When you buy an iPhone that was made in sweatshop, you contribute to it. There are alternatives but most people really don’t mind having slave labor build their iPhone.
Same for kids wasting their life on screen, most parents don’t mind as long as it keep them busy and it’s one less thing to care about.
I had a favorite benchmark, SWE-rebench, but sadly it's no longer maintained.
But yeah, I'll just take these benchmarks with a grain of salt. Only hands-on experience matters in the end, and these days it's very easy to switch models.
I don’t know if they did or didn’t try to kill bookmarks.
But on the commercial incentives, it does make sense for them to kill it.
Companies pay good advertisement so they show up first on the search result when a user want to find them.
Eg you want to go to carglass website: you type carglass in google and the company will most likely pay Google to show up first in the ads section because competitors are trying to sway away customers from them. If they didn’t outbid their competitors, even though the customer want to go to carglass, he will click on first result and may end up buying from their competitor without even noticing.
That wouldn’t happen if people had put carglass.com website in bookmark.
Google doesn't show ads on navigational queries because of the potential trademark-infringement legal issues that could result from recommending a competitor when somebody searches for a trademark.
Search for [carglass], no ads. Search for [facebook], no ads. Search for [spotify], no ads. Search for [american airlines], no ads. All of these would be very desirable queries to advertise a competitor on.
(IIRC they've gone back and forth on this depending on what the lawyers thought they could get away with; ISTR when Bing launched Microsoft would run ads on [google] that pointed to Bing. It's been a long time since I've seen an ad on a navigational query though.)
As a counter example, a couple days ago I was to install DaVinci Resolve on a new windows installation, so I searched "davinci resolve." Since I was used to having adblocker, I just clicked the top result, which is titled "DaVinci Resolve" but actually takes you to some awful shady thing like wondershare or softonic/uptodown, I forgot.
I feel like, navigational queries or no, Google should be held liable for allowing scam sites like that to advertise. I think there was a trending HN post recently on that front.
> As a counter example, a couple days ago I was to install DaVinci Resolve on a new windows installation, so I searched "davinci resolve." Since I was used to having adblocker, I just clicked the top result, which is titled "DaVinci Resolve" but actually takes you to some awful shady thing like wondershare or softonic/uptodown, I forgot.
I just tried this query on chrome / android.
This is the first link on the page at the very top:
When people tell me this stuff about their Google searches it drives me insane, I want to know why our experiences are so different. It is regional? Is it OS, is it browser, is it something else? I want to know!
You know your search results are personalized right? If I google something and my girlfriend googles the same thing, on the same network, we get different results.
>You know your search results are personalized right?
Of course.
But these results don't seem to change on a degoogled chromium on desktop, with no google account login personalization, no account logged in, in incognito, using: Ublock origin lite, User Agent / OS Randomizer, Privacy Badger, and using a VPN to other locales.
That's why I want to know, I've been curious about this for years, and nothing I've looked at seems conclusive.
My biggest guess is country or native language, but it's been hard to pin down. I think that may be the biggest factor (with local laws contributing maybe to how search results get displayed).
I'm curious! I want to know the driver behind the bad experiences people are having with it. (I've been using it literally since it was released.)
This a search domain term that means a search query where the user is trying to find a specific destination/product. I don't know how Google search works, but often a search will first go through an intent classifier and then use a different model based on the query type.
A "navigational query" is a query with a certain website in mind. So like if you search for [facebook], 99.99% of the time you are looking to go to Facebook, if you search for [hacker news] 99.99% of the time you are looking to go to news.ycombinator.com. Basically it replaces the address bar for people who don't know what URLs are. You aren't looking for information about the business, you want to go to the site/business itself.
Just saying, the author blames corporate internet and the algorithm that feed you content, but if you browse 2 articles before (this one: https://alicegg.tech/2026/06/12/popular-upcoming) he then praised steam for removing the “upcoming game list” that was purely based on the number of wishlist a game received and is praising the new algorithmic personalized calendar of upcoming “for you”. I find it a bit hypocritical.
People use these algorithms because they love them, and yeah someone posting once every 1 or 2 month is not going to make it in these days and ages. I don’t think the internet became worst, it just adapted to other needs.
Grapheneos with Google play services (sandboxed) only send 12 packets per hour, instead of around 350.
That’s great, but the sandboxed Google play services are by default set to normal in battery usage. Meaning it’s subject to dooze and other ways to artificially limit its background activity.
You neee to set the battery usage to unrestricted to reliably revive notification.
I wonder what the packet count would have been with unrestricted Google play services.
I'm less interested in the quantity, and more in the content of those packets. Do they contain my GPS movement history? A list of installed and opened apps? My contacts list, messaging text log, and list of visited websites?
The data I really care about protecting would fit into a few kilobytes, so knowing the phone sends, for example, a minuscule 2 MB/day, brings me no comfort.
Google play services on ‘stock’ android has access to absolutely everything. Including list of apps.
Obviously they send back this data to Google, for legitimate and advertising reasons. For instance, how would they backup apps you have installed if they didn’t know what you installed in the first place.
On grapheneos, it doesn’t have access to this data, but it can still gather a lot of data (just as regular apps on Android). Location (even without location permission), WiFi devices around, Gyroscope so they know if you are standing, driving or else, etc…
I don’t know in the USA but in France, if it’s deemed that you launched a lawsuit knowing very well it wouldn’t succeed, you are susceptible to get a 10k€ fine. Even jail in serious cases.
Yes. And you(r lawyer) can collect lawyer's fees and you can be made to pay the court fees, if you lose.
IME the American legal system is set-up to discourage litigation, though. A common tactic is to bury your opponent in the threat of heavy damages or jail-time to get them to settle for what you were originally after, which courts are perfectly happy to facilitate because it gets a potentially lengthy trial off their dockets. They'll punish (or be biased against) whichever party seems responsible for not accepting a "reasonable" settlement.
Algorithmic abuse of the system to extract payments already exists in the form of the debt collection industry.
A common tactic is to bury your opponent in the threat of heavy damages or jail-time to get them to settle for what you were originally after, which courts are perfectly happy to facilitate because it gets a potentially lengthy trial off their dockets.
That's not how litigation actually works in this country. Unless the plaintiff can actually prove damages, there's no threat of "heavy" damages. The primary economic concern is spending a lot of money on legal fees, which is the primary trigger for getting businesses to settle instead of litigate. It doesn't work so well against people though.
They'll punish (or be biased against) whichever party seems responsible for not accepting a "reasonable" settlement.
That is false and any judge caught doing that would be removed from the bench, immediately. The judge does not exercise any judgement over what a "reasonable" settlement is as they don't and wouldn't know what is reasonable until the facts have been presented...at trial.
Algorithmic abuse of the system to extract payments already exists in the form of the debt collection industry
The debt collection system doesn't use the court system at all. Assuming you are referring to individual debts, not corporate debts, debt collectors have a very low success rate in court since (a) they have to prove the debt is valid, (b) that they now own the debt contract, and (c) they did not violate any laws in the pursuit of collecting on that debt. A and C are actually really hard for a third-party debt collector to prove.
Indeed. Realistically a 'capable' small local LLM, even one that's definitely not as good as externally hosted ones will require a single 16GB GPU and access to basically all of the RAM on the GPU. That's not something people running Firefox on a $500 laptop with 8 or 16GB of total system RAM and a CPU-integrated basic graphics system have to spare.
No it doesn't. TODAY you can run with good performance a local LLM entirely on CPU that can do translation, summaries, search assistance, and all the things mentioned here.
On a 8GB system with a SSD it might be a bit slow due to having to keep experts on disk, but perfectly usable. On fast systems it may well be faster than the network round trip for many tasks.
> No it doesn't. TODAY you can run with good performance a local LLM entirely on CPU that can do translation, summaries, search assistance, and all the things mentioned here.
What specific LLM and how much RAM does it take up to load? How does it run on somebody's 8GB RAM $600 laptop they bought from Best Buy? What pp/s and token/s rate on that hardware?
How much ram has a fair amount of flexibility since MoE can be kept on flash and swapped in at a performance cost, and depending on how small a context you can use.
Well the linked thread mention 14700KF which is a 400$ processor and RTX 2070.
Beside its mentioned 5 or 6 gb of ram.
People run Firefox on pentium with less than 8gb of ram, just 4 or 5 years ago 4gb computer were still extremely common. Even Microsoft sold its surface devices with 4gb of ram, I was dumb enough to buy one.
To give a counterpoint, you could disable it on those low end devices.
But then people would complain about the size, like what happened with Google Chrome embedded llm.
I think there wasn’t any bad choice, mistral was the safest.
My post was an earlier post in this thread. I give figures for pure cpu, not using any GPU. The RTX 2070 figure is a separate figure (with rather insane speeds for that model!). You could contribute to the discussion by grabbing some numbers on lower power hardware if you have some available.
Though low memory devices like 4-8gb probably does need specialized inference to give fair numbers.
> you could disable it on those low end devices
it's not like it does anything if you don't ask for it-- and good thing, because the privacy invasion would be all the worse if it did!
> mistral was the safest
Sending the user's confidential information to third parties while falsely suggesting that it is private can cause them irrecoverable harm.
An alternative is that a feature is slow for some users on slow hardware, and there is a setting to make it faster at the expense of privacy and security that they can switch on. The large body of users that find the default config fast enough have no reason to flip the switch.
> But then people would complain about the size, like what happened with Google Chrome embedded llm.
People or astroturf accounts? :P but having read some of the commentary on it there was some pretty weird takes, like generalized complaints about AI and people thinking google was using their computer to serve other people. Google's business model generally prevents correctly marketing the functionality in any case: it's not like they're going to make a proper pitch for how important it is for your privacy when the rest of their business is centered on hoovering everything up. Mozilla is not so constrained :P
reply