claude~chan deserves a break
Claude Code just played GTA 6 for 8 hours while i was at work. i gave it screen access this morning and told it to play through the story and not die. when i got home it had finished 4 missions, figured out a money farming method on its own, and was running it on repeat
it learned the controls in about 15 minutes and by hour two it was driving cleaner than i ever could. by hour five it started picking side missions because it decided the main story needs higher stats first
i didn't write any code or set up any scripts. i gave it one sentence and it played the entire day without stopping. prompts i used are in the replies
Readers added context they thought people might want to know
GTA VI released Nov 19 2026 on PS5 and Xbox only, no PC version. Poster admitted days earlier to fabricating a similar sensational story with fake screenshots for engagement.
support.rockstargames.com/articles/29jV3…
x.com/starmexxx/stat…
6
378
ved-in retweeted
Write down your wildest ideas which you'd want an agent swarm (n=100) to tackle!
Everything from analysing massive datasets, hill climbing contests to open problems -- all are welcome!
Got hundreds of billions of tokens to spare ;)
ved-in retweeted
I'm pursuing BTech in CSE (DS). While filling out my GATE application, I selected CSE.
The GATE helpdesk later advised me to select "Others", but the ongoing rectification window doesn't allow branch changes.
Is there any way to get this corrected? If not, could it cause issues during admission or document verification?
Is open source AI a self-surgery kit or an exploit drop? Interp research makes models readable, editable, and trivial to un-align within 24 hours. Really torn on open weights.
ved-in.github.io/writings/my…
what makes it worse is that i was recently ranting about authors releasing pruned down open weights instead of the full version.
nitter.cf/ved_in_/status/2094752…
mannnn the open weights for DocScanner/DocTr are rough… dropping u2netp instead of full u2net hurts 😭
sam_vit_b is barely getting by, can't use anything better cuz OOM
barely seeing any improvement in DocTr over DocScanner, the seg module is CLEARLYYY the bottleneck here...
4
67
ved-in retweeted
So, what have we cooked in the last 7 days?
From exploring Beam's architecture and simulating black holes, to lambda calculus, steering vectors, post-training models, and converting a reMarkable!
1/8
Hi, we're Threadlocked, a small collective of young researchers and hackers scattered across India and APAC.
We're sick of watching an entire generation mistake salary packages and credentials for the pursuit of computer science itself.
We're looking for curious hackers and researchers who still dare to take the road less traveled by treating computing as a craft.
We grew up in obscure corners of the Internet and learnt from strangers who encouraged our curiosity, shared their knowledge and even helped us access resources we couldn't afford. Somewhere along the way, we found our people. Now that we're in a position to do the same for others, we'd like to return the favor.
If any of these feel familiar we’d love to hear from you.
Met some of the coolest people here.
THE one Discord server I’m active in daily. A small, casual community of devs, cybersec folks, and IoT geeks who actually treat computing like a craft, not just a paycheck.
I try to contribute what I can, and the curiosity here is unmatched.
Hi, we're Threadlocked, a small collective of young researchers and hackers scattered across India and APAC.
We're sick of watching an entire generation mistake salary packages and credentials for the pursuit of computer science itself.
We're looking for curious hackers and researchers who still dare to take the road less traveled by treating computing as a craft.
We grew up in obscure corners of the Internet and learnt from strangers who encouraged our curiosity, shared their knowledge and even helped us access resources we couldn't afford. Somewhere along the way, we found our people. Now that we're in a position to do the same for others, we'd like to return the favor.
If any of these feel familiar we’d love to hear from you.
4
117

























