's record on, say... security, data privacy
- Ctrl+F "security". Exactly 0 comment hits (on the first page).
- Ctrl+F "privacy". There are three hits, and two of them just mention "privacy policy" without making a comment about Meta's previous privacy violations.
- Ctrl+F "moderation". Exactly one hit about how Meta has historically been bad about moderation.
Let's look at the current post about Google launching this product (NotebookLM). Ctrl+F "kill". There are at LEAST ten comments about how Google will kill this product off soon. It's not comparable to the discussion re. Meta and its history.
Google killing off products is table stakes at this point. Imagine every time you clicked on an HN discussion about Python, you had to read ten comments about how Python is slow and if you want a faster language you should use Rust. "Python slow, Rust fast." Yes, we get that already. Can we just skip the memetic comments about "Google bad cuz they kill stuff" and read discussion with more meat?
I meant the Threads point as a hypothetical example not a literal one. I see annoying and repetitive comments on HN all the time. I don't have to upvote them or engage with them at all. I can merrily skip on by. Unless they're being abusive or threatening, I don't see what the problem is?
Clearly enough people think Google's record of hypeware-to-abandonware cycle is worth raising again and again because it's a frustrating aspect of the company's 'innovation' process and HN is a fairly influential place to keep that issue alive in the hopes that it might lead to some change or at the very least inform people who are considering investing a lot of time, energy, and capital into depending on such projects. It's also not exclusive to Google[1], though perhaps the one most notorious for it.
I've also found those "Python slow, Rust fast" threads genuinely useful because they prompt people to report on progress in Python on that front, or indeed even how the two languages can work together to play to each other's strengths.[2]
Eh, as a ride or die google reader fan, when there's six "DAE google reader" comments in a row, you all just read like lazy comments looking for some karma and detailing any (potential) interesting discussion before it starts. That's my main draw here, so definitely a reflexive down vote from me.
To warn users so they're aware that the lifespan of the product they're using may be very short, and therefore they may lose something they put a lot of investment in. It's empathy.
Nobody reading HN needs to be made aware of it, and those are the only people you'll be able to reach with a HN comment.
The thing is, these repetitive comments get spammed dozens of times on every Google-related post. They could be posted without reading the submission. Many of them probably are in fact probably posted just based on the title (seeing how many people posted basically exactly the same comments into a submission today that wasn't even about a product). These comments completely crowd out any other discussion just due to sheer volume.
There's 90 comments on this thread at the moment. Maybe like 5 are discussing anything about the article itself. The rest are just totally predictable and generic comments, which is something the comment guidelines specifically ask to avoid. (And yes, I'm now also participating in that.)
Now, fair enough. Maybe there really is nothing to be said about the concept of the product. There's definitely not much to be said yet about the execution given nobody here will have been able to use it, so that definitely limits the scope of meaningful discussion to be had. But since the spam won out, we'll never find out.
If there's nothing of substance to talk about, we could just talk about nothing and let the obviously boring and uninteresting submission die in silence.
And if you need to warn your family and friends not to use this product because you think it will be cancelled too soon, you'll probably reach them more effectively via methods other than HN.
The HN Guidelines[1] also say that commenters should:
"Respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith."
Just because people have posted similar comments about Google's penchant for killing projects, doesn't mean that they are "spamming". Consider, instead, that it may just reflect a groundswell of opinion borne out of bitter experience and disappointment about the company's record and their genuine concerns that this is yet one more tombstone in the expanding Google graveyard.[2]
Sure, let's take it as a given that everyone who posts wholly original messages about "lol this product will be canceled" are doing so purely because of the harm they have personally suffered due to a product cancellation, and they just missed the previous messages on that subject. So what? It's still turned this thread into a wasteland and prevented any productive discussion.
Posting in good faith isn't a free pass to post off-topic, repetitive ir low-effort stuff. Just think about what the outcome of that kind of policy would be if applied across the board.
For example, a lot of people have very strong opinions on divisive political issues. Should they be picking submissions that are tangentially related to those issues and hijacking the comments section for discussion about that subject? No, of course not. And I'm not saying that because I disagree with their hypothetical political opinions. I'm saying it because if that becomes the norm, we can't have a discussion about the actual thing that 200 people thought was interesting enough to upvote, because all the discussion will be a retread of that political issue.
It's as if Rust evangelists posted comments about "this bug would never happen in Rust" in response to a technical deep dive to a security bug. Yes, I'm sure they'd be right. Yes, I'm sure they'd be posting that in good faith. There's even a small chance they'd be posting it because they think there's still HN readers who haven't heard the good word of Rust. But it'd also be a shallow comment, variations of which would be posted hundreds of times, and totally generic in that the exact same comment could have been posted to any of a hundred other submissions about a security issue.
> So what? It's still turned this thread into a wasteland and prevented any productive discussion.
How exactly has it "prevented any productive discussion"? I've seen plenty of comments about the merits and limitations of its Bard foundation, the design choices, the potential business model, etc.
The example you give of political opinions doesn't feel like a fair comparison. We're talking about the company's record on supporting its projects in the context of a new project announcement not a divisive socio-political controversy.
A better example might be: If Red Hat announced a new open source initiative, would it not be relevant to bring up concerns about its commitment to the new FOSS project in light of the impacts its recent announcement about CentOS is having on other companies and projects that depended on it?
My favorite were comments on a post about Google Zanzibar, which is an internal Google infrastructure thing that they published a research paper about. Other companies started adopting that method, and there were HN comments warning about Google killing products. I don't think Google is going to cancel a research paper!
If a company behaves badly then I think it’s warranted to bring up their bad behavior whenever that company is the topic of discussion. In the U.S. we mostly have regulatory capture and so people need to fight back or otherwise hold companies accountable.
I don’t know if Google qualifies in my mind for this sort of action but companies like BP, Exxon, and others do.
If that's true, then you're only exacerbating it by attacking "said thread".
Furthermore, threads raising the issue of googles seemingly erratic depredation of their products and threads discussing the product itself aren't mutually exclusive. You can do this neat thing called scroll down, and not read these threads if they bother you so much.
That's the thing I don't get. Why don't people just ignore the comments they don't like and don't want to boost? Feels like some kind of Stack Overflow-esque over-persnickety moderation syndrome.
The value is that it provides context around googles substantial issues with long-term support for their products. So anybody who's new to hacker news or even new to the tech world can see that this is still an ongoing issue, and can avoid getting burned like so many of us before.
And given the forum, I like to think that there's probably a fair number of Googlers who see these comments, so it's always on the forefront of their minds when they may bring it up to their managers and their managers managers.
It’s not spam to some when they are reminded that google never was, a “do no evil” corporation.
Google has tarnished itself. Yes it’s the biggest. But this is because of a well documented cycle - and when evidence of this horseshit pops up “yet again” it’s important to remind everyone to not get too attached to anything google.
Bard isn't a model, it's a product. Bard was launched with PaLM originally, the same model that the one employee claimed was sentient. At Google I/O they announced PaLM2 and switched over to Bard being backed by it, while also making Bard open to the general public on the same day.
Google is currently in the process of training Gemini, which is positioned to be their GPT-4 killer, as it's multi-modal, and presumably quite a bit more powerful than GPT-4 is. PaLM2 is more analogous to GPT-3.
NotebookLM is a very different product than Bard. Bard has a vague understanding of the information on the internet, and NotebookLM will, presumably, have some level of understanding of your personal documents. It will almost certainly be using PaLM2 until Gemini is ready. That will be neat because Gemini will be able to look at images and read visual charts in your documents. PaLM2 is only able to comprehend text.
Right - to clarify I meant the model used by bard. I was too lazy to see what bard currently uses since it’s transparent to the user, to your point. Multi modal models being used would be killer.
But the monetization thoughts remain, this stuff isn’t cheap.
> I wish the folks who clearly do not like Google would just not use their products instead of spamming every thread about how they will kill the product, true or not.
Me, upon reading something about this: "Oh, this sounds interesting, I would like to have a go at something nea--"
observes it is released by Google
"Well, never mind, I'm not going to invest the energy into it until it reaches its third birthday."
Google could simply stop taking a scythe to products at a rapid clip and shed their reputation if they want to benefit from not being slagged for it everytime they release something new.
While I understant the feelings about Google wrt product longevity, I fail to see why that would prevent you from using this specific product as the 'investment' seems to be you uploading a Google doc and the payoff is instant.
I feel like most social media would be much more usable if there is some kind of browser extension that automatically clusters the comments and hides the same opinion/sentiments that have been repeated a dozen times in the same thread(you can label the clusters with comment counts of the same topic).
edit: Actually before someone submits a comment, they should be informed that how many people have said the exact same thing, like a repost warning.
> I wish the folks who clearly do not like Google would just not use their products instead of spamming every thread about how they will kill the product, true or not.
It seems weird that you think these people do not like google. Then why would they care? The uncertainty of the "google graveyard" only affects you if you do like and use their stuff. And fans in particular can be very gloom and doom about anything, e.g. see any esports game update and how it will completely break everything.
I want to like Google’s products, and used to be quite the Google evangelist, but the lack of commitment is one of several reasons I won’t invest time in them now.
——
Anyway,
It’s not clear which model they’re using for this. I assume whatever Bard is using, but who knows. This is relevant because depending on the intended experience the latency will matter.
Overall it’s not a bad idea, but I do wonder what the monetization path will be for Google. I imagine this will be part of workspace. Perhaps they will add more tiers to include these offerings.
I wish they shared a bit about how this will be differentiated from Bard. Is this simply a new front end to Bard? It’s really an open question. I haven’t seen many products that use LLMs that are better than the prompt response UX.
The most interesting thing about this blog post is the “source grounding.” I’m curious if there’s actual engineering behind it, or is it prompt tweaking contextualized behind the scenes on a given doc.