Submit your Application to Partner with the IndiaAI Safety Institute! 4 Days Left! Are you working at the forefront of AI safety and governance? Partner with the IndiaAI Safety Institute and contribute to national efforts to build a safer AI future. Selected institutions
RESEARCH
-
Blurring images with code helps LLMs recognize faces like squinting
By
–
I think it’s literally blurring the image with code tools, and it recognizes the face after — the LLM equivalent of a human squinting or viewing it at a distance
-
LLMs Solve Cold Start Problem; Real Value in Continuous User Learning
By
–
All that LLMs do is solve the cold start problem. The real action is when AIs learn continuously from their users.
-

ChatGPT o3 may only see image after blurring via tool use
By
–
Possible, but a few details suggest otherwise: 1. ChatGPT o3, also with search, doesn’t seem to be able to do it (0/3 attempts) 2. It doesn’t work if you just ask “What is this?” 3. The reasoning trace summary suggests it only sees it after blurring (presumably via tool use)
-
JAX Usage as Indicator of High-Quality Research Papers
By
–
You can tell a paper is going to be solid when you follow the link to the repo and see it uses JAX
-
2025 AI Renaissance: Are You Prepared for the Future?
By
–
2025 Is Poised To Usher In The AI Renaissance—Are You Prepared?
#AI #AIio #AIInnovation #ML #DataScience #Futureofwork @timnitgebru @oriolvinyalsml @ceobillionaire @soumithchintala @waitin4agi_ @sallyeaves @bernardmarr -
AGI Alpha: Humanity’s Next Giant Leap Forward
By
–
[ We Choose To Ascend With AGI ] Like the Apollo Moon landing, "AGI Alpha" is our next giant leap – the greatest opportunity of our time. Together, we choose not merely to witness the dawn of AGI, but to ascend with it. Web : http://
agialphaagent.com #AGIALPHA -
Your Brain On AI: ‘Atrophied And Unprepared’
By
–
Your Brain On AI: ‘Atrophied And Unprepared’
#AI #AIio #AIInnovation #ML #DataScience #Futureofwork @demishassabis @Ronald_vanLoon @TamaraMcCleary @geoffreyhinton @goodfellow_ian @jeffdean @erikbryn -
HLE Questions Structure and Public Accessibility Challenges
By
–
Es difícil pues las preguntas de HLE no son públicas, salvo unos pocos ejemplos, y no tienen una estructura similar entre ellas.
-

Grok 4 Shows Exceptional Performance on Humanity’s Last Exam
By
–
Si se confirman estos resultados de Grok 4 -y según cómo se hayan calculado- estaríamos ante un modelo muy potente. Un 45% en Humanity's Last Exam es una salvajada. Más del doble de los modelos actuales (o3/Gemini)
