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
71 frames of our incredible Karnataka as a tribute to the 71st Kannada Rajyotsava!
71ನೇ ಕನ್ನಡ ರಾಜ್ಯೋತ್ಸವ!
ಒಂದು ರಾಜ್ಯ, ಹಲವು ಜಗತ್ತು!
#KarnatakaRajyotsava
I am disturbed to see Prof. Suryanarayana Doola being vilified over the recent controversy at IIT Bombay. I know Surya personally. He is one of the finest faculty members I have known in the IIT system, and a gentleman of the highest integrity. Whatever concerns have been raised should be examined fairly. A public campaign against an individual serves no one.
With over 25 years of experience at IIT Bombay and IIT Delhi, here is my take on the entire controversy and also a solution, which I believe we must look into.
The problems we are seeing are all primarily because of the one size fits all approach in IITs.
With students coming through an intensive coaching culture, and classrooms bringing together students whose JEE Advanced ranks range from single digits to five digits (in the same class -for example CSE in IIT Delhi), we now have highly heterogeneous classrooms across the IIT system. Language also further adds to the challenge.
The question to ask therefore is, with such heterogeneity, can we expect every student to learn at the same pace?
There is nothing wrong with opening the doors of our best institutions to students from different backgrounds. The question is whether we have addressed this heterogeneity in our teaching-learning methods.
A faculty member cannot pitch every class at several levels at once. Move too slowly, and some students lose interest. Move too quickly, and others are left behind. In my classes, I have also seen students with excellent JEE Adv ranks struggle. Coaching had prepared them to solve familiar types of questions quickly. Many are lost when they see a question which they haven't solved before.
This is not a difficulty confined to any admission category. The move from coaching to independent learning can be hard for anyone.
Here is a solution.
I feel we should consider moving our https://t.co/KVfDlMnXUZ programmes to five years in IITs. Those who meet the credit requirements could complete it in four years, or perhaps three and a half. Others could spread out their courses over five years and receive more academic help, including recorded lectures, online material and support with language where needed. The pace would be decided by performance within the IIT, not by a student’s admission category or entrance rank.
Calling it a five year programme for everyone has a psychological advantage. Students who need more time would simply be following the programme as designed, while those who are ready could complete it sooner. No one would be singled out for taking a fifth year.
We must uphold academic standards. We must also recognise that students do not all arrive at the same point, or learn at the same pace. Giving them the time and support to meet those standards is a responsibility of the system.
@iitbombay #IITs @iitdelhi
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: https://t.co/OGyPb7yaYt
Anthropic’s Economics team is sharing a new model of how AI might affect economic growth, jobs, wages, and more by 2030.
Explore the scenarios, tell us what you think will happen, and see how your answers compare to more than 10,000 Americans. https://t.co/AvQlEZNxR0
No, this is not a protein powder.
Here is a quick heuristic for considering anything to be an excellent protein source - at least 20g of protein per 100 kcal. 15-20g is very good. 10-15 is good. Below 10, it's really not a protein source
If you do the math on this powder, it comes to 4.2g of protein per 100 kcal. It is mostly a carb and fat source. In contrast, whey isolate gets you 22-24g per 100 kcal.
A deep analysis of what may have lead to the Nepal glacial collapse and a breakdown of how the flood overtook communities in a matter if minutes. #Nepal#flood#weather
We created a pitch deck to tell a handful of VC firms about us and what we were up to (a fun experience!). Here’s a few slides about our background and some of the things we’ve worked on from the pitch deck (it was fun putting together the list of people in our teams who have gone on to found a whole range of exciting companies). We are delighted to have selected @radicalvcfund
and @khoslaventures to lead our initial funding round, along with participation from @lightspeedvp, @kleinerperkins, Doerr Capital (@johndoerr), and Alphabet (@Google). We’ll be working with them to close our seed round over the next few weeks.
Announcing Discovery Loop!
I am very excited to announce that, along with my longtime friends and collaborators @Sanjay_Ghemawat, @OriolVinyalsML and @quocleix, we are founding Discovery Loop (@DiscoLoopAI), a Public Benefit Corporation whose mission is to automate machine learning, science, and engineering to accelerate discoveries and progress. The four of us have worked together for 14 to 30 years, and have helped build some of the world’s most used products, infrastructure and AI models, and we’re excited to turn our attention to this ambitious endeavor.
♾
Learn more at: https://t.co/Rv3LMdLluK
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.
AI will transform every industry, power every company, and be built by every country.
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.
The world needs both frontier closed models and frontier open models.
https://t.co/AUKzoQ5Ikb
Not just encroachment removal. We have also started repair of damages to the footpaths to improve walkability. We will spend ₹ 70 CR on repairs to footpaths.
ಕೇವಲ ಒತ್ತುವರಿ ತೆರವುಗೊಳಿಸುವುದಷ್ಟೇ ಅಲ್ಲ. ಪಾದಚಾರಿ ಮಾರ್ಗಗಳಲ್ಲಿರುವ ಹಾನಿಗೊಳಗಾದ ಭಾಗಗಳನ್ನು ದುರಸ್ತಿಪಡಿಸುವ ಕೆಲಸವನ್ನೂ ಆರಂಭಿಸಿದ್ದೇವೆ, ಇದರಿಂದ ಜನರು ಸುರಕ್ಷಿತವಾಗಿ ಮತ್ತು ಸುಗಮವಾಗಿ ನಡೆದುಕೊಳ್ಳಲು ಅನುಕೂಲವಾಗುತ್ತದೆ.
ಪಾದಚಾರಿ ಮಾರ್ಗಗಳ ದುರಸ್ತಿಗಾಗಿ ₹70 ಕೋಟಿ ವೆಚ್ಚ ಮಾಡಲಾಗುತ್ತಿದೆ.
#GBA #NammaBengaluru