> Okay, let's tackle this roast based on the user's data. First, I need to pick the three most unusual bits. The user is using Safari 17.6 on Mac OS 10.15.7. Wait, macOS 10.15 is Catalina, which is pretty old now—like from 2019. That's a bit outdated, so that's unusual. Next, the screen resolution is 1024x1366. That's a taller screen than usual. Most common resolutions are wider, so maybe they're using an older iPad in portrait mode or a rotated display? That's quirky. Referrer is empty, meaning they came directly here. No referral site to make fun of, but I can joke about them typing the URL manually. Also, their location is Calgary, Canada. Maybe link the cold weather there to their outdated OS, like keeping things frozen in time. Combine all that into three concise sentences. Start with the outdated OS, then the screen resolution, mention the direct visit, and tie in the location with the OS. Make sure it's funny but not mean. Check the structure: three sentences, each targeting a different aspect. Let me put it all together.
> Roast: Ah, a Canadian Safari-saddled NPC from Calgary, still clinging to Mac OS 10.15 like it’s a Blockbuster membership, browsing on a screen taller than your mountains but somehow arriving here directly—did your dial-up finally finish loading the URL? (Three quirks: ancient OS, vertical screen ratio, empty referrer + location roast.)
It guessed correctly I’m using an iPad and maybe the resolution but somehow got the OS wrong (i.e. if I’m on an iPad I’m not using MacOS). I am still on iOS 17 so it got that I’m not on the latest but made a mistake since 10.15 is the latest MacOS release. The location is also way off.
While I'm incredibly glad we're getting open source LLMs with chain of thought, I'm not really impressed by its current use cases. LLMs are all same-y and it feels like we're getting loafs of cheesy language spam without much sci-fi future.
I expect for this kind of tech to have really novel use cases. For it to sit between me and the internet and remove the ads, nuke time wasting clickbait, and obliterate low-information irrelevant noise. For it to be my personal bodyguard that protects me from any and all forms of attention stealer.
I want this to be tech to give birth to an anti-Google, anti-social media algo, anti-advertiser terminator from the future. Something that torpedoes the previous paradigm and that does it so quickly that the old purveyors can't adjust in time.
It would be delightful if in 10 months from now, internet advertising no longer worked at all and that everyone adopted these same protections for themselves.
It seems like this tech is capable of doing it already.
Please someone make this. I want advertisers to pay me if they want access to my brain.
I've been working on exactly this. Great to know there's some demand for it.
It's now my daily driver for web access. It monitors for content I'm interested in (that's how I found your comment), handles all my searches and feeds, can dynamically adapt its interface, and is working on integrations to submit content for me so that I don't have to leave that interface to write these replies.
Hoping to release it early March but I'm a bit stuck on how to position it. I'm not a marketer and I think it'll just get drowned in the sea of "agent" slop.
Having a full-on buffer between yourself and the internet is a breath of fresh air but I have no idea how I would have found such a thing if I hadn't made it.
Maybe a curator or a bouncer or something. Open to suggestions if anyone has some.
A browser extension? Mega-awesome AI Adblock is like the best possible thing that could happen to the Internet right now and that seems to be good player for the cassette.
Is there a GitHub or Discord or anything where I subscribe to the project?
This sounds great. In a sea of paid ad placements and algo-recommended content, it’s hard to “surf the web” like I did in the early 2000s. It was a challenge to sit on googles homepage and think to myself “Hmm, what do I want to look up today?”. It required forethought of what I wanted to be served for content, and I always received proper content back to me after a search.
Anything that gets me closer to that original intention requirement before getting served content is a must-have in my books! Sign me up!
> It would be delightful if in 10 months from now, internet advertising no longer worked at all
Internet advertising will work as long as the internet exists. Classic targeted advertising on the other hand will no longer work but all it means is that it will be much more subtle and you will struggle to tell it apart in a way that you don't nowadays. Sounds a far more dangerous alternative
> Internet advertising will work as long as the internet exists.
I don't know. You're probably right, but more and more it seems like mainstream websites aren't meant for human consumption. The less ads a page have, the more likely it is to have relevant content.
Classical targeted ads, where some sales person actively goes out and buy ad space on a few select locations, I feel like that's more valuable than ever, but not heavily used anymore.
> It would be delightful if in 10 months from now, internet advertising no longer worked at all and that everyone adopted these same protections for themselves.
My bet is that hosted LLMs will have advertising baked in and also understand you in a much deeper way so as to manipulate you even more effectively.
iPadOS’s Safari user agent identifies itself as macOS, that’s why. Prior to iPadOS it did contain the string iPad; that was removed. So it can’t tell if it’s an iPad or a Mac with a small vertical screen, although a human would consider the former more likely.
As far as my Chromium is concerned, it identifies as a...: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36
The only thing it doesn't identify as is Internet Explorer.
(Yes it's an old version, no I will update when I feel like it. Shush.)
Don’t know about your particular problem, but I’m pretty sure it was largely a response to iPads being sent to mobile sites with horrible usability thanks to navigator.userAgent.match(/iPhone|iPad|iPod/) or the server-side equivalent. The problem was amplified by the release of the 12.9’’ iPad Pro, so self-identifying as a Mac made sense.
Well I’m glad Apple prioritizes us over you the. I don’t care that you want it a different way. Apple forcing the end of bad practices in this instance is a win.
It is not fair to judge the model performance on those guesses, if you have a look at the website you will see that the model is being told those parameters (os, screen size, etc). The model is not guessing those. Only thing that it might be guessing is your location
> Roast: Ah, a Canadian Safari-saddled NPC from Calgary, still clinging to Mac OS 10.15 like it’s a Blockbuster membership, browsing on a screen taller than your mountains but somehow arriving here directly—did your dial-up finally finish loading the URL? (Three quirks: ancient OS, vertical screen ratio, empty referrer + location roast.)
It guessed correctly I’m using an iPad and maybe the resolution but somehow got the OS wrong (i.e. if I’m on an iPad I’m not using MacOS). I am still on iOS 17 so it got that I’m not on the latest but made a mistake since 10.15 is the latest MacOS release. The location is also way off.