HN def is a community, in some sense of the word, but I also dont like the phrasing and emphasis on "we" having done something or we having a problem and whatnot.
HN promotes debate and individual opinions, so statements like in parent just have bad connotations to them.
> dont like the phrasing and emphasis on "we" having done something
Fair enough. I cited two we’s.
In the first—where “we, as a community, have been critical of crypto”—I juxtapose this community at large’s scepticism of crypto with its giving the benefit of doubt to people similar to crypto’s worst in AI.
In the second—where “we have a wider problem”—I intended to reference Silicon Valley at large. I did not intend to imply that Sam Altman is HN’s problem to solve. (Though I welcome the help, if you have sway in D.C. or your state capital. I am not sure I have the bandwidth to take on a political project of this magnitude.)
> I did not intend to imply that Sam Altman is HN’s problem to solve.
In consideration that Sam Altman was the former president of YCombinator, and Hacker News is financed by this company, it would be true hacker spirit (as in Hacker News) to subvertingly use Hacker News to "solve" the Sam Altman problem. :-D
Now imagine it happening for literally every interaction. Opened a context menu? Oh give me a second to check the LLM server for useful actions I can offer you. Want to close a window? One sec just wanna analyze the content first to see if you reealy want to close it.
I am so happy that most gaming is now possible on Linux and I dont have to use Windows at all.
I left Windows for Linux around when Win 10 came out. What prompted was the whole shenanigans to remove features I didn’t want and frustration with the settings (I like to change stuff every once in a while). I throw myself in the world of Arch, i3wm, and ricing. I’m using macOS now, but it’s quite nice to know there’s an haven out there once the abuse is too much.
I'd be surprised if any legal department in any company with one will not freak the f out when they read this. They will likely loose the biggest customers first, so even if it is 1% of customers, it will likely affect their bottom line enough to give it a second though. I don't see how they might profit from an in-house LLM more than from their enterprise-tier plans.
Their customer support will have a hell of a day today.
Whats baffling to me is why companies think that when they slap AI on the press release, their customers will suddenly be perfectly fine with them scraping and monetizing all of their data on an industrial scale, without even asking for permission. In a paid service. Where the service is private communication.
I am not pro-exploiting users' ignorance for their data, but I would counter this with the observation that slapping AI on product suddenly makes people care about the fact that companies are monetizing on their usage data.
Monetizing on user activity data through opt-out collection is not new. Pretending that his phenomenon has anything to do with AI seems like a play for attention that exploits peoples AI fears.
I'll sandwich my comments with a reminder that I am not pro-exploiting users' ignorance for their data.
Sure - but isn't this a little like comparing manual wiretapping to dragnet? (Or comparing dragnet to ubiquitous scrape-and-store systems like those employed by five-eyes?)
Most people don't care, paid service or not. People are already used to companies stealing and selling their data up and down. Yes, this is absolutely crazy. But was anything substantial done against it before? No, hardly anyone was raising awareness against it. Now we keep reaping what we were sawing. The world keeps sinking deeper and deeper into digital fascism.
Companies do care: Why would you take additional risk of data leakage for free? In the best case scenario nothing happens but you also don't get anything out of it, in the worst case scenario extremely sensitive data from private chats get exposed and hits your company hard.
Companies are comprised of people. Some people in some enterprises care. I'd wager that in any company beyond a tiny upstart you'll have people all over the hierarchy that dont care. And some of them will be responsible for toggling that setting... Or not, because they just can't be arsed to with how little they care about the chat histories of the people they'll likely never even going to interact with being used to train some AI.
i mean, i am in complete agreement, but at least in theory the only reason for them to add AI to the product would be to make the product better, which would give you a better product per-dollar.
Well I really hope this massively blows up in their face when all of Europe goes to work just about now, and then North America in 5-8 hours. Let's see if we have another Helldivers 2 event that makes them do a hard backpedal after losing thousands of large customers that will not under any circumstances take the chance.
I have a friend with a law firm who just called me yesterday for advice as he's thinking about switching to Slack from Teams. I gave him a glowing recommendation because it is literally night and day, but there is no way in hell he takes any chance any sensitive legal discussions leak out through prompt hacking. He might even be liable himself for knowingly using a tool that spells out "we read and reuse your conversations".
Out of those, Mattermost was the easiest to setup (just need PostgreSQL and a web server, in addition to the main container), however not being able to easily permanently delete instead of just archiving workspaces was awkward. Nextcloud Talk was very easy to get going if you already have Nextcloud but felt a bit barebones last I checked, whereas Rocket.Chat was overall the more pleasant option to use, although I wasn't the biggest fan of them using MongoDB for storage.
The user experience is pretty good with all of them, however in the groups that I've been a part of, ultimately nobody cared about self-hosting an instance, since most orgs just prefer Teams/Slack (or even Skype for just chatting/meetings) and most informal groups just default to Discord. Oh well.
The problem is not technical, but social with these platforms.
i.e. How do you convince 40+ people from 5 countries to add yet another memory resident chat application and fragment their knowledge to another app/mental space?
This gets way harder as the community becomes more dynamic and temporary (i.e. high circulation like students). I gave the good fight last year with someone, and they just didn't flex a nanometer citing ergonomics of Slack is way better than alternatives, and didn't care about data mining (was a possibility back then) or keeping older messages at ransom.
> i.e. How do you convince 40+ people from 5 countries to add yet another memory resident chat application and fragment their knowledge to another app/mental space?
If it's a company, you can just be like: "Hey, we use this platform for communication, you can log in with your Active Directory credentials."
It also has the added benefit of acting as a directory for every employee in the company, so getting in touch can be more convenient than e-mail (while you can also customize the notification preferences, so it doesn't get too spammy), as opposed to the situation which might develop, where some teams or org units are on Slack, others on Teams and getting in touch can be more messy.
If it's a free-form social group, then you can throw that idea away because of network effects, it'd be an uphill battle, same as how sometimes people complain about people using Discord for various communities, but at the same time the reality is that old school forums and such were also killed off - since most people already have a Discord account and there's less friction to just use that.
Either way, I'm happy that self-hosted software like that exists.
That's a big if, and the answer is "No" in my case. If it was, that comment wouldn't be there.
It's not a "social group" either, but a group of independent institutions working together. It's like a large gear-train. A lot of connections between small islands of people. So you have to work together, and have to find a way somehow. So, it's complicated.
> Either way, I'm happy that self-hosted software like that exists.
Me too. I happen to manage a Nextcloud instance, but nobody is interested in the "Talk" module.
> But you can opt out, right? So what’s the problem?
This thinking is the problem. "Oh, we just added your entire private/privileged/NDA/corporate information to our training set without your consent. What's the problem?"
Opt-out must be the default.
Edit: By "Opt-out must be the default." I mean: no one's data must be included until they explicitly give consent via an opt-in :)
Especially since once it has been trained, it is in the model, and I am not aware of any way anyone has discovered to later remove from the model single or selected training data points, except for re-training/re-learning the model. So basically the crime might already be done.
But I also know that so many businesses are too sluggish to make a switch and employees incapable of understanding the risk. So unfortunately not all of Europe will switch away. But I hope a significant number gives them the middle finger.
There's something called "machine unlearning" being worked on to address these issues.
This doesn't mean that I support Slack or any opt-in without consent training model. On the contrary. I don't have any OpenAI/Midjourney/etc. account, and don't plan to have one.
Worth noting: This is a legal requirement in Europe
The GDPR mandates that consent is given affirmatively, with this kind of "oh we put it in the EULA nobody reads" being explicitly called out as non-compliant.
The way to opt out is by contacting support,in an era where opt ins and outs should be handled by a toggle button.
Either they don't expect many people to wish to opt their slacks out, or they're aware of the asynchronous friction this introduces and they don't care.
What you say is true, and I agree, but that is the emotional human side of thinking. Purely logically, it would nake sense to compare the two systems of control and use the one with fewer human casualities. Not saying its gonna happen, just thinking that reason and logic should take precedent, no matter what side you are on.
It definitely seems like a matter of simple math. But, I'm not 100% sure it's always the most logical choice to defer to statistics.
By definition, stats operate at the macro level. So, for instance, I may be a safer driver than the AI average. Should I give up control? I suppose it's also a matter of degree and there's the network effect to consider (i.e. even If I individually beat the average, I'm still on the road with others who don't).
So it gets a little more complicated and I'm also not sure the aversion to relinquishing control is strictly "emotional" (as in the irrational sense). There's something about the potential finality of a failure that goes along with autonomy and agency over one's own life. The idea that a machine could make a mistake that ends your life, and you never had a chance or say in that outcome is off-putting in ways that feel more rooted in rationality and survival than in emotion.
It's not just opt-in, it's a non-default option you have to actively seek out and enable with every new conversation you start. So yes, by default, without additional steps taken, telegram is not e2e encrypted.
Wouldn't large body collisions be a fairly common occurrence in young planetary systems though?
Of course any additional condition makes life rarer, I'm just thinking this one might not make it rarer by as many orders of magnitude as it might look like at first glance.
Out of what, 500-1000 billions of planets just in milky way? I dont think folks do realize how big those numbers are. And there is no reason to ignore rest of the universe if we talk about probabilities and statistics
It's just another fraction to multiply in the drake equation. Start with planets, cut down to rocky ones, only in habitable zone, only rotating a certain way, only with plate techtonics, etc etc.
Or using your numbers,
1/4 factor eliminates 750 million possible candidates. That's not a happy thought.
I think the point is that we know the numerator today is one, and no matter how much you cut down the denominator, increasing the numerator to two would be a huge deal. Conversely, given that we think the denominator approaches a small infinity, it seems implausible that the numerator is actually one.
But, the 2 may be on the other side of the universe, or even past the cosmological horizon. And since FTL speed travel is almost certainly impossible, it may very well always be 1 to the best of our knowledge.
A single 1/10 factor makes it 1 trillion. 10-ish more factors and we're down to a small number of planets before we even consider the emergence of life and likelihood of propagation.
There are likely many more 1/10 factors: Habitable zone, diurnal cycles, billions of years of geological peace, sufficient water, a stable moon (maybe only one), a particular spectra of the star ...
all those were "well they seem to happen to 1/10 or 1/100 planets" individually. They cut down the space quickly when combined.
Even 1 in a million (planets) would allow for 100 thousand solar systems in the 100 billion stars strong galaxy to have at least one tectonic active planet; not counting moons.
As stated in another thread, it's yet another 80% reduction in the number of habitable worlds. On top of all the factors, the exponential decay is pretty steep.
That's just the nature of the drake equation. It's very much a geometric series and if all factors have to line up with 1/10 odds, you only need 10-ish "vital things" to cancel basically all chance of life except earth.
And here we have a hypothesis which might be about 1/10 odds and might be vital. 9 left. Water? Diurnal duration? Spectrum of star? A billion years of peace? A moon? A magnetic field? There's potentially lots of factors.
"basically" is doing only what I suggested: a geometric series.
Here's exactly what work it's doing:
Pretend there's 10s of trillions of planets. That's 10-13 zeros depending on whose estimate you trust most.
1/10 factor cancels one zero.
at most 13 factors accumulated means you have 1 habitable planet out of all those planets.
We have just hypothesized a 1/10 factor in this thread - that leaves 12 more - and I've lised 6 more off the top of my head.
It's just a fermi question - ballpark estimates like that are a way of thinking of the relative scale. A 1/10 chance seems like it leaves a lot of planets left (as you say - 100s of billions), but there are already many 1/10 factors floating around.
The Milky Way alone has about 2.5 × 10^11 stars. The Andromeda Galaxy has around 10^12. Let's take 0.5 × 10^12 stars on average per galaxy.
There are about 2.5 × 10^11 galaxies in the observable universe.
This gives us around 10^23 stars in the universe to fiddle with. Assume every star has an average of 2 planets; some have more, some have none.
This is a pretty large number to trim down.
I'd argue the Drake equation is excessively conservative. Note that when microbial life first emerged on Earth 4.1 billion years ago, the Earth's atmosphere was rather reducing, and the Sun was around 30% less luminous than it is today. There was free water, but no free oxygen, and an extremely high-pressure CO2 atmosphere.
The universe is arguably extremely young; the longest-lasting stars will only burn out around 10^13 years from now, and the universe is barely 10^10 years old. It's fair to say that many sun-like stars haven't even formed yet.
Right but the unstated assumption that there is no other positive path isn’t any more well-supported than the converse. For example, observe the variety of life we have locally. I’m thinking particularly of the various life (or life-like but I digress) that exists in extremes like thermal vents or under-explored places like deep soil. So maybe it does eliminate 1/10 but maybe we forgot to add the other 1/10 for life that wants to
live at 100C (or whatever) — I’m pushing back on your assuredness, not the math.
https://www.youtube.com/watch?v=ikg3-GQLg3g