The most prestigious law school admissions discussion board in the world.
|
 |
 
|
chatGPT "Atlas" browser actually seems kinda nice if it has adblock
| woke dog | 10/21/25 |
| screenman | 10/21/25 |
| woke dog | 10/21/25 |
| screenman | 10/21/25 |
| woke dog | 10/21/25 |
| screenman | 10/21/25 |
| woke dog | 10/21/25 |
| screenman | 10/21/25 |
Poast new message in this thread
Date: October 21st, 2025 6:00 PM Author: woke dog having an LLM have constant real-time access to your current webpage/workspace and browsing history would actually be pretty useful
obviously unusable if adblock doesn't work with it, though. i think they'll allow it at first for a while, at least |
Date: October 21st, 2025 6:03 PM Author: screenman using it now
already tracking behavior - when I close a tab it just automatically opens a new one *typing xoxohth.com + Enter* then loops through poasts, blank bumping poasts based on my previous blank bumps.
|
 |
Date: October 21st, 2025 6:06 PM Author: woke dog i'm gonna have it self-automate a daily routine of switching my moniker every morning to one of my 900+ saved monikers (http://www.autoadmit.com/thread.php?thread_id=5788434&forum_id=2/#49364805) |
|
Date: October 21st, 2025 6:19 PM Author: screenman 1800000 (http://www.autoadmit.com/thread.php?thread_id=5788434&forum_id=2/#49364819) |
|
Date: October 21st, 2025 7:30 PM Author: woke dog the "agent" seems entirely useless though. nobody is going to use that, lol
"Uh, yeah, uh, chatgpt can you order me some sysco slop, thank" |
|
Date: October 21st, 2025 7:37 PM Author: screenman it's terrible at browsing - goy popups everywhere so it takes forever to actually do anything.
agent to agent chat will enable instant purchase of slop - ai developers using high frequency trading techniques to get food orders sent to the kitchen within 10ms
"I'm hungry, order my favorite meal and grab a matcha latte at Starbucks with triple sugar added too"
11ms ping "your 4500 calorie meal is on the way, Stacy" |
|
Date: October 21st, 2025 7:43 PM Author: woke dog yeah, the next step is definitely gonna be like that (AI communicating and transacting directly with AI), but we're not close to that yet imo. LLMs are waaaaay too vulnerable to prompt injections/manipulation in their current state
i do think people are gonna love this shit though once they get it rolling and smoothly functioning. people just love the narcissistic circle-jerking, man. "I Love my AI assistant. It just Gets Me, you know?" |
|
Date: October 21st, 2025 7:46 PM Author: screenman (http://www.autoadmit.com/thread.php?thread_id=5788434&forum_id=2/#49364932) |