I'm in love with this sentence:
"if you don't connect with bigger things (sunlight, history, art, nature, love), you will make the ill mistake of believing you are the big thing."
Indians abroad used to largely blend in and remain inconspicuous. That began to change with Modi, who fostered a culture of unapologetic Indian pride, encouraging Indians to believe that India is the greatest country on earth and that Indians are the greatest thing since sliced bread.
In Modi's defence, it was good that he encouraged Indians to shed their insecurities and inferiority complex. But the flip side is that it has given way to arrogance and an obnoxious chest thumping nationalism that doesn't travel particularly well. There's a difference between being proud of where you come from and expecting everyone else to share your enthusiasm.
This work is simple, clear, compelling, but embellished. I first found it shocking, the notion that a model can natively understand neural network weights and productively edit them! That's really not what's happening, & many here are annoyed. BUT, it's still notable & worth understanding.
I've done a bit of analysis & auditing, here are some thoughts:
1/ This paper does *not* show an agent can edit its weights ala RSI. It does *not* show a model can edit in new capabilities. It DOES show that a frontier model can remove "bugs" (reasoning loops), using the same ablation technique researchers have used to remove safety refusals (Exp 8 - Qwen). So yes, renewed concern about frontier LLMs getting access to their own weights & editing safety/CBRNE/alignment features.
2/ This paper also does *not* show that an agent can "natively" read, deeply understand, and fix model weights. What is happening is that the model can cleverly deduce the same things that mech-inturp researchers do and fix them in mostly ham-fisted ways. Still cool, but not as groundbreaking. Evidence in thread below...
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: https://t.co/bOWCcFhCVq
“Yeah, there are these companies called neoclouds. They buy GPUs from NVIDIA and then rent it back out to the hyperscalers. Where do they get money for the GPUs? Excellent question. Turns out they get the money from NVIDIA themselves, along with a bunch of private credit lenders whose loans are backed by the GPU itself. Elon just said all of them might go to zero soon. Anyways, what are you doing later tonight?”