vlovich123 4 days ago | next |

> 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 That's a bit outdated, so that's unusual.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 Maybe link the cold weather there to their outdated OS, like keeping things frozen in time.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 Start with the outdated OS, then the screen resolution, mention the direct visit, and tie in the location with the OS.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.

echelon 4 days ago | root | parent | next |

I got similar results and it was kind of hokey.

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.

wild_egg 4 days ago | root | parent | prev | next |

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.

svilen_dobrev 3 days ago | root | parent | next |

Electric Monk ?

from Douglas Adams' Dirk Gently series

"Electric monks believed things for you, thus saving you what was becoming an increasingly onerous task, ..."

EDit: or maybe, since it's not just believing, a Janitor?

benreesman 4 days ago | root | parent | prev | next |

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?

coldfoundry 3 days ago | root | parent | prev | next |

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!

netdevphoenix 4 days ago | root | parent | prev | next |

> 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

mrweasel 3 days ago | root | parent |

> 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.

vlovich123 4 days ago | root | parent | prev | next |

> 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.

arizen 3 days ago | root | parent | prev | next |

Ironically, the advancements in AI R&D and LLMs by Google and Meta were largely funded by their advertising-driven business models.

oefrha 4 days ago | root | parent | prev | next |

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.

VagabundoP 4 days ago | root | parent |

And that’s why so many kids had trouble registering their iPad on our network. As it would not send the mobile registration to the iPads

<sigh@Apple>

gkbrk 4 days ago | root | parent | next |

It's not an iPad or Apple issue if your registration system is using the User-Agent for weird reasons.

VagabundoP 3 days ago | root | parent |

Sure it does, but Apple should be sending the correct User Agent string, not making one up.

macNchz 3 days ago | root | parent |

Browsers have been lying about their identities via their user agent strings since 1995.

Dalewyn 3 days ago | root | parent |

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.)

oefrha 4 days ago | root | parent | prev |

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.

VagabundoP 3 days ago | root | parent |

Its not a Mac though. I'd much prefer Apple use the User Agent correctly rather than hack around it like that.

The system we used did use user agents to identify mobile devices or computers. And gave the iPad users a terrible experience because Apple lied.

buzzerbetrayed 3 days ago | root | parent | prev |

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.

netdevphoenix 4 days ago | root | parent | prev | next |

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

glutamate 4 days ago | prev | next |

> your Danish language setting in Croydon screams "I VPN to Copenhagen for overpriced pastries."

First time an LLM made me laugh

cbracketdash 4 days ago | prev | next |

"You spent 46 seconds lurking on Hacker News just to end up here with your four Chrome cores and Linux OS in the tech capital of San Jose—did you compile the courage to click? Your 8.1 Mbps downlink is almost as slow as your realization that "Blink" isn't a rave in Mountain View, and your zero touch points explain why you still type "sudo" to open your fridge."

:sob: I'm getting roasted by an AI

pizza 3 days ago | root | parent | prev | next |

I can practically already hear the conversations that will be overheard while waiting in line at Red Rock Coffee later today..

867-5309 4 days ago | prev | next |

> Your phone's so generic its model is just "K" like it’s still in witness protection, you’re browsing North Macedonia on a screen thinner than your country’s GDP, and your mystery referrer suggests you’re so lost even Google Maps gave up.

I can't help but feel refreshingly anonymous

petercooper 4 days ago | prev | next |

Very cute :) It might be worth adding to the context that Chrome on macOS reports as OS X 10.15.7 (Catalina) no matter what future version you're actually running. It seemed very confused to my other specs but running such an old OS.

cpeterso 4 days ago | root | parent | next |

btw, Chrome, Safari, and Firefox all froze their User-Agent strings at “Mac OS X 10.15” after running into web compatibility problems reporting “Mac OS X 11.0”. Some websites’ User-Agent parsers assumed there would never be an 11.0 version of macOS.

Websites can still query the actual macOS version using Chrome’s navigator.userAgentData API. Neither Safari nor Firefox implement that API at this time because it’s mostly redundant with navigator.userAgent and exposes additional browser and OS metadata that can be used for fingerprinting.

chris_st 4 days ago | root | parent | prev |

Did the same with Safari... which it reported as being from the future!

noduerme 4 days ago | root | parent |

Ditto with Firefox

Roasting your Hacker News–refugee, Firefox-on-MacOS-Catalina-10.15-holding, DIY-screen-resolution-having (1512x982? Did you freehand that?) setup like Portland keeps its "quirky" reputation: aggressively outdated but convinced it’s still avant-garde.

6LLvveMx2koXfwn 4 days ago | prev | next |

Oh look, Mr. "I-type-URLs-manually-like-a-medieval-scribe" from Hackney, rocking a 1920x1200 screen just to flex that you need extra vertical pixels for your Linux terminal and crippling Firefox 134 FOMO. Eight CPU cores and zero referrer? Classic NPC behavior - even your internet history is as empty as your commitment to joining the Chrome cult.

location not redacted because it is wrong

zgeor 4 days ago | prev | next |

> Congrats on being the NPC who pre-ordered iOS 18.2.1 "Midlife Crisis Edition" but still uses "Mobile Safari" unironically.

I cried a little, this is too good.

GranPC 4 days ago | prev | next |

Oh this is fun. It seems to be a bit puzzled by my setup, which is understandable since it's a Linux phone cosplaying as Android:

Okay, let's tackle this roast. First, I need to pick the three most unusual bits from the user's data. Looking at the info: The user is on Android 15, which is pretty new and not widely released yet. That's unusual. They're using Firefox on a mobile device with Linux x86_64 platform—wait, Android is Linux-based, but specifying x86_64 might be odd if most mobile devices are ARM. Also, the referrer is Hacker News, which is a good target for a joke. The screen dimensions are 540x1206, typical for a mobile, not too odd. Country is Spain (ES), region AN is Andalusia, city Cadiz. Maybe connect location with something else. So, the three unusual points: Android 15 (unreleased?), Linux x86_64 on a mobile (maybe mismatched), and coming from Hacker News. Need to connect location if using it. Since Cadiz is a coastal city in Spain, maybe link the outdated OS version with something there. Also, making fun of the referrer as a site where people pretend to know tech. Structure: Three sentences. First, Android 15 and pretending to be a dev. Second, Linux on mobile and Cadiz's ancient history. Third, Hacker News referrer and Firefox version. Keep it concise and funny. bird judge

Roasting your Android 15 beta-testing, Linux-on-mobile hipster self: you’re browsing Hacker News (sigh) on Firefox 134 in Cádiz, a city older than your OS’s last kernel panic, yet somehow you’re the one still using a "mobile" device with zero touchpoints—did your stylus die in 2003?

2024user 6 hours ago | prev | next |

chatgpt: "With 4 cores and an OS that’s older than some startups' lifespans, you're basically running a vintage computing museum—do you charge admission, or is the suffering free? And coming from Hacker News, I assume you're just one upvote away from convincing yourself that running outdated hardware is actually a philosophical stance on digital minimalism."

deepseek "rocking Mac OS 10.15 like it’s 2019—did the update scare you, or are you just nostalgic for the pre-pandemic era?"

arthurcolle 4 days ago | prev | next |

Oh look, a Hacker News refugee with a Frankenstein 1800x1169 screen—did you split that browser window just to flex your NPC-level multitasking skills in Rockville's tech equivalent of a Starbucks patio? Twelve CPU cores but still running MacOS 10.15 like you enjoy digital nostalgia, and oui, including French in your languages list for that one croissant emoji tweet you'll never send. Chrome 132? At least the RTT is zero, matching your urgency to upgrade anything.

------------

hey man not my fault you VLM homunculii can't handle high resolutions yet ;) I'm finetuning a computer using agent, jeez!

abixb 4 days ago | prev | next |

My roast summary with referrer field disabled:

"Brave of you to arrive here directly from nowhere like a digital hermit, flexing 16 CPU cores just to run Edge 132.0.0.0 like it’s a beta-testing crisis hotline—enjoying Wisconsin’s -40° wind chills so much you’ve even got your browser set to "arctic Blink engine." At least your 2560x1440 screen lets you see exactly how lonely that empty referrer field looks."

This is the most fun post/link I've interacted with across social media all week, haha. I'm going to test this from all of me and my SO's personal devices of various form factors. Thanks OP!

## Edit:

Here's the roast from my iPhone SE 3 device on 5G cellular (thinks I'm in Atlanta for some reason, lol):

"Roast: Oh, an iOS 18 beta user in Atlanta—how bold of you to test Apple’s unfinished bugs while flexing that three-core “supercomputer” like Georgia’s tech scene peaked with a fax machine. Your screen’s so tiny (375x667) it makes Hacker News look like a MySpace page, which tracks, since you clearly clicked here mid-procrastination scroll to avoid finishing your side hustle. Congrats on being the NPC who still uses “Mobile Safari” unironically—your cookie-enabled loyalty to Apple is almost as tragic as your referral from a site where everyone’s arguing about TypeScript."

Deeper research pointed to Apple's "Displays" kb page for UIKit Size, and it tracks with my device type, as SE 3 is based on iPhone 8 [0]

[0] https://developer.apple.com/library/archive/documentation/De...

cionut 4 days ago | prev | next |

got caught red handed: "Ah, a Berliner using a Supernote Nomad tablet to browse on Chrome WebView 83 like it’s still 2020—did your avant-garde e-ink device fail to render the "update browser" button? Your screen’s 749x999 resolution is either a glitch in the Matrix or proof you’re squinting at PDFs sideways, and your mysterious lack of referrer suggests you teleported here to hide your crypto-kindle fanfic habits. Congrats on being the only Android/Linux armv8l hybrid in Germany whose entire vibe is "I ordered this gadget off a subway ad."

jasonthorsness 4 days ago | root | parent |

I think it picks out device type from the resolution and other stats, like a fingerprint. Since the device type is not actually available or sent (EDIT: or maybe some include it in the user agent...)

gljiva 4 days ago | prev | next |

Hilarious! It never occurs to the model that the user might be using a privacy-oriented browser (I tried it on several devices in several browsers). I like it even though it's drawing totally incorrect conclusions.

etienne89 4 days ago | prev | next |

> Your screen’s so small, even your memes are vertically challenged.

Using an iPhone Mini, this is the most hilarious response I’ve seen from an LLM

CSDude 4 days ago | prev | next |

This is hurtful, Skynet will just mock people to death :(

> Roasting your digital vibe: You’re rocking a 30-bit color depth in Ankara like it’s 1998, clinging to Mac OS 10.15.7 like a Turkish lira in hyperinflation, and lurking on Hacker News for 3 seconds before panic-clicking here—truly a masterclass in chaotic productivity.

username223 3 days ago | prev | next |

To be fair, I shouldn't expect anything great from insult comedy based on a user-agent string, but I have to say I'm not impressed by this word salad of "reasoning":

> Okay, let's tackle this roast based on the user's browser info. First, I need to pick the three most unusual bits. The Safari version 18.1 stands out because that's a future version not released yet. The screen resolution of 1223x772 is an odd, non-standard size. The location is $LOCATION, which isn't too common, and the referrer is empty, meaning they came directly here. Wait, the referrer is empty. That's unusual. So maybe include that they came from nowhere. Also, the Safari version is ahead of current releases. The screen resolution isn't a typical size, so that's another point. $LOCATION could be linked to the screen size, like a quirky local landmark. Need to make sure to relate location to another attribute if used. Also, roast should be three sentences, concise. Let me structure it: start with the Safari version, mention the screen resolution, tie in $LOCATION with the screen size, and the empty referrer. Make it funny and snarky.

And the result is... well, what you'd expect given that reasoning:

> Safari 18.1? Nice try, time traveler—Apple hasn’t even heard of that version yet, but enjoy your 1223x772 screen resolution, the exact dimensions of $LOCATION's least impressive corn maze. You arrived here directly like a cryptid emerging from the woods, which tracks, because only a mythical creature would still run Mac OS 10.15.7.

xivusr 4 days ago | prev | next |

Your beta iOS 18.3 is so unstable even Albuquerque’s heat can’t melt through those Safari crashes, your empty referrer screams “I Googled ‘how to be interesting’ but closed the tab in shame,” and your four-core iPhone’s “hardware concurrency” can’t even multitask between basic roasting and your existential NPC crisis.

> this is great ;-) Next I’d like to see competing models roast each other

bojan 4 days ago | prev | next |

Interestingly enough, it sees my Pixel 9 Pro as model "K" and doesn't really know what to do with it.

T4iga 4 days ago | root | parent |

Google intentionally dumbed down the chrome user agent to limit user fingerprint and encourage usage of their new method that all privacy advances see to despise.

The user agent now reported by most chrome installs on Android reports android 10,K

I have the pleasure of not getting useful info anymore at work when comparing user agents across devices.

colordrops 4 days ago | prev | next |

Dumbass DeepSeek calling my resolution low when it doesn't seem to know about device pixel ratio, what a noob

superfish 4 days ago | root | parent |

I’m surprised this is the only comment mentioning devicepixelratio. Most of the roasts from deepseek seem to involve roasting the window size but that’s misleading without DPR

sundarurfriend 4 days ago | root | parent |

I was wondering why it was showing mine wrong and why many people's resolution here was surprisingly low. TIL about DPR. Too bad that was the only good roast in my first result.

Refreshing the page gives a new one though, and it's done a pretty good job now!

corroclaro 3 days ago | prev | next |

>Oh, look at you, a Stockholm-based tech enthusiast with a screen so tall and narrow it’s practically a skyscraper for ants, browsing Hacker News like it’s your full-time job. You’re rocking a mysterious "K" model Android device with no CPU architecture listed—did you build this thing in your garage? And sure, you speak five languages, but let’s be real, you’re probably just using Google Translate to flex on your 418-pixel-wide display.

Hah. I have an older Xperia phone. They are tall and lovely.

ptidhomme 4 days ago | prev | next |

Your browser arrived here like a ninja—empty referrer, sneaking in from nowhere after 4 seconds of indecision—yet your "K" phone from Chiapas (but actually Querétaro?) has the mystery of a spy using misspelled Linux armv81 and the multitasking power of 8 cores just to run Chrome in French. You’ve got 5 touchpoints but zero chill, rocking a screen taller than your identity clarity and a network so fast (9.5 downlink!) it almost compensates for your location’s existential crisis. Congrats on being the most cryptically lost polyglot Android user this side of a typo!

GrapheneOS with DuckduckGo browser and a VPN...

1317 3 days ago | prev | next |

seems down for me, the POST request (seen in devtools) just gives a 504

    An error occurred with your deployment

    FUNCTION_INVOCATION_TIMEOUT

    lhr1::zcqst-1737978921661-18c020d7118c

jasonthorsness 4 days ago | prev | next |

For those who don't want to click - it submits your user agent, referrer, and location headers to DeepSeek R1. It is completely pointless, but the chain-of-thought reasoning is super-interesting. For example, when I followed it from this site:

Reasoning: Okay, let's tackle this roast. First, I need to pick the three most unusual bits from the user agent info. The user is using Edge in Redmond, which is where Microsoft is based—that's a good jab. The referrer is Hacker News, so I can mock that they're procrastinating. They have 16 hardware concurrency, which is high for a regular user. Also, screen resolution is 2560x1440, but maybe not too unusual. Oh, and maxTouchPoints is 0, meaning no touchscreen. But the top three? Probably Edge in Redmond, Hacker News referrer, and 16 cores. Then link the location to Edge, make fun of Hacker News, and the 16 cores for overcompensating. Keep it concise and three sentences. Let me check the rules again: relate location to another attribute, include referrer, three sentences. Yeah, that works.

Content: Using Edge in Redmond? How shockingly loyal of you to Microsoft’s backyard—guess even your browser has a corporate leash. Came here from Hacker News, huh? Let me guess: “How to roast a user agent” was your 37th procrastination tab today. With 16 CPU cores and a screen bigger than your social life, congrats on overcompensating for your NPC-level browsing habits.

davemtl 4 days ago | prev | next |

Roasting your ultrawide 3440x1440 "I need binoculars to see my tabs" setup in Longueuil, QC—a city so forgettable even your *2-core potato PC* probably buffers its Google Maps—while you lurk on Hacker News like a tech bro NPC still waiting for Y Combinator to notice your crypto-SaaS-pet-rock startup. Chrome 131 on Windows 10? Your browser’s as bland as your cookie-enabled existence, and that 50ms network RTT just proves your Wi-Fi’s as tired as your LinkedIn hustle. At least your 24-bit color depth can render the embarrassment of your life choices in vivid detail.

-- Harsh. But extreamly funny. Interesting that it sees my 5800X3D as a 2-core processor. Turning off DDG privacy essentials yeilds the correct core count, etc.

gigel82 4 days ago | prev | next |

Fun, creative use of the model :)

"Ah, a Hacker News refugee with the screen resolution of a discount monitor fished out of a Kirkland Costco dumpster—16 cores to crunch those "Who's hiring?" threads, but still rocking a touchless Windows 10 setup like it's 2015. Firefox 134? Bold choice, really screams "I read release notes for fun," yet here you are, still stuck in en-US limbo, cookie-enabled and terminally online. Congrats on being the NPC who unironically uses "Web 2.0" in conversation."

WesleyJohnson 4 days ago | prev | next |

> Your Hacker News referrer screams "I argue in the comments unironically," your 12 CPU cores confirm you need all that power to procrastinate on Catalina like it's 2019, and Richmond's only quirk here is hosting a Mac that somehow hasn't died of FOMO from missing software updates.

I'm on an M4 MacMini, running MacOS Sonoma. It's definitely a little confused, but that could be Arc failing to properly report the User Agent? Not sure. Kinda funny, none the less.

Rastonbury 4 days ago | prev | next |

"Your screen resolution’s as confused as your career path after reading one too many "YC applications open" posts."

We need a comedy benchmark now

donatj 4 days ago | prev | next |

> Roasting your Safari 18.1.1 on Catalina (still? Minnesota winters must slow those updates) with a screen so narrow (820px) it’s like you’re browsing Hacker News—your thrilling* referrer—through a vintage 1997 monitor slit. Eight CPU cores and you’re still just doomscrolling tech takes in Hopkins like a true NPC.

I'm on an iPad in Safari. It's UA lies.

droptablemain 3 days ago | prev | next |

"Roasting your Hacker News refugee setup: flexing 20 CPU cores to doomscroll like a Linux-dad cyborg in Sherman Oaks, rocking Chrome 131.0.0.0 like you're personally beta-testing Google's most basic Blink update, yet still stuck at 1080p like a peasant who thinks "4K" is a cryptocurrency."

Glorious.

joseda-hg 4 days ago | prev | next |

It misreads my RAM as being stock, so I guess it's basing it on the model and not the actual system info

Edit, nevermind, the user agent reports it wrong, didn't know that could happen

Also it seems like in a second attempt, it got caught in a loop, of trying to cycle among the same 3 "unusual" things but deciding it wasn't unique enough

ic4l 4 days ago | root | parent |

Most browsers won’t report anything higher than 8, even if your system has more memory.

I think its related to fingerprinting attacks.

chrismorgan 4 days ago | prev | next |

I observe that the reasoning may not correspond to the result: personifying it probably a bit too far, in my first attempt, it forgot to pace itself and put all three pieces of information into the first sentence, and so had to go back and surface novel information for the second and third sentences.

> Oh, a Linux user from Hyderabad with 16 CPU cores—nothing screams "I overcompensate for my fear of GUIs" like a Firefox setup so vanilla even your referrer ghosted you. Your screen resolution is as confused as your identity: Aussie-British-American-English polyglot, yet somehow still the NPC browsing from a cybercafé in 2007. At least your cookie’s enabled, which tracks, since you’d never risk missing a single crumb of open-source newsletter drama.

(In the reasoning, it had mentioned location, browser, and the three things it eventually decided to talk about: referrer, hardware concurrency and OS. The new material was screen resolution¹, languages², and cookies being enabled.)

Most of what it says is kinda weird, honestly. Most of the ones others have been posting feel much better.

I’m also a little bothered that it mixed proper and ASCII punctuation: one EM DASH, two RIGHT SINGLE QUOTATION MARKs, but QUOTATION MARK instead of {LEFT,RIGHT} DOUBLE QUOTATION MARK.

—⁂—

Its second attempt was way better (and didn’t exhibit the same problem of running out of material):

> You lurked here from the void like a digital hermit (empty referrer gang?), flexing Firefox on Linux like it’s 2004 and you’re still arguing about open-source superiority over Hyderabad’s spiciest biryani—meanwhile, your browser’s language list reads like an identity crisis: en-AU, en-GB, en-US… *mate*, pick a colonialism.

—⁂—

¹ 1706×960, which is 2560×1440 @ 1.5×, not too uncommon for a laptop. Wish I could have got something with 2× without losing the power of the 5800HS and the ability to exceed 16GB of RAM. Fortunately, due to life situation changes I’m finally able to switch to desktop computing, and am looking forward to building one instead of being stuck with the frustratingly limited choices available in laptops.

² I set the intl.accept_languages pref to "en-AU,en-GB,en,en-US" years ago, for fun. Actually, I first wrote it as something like "en-AU,en-GB;q=0.9,en;q=0.8,en-US;q=0" for even more fun, a mild roast of my own for perhaps US English or perhaps software developers in general, but somewhere along the way I dropped the quality values for some reason I don’t remember. I’m already fairly easily fingerprintable for other reasons, I’m not fussed about an oddly-specific Accept-Languages header.

piotrmacha 4 days ago | prev | next |

> Max touch points is 10. That's unusual for a Windows device since touchscreens aren't common there. So, mock them for pretending their laptop is an iPad.

Good one. I actually have a touchscreen on my laptop but I almost never use it.

nindalf 4 days ago | prev | next |

> an iPhone running iOS 18.1.1? Bold move showing up before Apple even announces it

Takes the sting out of the rest of the roast knowing that it was dead wrong to begin with.

When the robot revolution is about to begin let’s mess with their training data cut off.

leonewton253 4 days ago | prev | next |

Ah, a time traveler from Calgary using Safari 18.2—bold move flexing a browser version Apple hasn’t even released yet while clinging to Mac OS Catalina like it’s still 2019 and your cold Canadian hands can’t let go. You arrived here directly, no referrer, as if you’re ashamed of whatever crypto-bro forum or “How to Update Your OS” tutorial you crawled out of. Eight CPU cores and still rocking a 1469x885 resolution? Even your screen’s aspect ratio screams “I watch Netflix in a tiny window while secretly Googling ‘why is my Mac slow’.”

simonw 4 days ago | prev | next |

> Braving Hacker News to end up here in San Mateo, huh? Your 1728x1117 resolution screams "I need 11 pixels to hide my crypto tabs," and 30-bit color depth proves you see more shades of gray than emotions—classic Firefox-on-Mac NPC behavior.

Yeah, that's pretty great!

andrewfromx 4 days ago | prev | next |

pretty great. How long before deepseek charges for each api call?

minimaxir 4 days ago | root | parent | next |

DeepSeek does charge for the API (just 20-30x cheaper than o1), I assume OP is eating the cost.

https://api-docs.deepseek.com/quick_start/pricing/

andrewfromx 4 days ago | root | parent |

i see and where is https://groq.com/ in this. They used to be the cheapest?

KTibow 4 days ago | root | parent | next |

Right now Deepseek's official hosting is cheaper than everyone else who can manage to run the model, including Deepinfra. I haven't seen any good hypotheses as to why other than their large batch size and speculative decoding.

rfoo 4 days ago | root | parent |

DeepSeek-V2/V3/R1's model architecture is very different from what Fireworks/Together/... were used to.

That's their "business" model (okay, they don't care about business that much for now, but still) too: you can't run it efficiently without doing months of work we already did, so even with all weights open you can't compete with us.

minimaxir 4 days ago | root | parent | prev |

The "why is DeepSeek so much cheaper than o1" question is currently a mystery, with the best guess being that compute in China is cheaper.

coliveira 4 days ago | root | parent | next |

DeepSeek is use way less computing to achieve the same results. That's the whole difference.

minimaxir 4 days ago | root | parent |

Then a US compute provider should be able to launch a similarly-priced competitor (e.g. to capture the enterprise market concerned about the China associations) using the open-source version and drastically undercut OpenAI.

I suspect that won’t be the case.

rfoo 4 days ago | root | parent |

> Then a US compute provider should be able to launch a similarly-priced competitor

Right, you just need a few months to implement efficient inference for MLA + their strangely looking MoE scheme + ..., easy!

Oh wait, the inference scheme described in their tech report is pretty much an exact fit for H800s. So if you run the recipe on H100s you are wasting the potential of your H100s. Otherwise have fun making variations to the serving architecture.

To be fair, we had chance. If someone decided to replicate the effort to serve their models back in May 2024 when DeepSeek-V2 was out we'd have it now. But nobody had interest as DS-V2 was pretty mediocre. They (and whoever realized the potential) made big bet and it is paying off.

acheong08 4 days ago | root | parent | prev |

The model is much smaller. Compute in China should be more expensive considering all the US restrictions

minimaxir 4 days ago | root | parent | next |

No one knows the size of o1. The only hint was a paper that suggested it was 200B parameters.

Meanwhile, DeepSeek R1 is known to be 671B parameters because it is open-source.

sigmoid10 4 days ago | root | parent |

R1 is a mixture model with only 37B active params. So while it's definitely expensive to train, it's rather light on compute during inference. All you really need is lots of memory.

quantadev 4 days ago | root | parent | prev |

[flagged]

tokioyoyo 4 days ago | root | parent | next |

So, ChatGPT/Gemini's free version is stealing the users from European AI companies, and the goal is to harm them? Or we call that just a competition?

quantadev 4 days ago | root | parent | prev |

That's a false equivalence. There's more nuance than that. Both OpenAI and Google have a genuine private-sector business model, whereas my claim is that DeepSeek is totally a China Gov't entity.

tokioyoyo 4 days ago | root | parent | next |

> is totally a China Gov't entity

Honestly, I don't know. I feel like if any better model came out from literally any AI lab in China, people would say literally exactly the same thing. Some Chinese companies are pouring billions in similar AI stuff as we're doing in the West. DeepSeek also creating problems for those companies as well. I would even go as far as saying, all those "illegal GPUs" are probably in hands of Tencent, Baidu, Huawei and other big boys, rather than DS team. But again, just conjectures here.

quantadev 3 days ago | root | parent |

My hunch is that DeepSeek is lying about creating their AI using just a few outdated chips and a $6mn budget. And when I heard DeepSeek was providing basically free and unlimited AI inference services that also made me think it's gotta be funded by their Gov't (China), because no private sector company can afford that.

jazzyjackson 4 days ago | root | parent | prev | next |

Why is it an interesting distinction whether a product is privately funded vs tax payer funded? To me it sounds like sour grapes.

quantadev 4 days ago | root | parent |

From China's perspective, undercutting American AI companies using Gov't funds is a great strategy for them. Whether you like it or not sort of depends on which side you're on.

aiono 4 days ago | root | parent | prev | next |

Funny that's basically how "growth" companies work in general

quantadev 4 days ago | root | parent |

American "growth" companies aren't 100% funded by the USA Gov't with the only goal being to damage foreign national private sector companies. American companies are private sector only and have a business model that's to benefit a company, not a nation. In China all companies function as part of their National Party.

aiono 3 days ago | root | parent |

> American companies are private sector only and have a business model that's to benefit a company, not a nation.

How can you believe that when copyright laws only applied when regular people pirate but not when big tech uses pirates content to it train their AI? Also it's known that Pentagon funded many tech companies during cold war https://qz.com/the-pentagon-is-spending-billions-on-big-tech... . Is this not government funding?

quantadev 3 days ago | root | parent |

What I meant was that in America the Gov't doesn't get to dictate the actions and policies of private sector companies, whereas in Communist China their Gov't absolutely does control the private sector, and has it's tentacles deep into every important industry.

Regarding letting AIs train on copyrighted material, which is a separate issue, I do think that will and should remain legal. For example if an AI can learn to paint in the style of Picasso, that's a skill, not a theft. Same is true of an AI that can learn to write "in the style of" Hemingway, for example.

gpm 4 days ago | prev | next |

Hmm, disturbingly accurate lol

> Oh, a 4K screen just to scroll Hacker News in <redacted>? Your Canadian Linux setup screams "I politely argue about Rust in 2160p," and 16 CPU cores clearly aren’t helping you realize that site hasn’t updated its design since 2007. Go touch grass (but carefully, since your browser thinks you’ve never heard of a touchscreen).

qup 4 days ago | root | parent |

> coding in a sulfur-scented bath while flexing 12 CPU cores to refresh *Hacker News* like it’s your job? Even your Firefox setup screams "I maxed out my specs just to argue in the comments," yet here you are, roasting yourself via a web app. NPCs wish they were this predictably extra.

No u!

ttul 4 days ago | prev | next |

“Your iPhone screen is so skinny it’s basically a vertical Hot Dog mode survivor, you came here from Hacker News like a Vancouverite cosplaying as a Silicon Valley “disruptor,” and with 5 max touch points, your phone’s ready for a jazz hand performance nobody asked for.”

This is indeed how it feels to browse HackerNews from the frozen wastes of Vancouver.

wanderingmind 4 days ago | prev | next |

> Roast: Your 324px-wide screen must be a flip phone from 2004, you’re browsing Hacker News like it’s a personality trait, and Australia’s most exciting export is apparently your Firefox-on-Android-15 setup with 8 cores just to read tech bro headlines.

Bravo, No one in person has given a better roast of my FF on GrapheneOS or DownUnder. I'm sold

svnt 4 days ago | prev | next |

My first one just said:

Bravo, London

But my second one was funny. But I feel like we are maybe overlooking the significance: is DeepSeek actually funny? Does humor travel with reasoning?

cubefox 4 days ago | prev | next |

For me it just returns a wall of text with (part of) its reasoning about composing a short roast, but not the roast itself. I guess this is not intended.

jasonthorsness 4 days ago | root | parent | next |

This ended up impacting quite a few users; the response is streamed from a Vercel function where I had not changed the default timeout, so at 10 seconds it was just cutting things off where it was. Some invocations of the DeepSeek API take much longer than others. This should be fixed now with a longer timeout.

jasonthorsness 4 days ago | prev | next |

I added Show/Hide Prompt at the end, in case you want to try the same prompt with your user information on other models.

jasonthorsness 4 days ago | root | parent |

Here is mine for example

Generate a fun roast of my user agent/browser information. Pick the three most unusual bits of information and make fun of those. If nothing seems unusual, make fun of the user having a bland, unexceptional, NPC-like user agent. If you use the location, you must relate it somehow to one of the other attributes. Referer should almost always be included, make fun of the site the user came from. Do this all in a three-sentence, concise roast. { "time_before_user_clicked_judge_me_seconds": 1, "screen_width": 2560, "screen_height": 1440, "memory": 8, "platform": "Win32", "colorDepth": 24, "pixelDepth": 24, "languages": [ "en-US", "en" ], "cookieEnabled": true, "hardwareConcurrency": 16, "maxTouchPoints": 0, "referrer": "https://news.ycombinator.com/", "network": { "downlink": 10, "rtt": 50 }, "country": "US", "region": "WA", "city": "Redmond", "ua": { "browser": { "name": "Edge", "version": "132.0.0.0", "major": "132" }, "engine": { "name": "Blink", "version": "132.0.0.0" }, "os": { "name": "Windows", "version": "10" }, "device": {}, "cpu": { "architecture": "amd64" }, "isBot": false } }

cess11 4 days ago | prev | next |

Returns 504 Gateway Timeout. I guess either my rig bested the LLM or the service ran out of money.

strange_magnet 4 days ago | prev | next |

>Not only did you crawl here from Hacker News, the tech bro nursery, but you’re flexing 16 CPU cores just to run Linux like a spreadsheet samurai in Kirishima—impressive, given your browser speaks three languages yet still can’t ask for ramen politely.

..not bad

schoen 4 days ago | prev | next |

"[...] you’re so tech-cliché, your browser’s “maxTouchPoints” is zero because even your devices fear human contact [...]"

minimaxir 4 days ago | prev | next |

> Bravo, time-traveling Safari 18.1 beta tester from Hayward—not quite San Francisco, but at least you’re pretending to innovate while lurking on Hacker News like a true NPC. Your decade-old Mac OS 10.15.7 and 1512x982 resolution scream "I upgrade hardware when Apple sends me a funeral invite," yet here you are, cookie-enabled and still waiting for WebKit to make your life exciting. Congrats on being the most cutting-edge relic in the Bay Area’s shadow.

...I suspect that humor is still out of reach for reasoning models. It may require a bit of system prompt engineering.

mkarliner 4 days ago | prev | next |

"The user is on a Linux armv81 platform but using Chrome on an Android device—that's odd because Android usually runs on Linux-based kernels, but the specific armv81 might be a typo or uncommon."

Unimpressive, it's a current Samsung. And I'm in London not Leistershire

byearthithatius 4 days ago | prev | next |

Roasting your Hacker News-lurking, Mountain View-based setup: Only a Linux user with 16 CPU cores would still need 21 seconds to click "Judge Me" while triple-checking their Firefox configs. Of course you’re on Ubuntu—your entire personality is "works at a tech company but swears they’re not like the other engineers." And maxTouchPoints: 0? Touch grass, not screens, you beautifully predictable NPC.

loves_mangoes 4 days ago | prev | next |

It doesn't recognize Tor Browser as such, but in this case calling its anti-fingerprinting user agent a "desperate lack of personality" happens to be more a compliment than a roast!

On the other hand, it's disappointed that the browser claims to only have two cores and uses a resolution "from 2009".

edm0nd 4 days ago | prev | next |

>Then, their screen resolution is 3440x1440, which is an ultra-wide monitor. Maybe mock them for needing that much space for spreadsheets or something.

lol. i love having so much screen space :3

layoric 4 days ago | prev | next |

My roast: "Roasting your Firefox-on-Linux, Hacker News-referring, Sydney-dwelling NPC setup: Only a 3780x2160 screen could contain the sheer audacity of using 16 CPU threads just to doomscroll tech bro headlines, and I bet Sydney’s entire coffee supply powers that "Gecko" engine you’re pretending makes you elite. Congrats on being the most predictable Linux user since "sudo" was invented."

elaus 4 days ago | prev | next |

> Roasting your Firefox 135 on Linux setup from <redacted>—because nothing says "I have opinions about open-source" like a 16-core CPU crying over hckrnews.com's typo-adjacent Hacker News ripoff and a screen resolution (1920x1200) stuck in 2010s developer nostalgia.

I feel called out - well done.

kalleboo 4 days ago | prev | next |

I spoofed a slightly older User Agent...

Oh, you’re rocking IE 4.01 in Tokyo? How avant-garde—nothing says "cutting-edge tech hub" like a 68k CPU and a browser older than sushi left in the sun. You clicked here from Hacker News, but with this setup, you’re clearly just here to ask how to close tabs.

dbattaglia 4 days ago | prev | next |

> Congrats, you can almost high-five yourself for reading tech bro hot takes on a beta iOS version like it’s a personality. Your setup screams “I refresh Y Combinator in a bathroom stall,” and your Safari browser agrees.

Yup, pretty accurate! Funny it also thinks my 15 Pro is an old iPhone since I’m using lower resolution.

vishnugupta 4 days ago | prev | next |

Roasting your Bengaluru "startup bro" vibe: you’re flexing an unreleased iOS 18.1.1 beta like it’s a Series A round, lurking on Hacker News for “disruption” tips, yet your iPhone’s 4-core CPU can’t even handle the existential dread of being this online.