OpenAI is huge in India. Its models are steeped in caste bias. by Nilesh Christopher

Singha’s experience is far from unique. An MIT Technology Review investigation finds that caste bias is rampant in OpenAI’s products, including ChatGPT. Though CEO Sam Altman boasted during the launch of GPT-5 in August that India was its second-largest market, we found that both this new model, which now powers ChatGPT, and Sora, OpenAI’s text-to-video generator, exhibit caste bias. This risks entrenching discriminatory views in ways that are currently going unaddressed.

India needs its own foundation model. We can’t just be data providers.

Inside China’s New Wave of Conceptually Innovative Bookstores by By Jean-Yves Mollier and Patricia Sorel
October 2, 2025

Once inside, customers are further surprised by a vast mirror covering the floor, which doubles the visual effect of this store created for the third millennium. Architect Li Xiang, founder of the X+Living design firm in Shanghai, visitors the impression of entering a rocket, whose speed is represented by the spiral staircase that takes them on a world-spanning trip. As at Hangzhou, Li Xiang divided the Shenzhen store into four distinct sections: a “concept” zone, a “forum,” a more conventional area aimed at children, and finally a conference room for events.

These spaces can be more than stores/libraries. They can be communal spaces.

We need more such spaces!

Apple sidelines lighter Vision Pro to prioritize smart glasses by Jay Peters

With its glasses, Apple will be late to the game. Meta just announced a bunch of new glasses of its own, including a second-generation version of the Ray-Ban Meta glasses with much better battery life that are available now, a new pair of Oakley-branded glasses designed for athletes launching this month, and the Meta Ray-Ban Display glasses, which my colleague Victoria Song called the best smart glasses she’s ever tried.

I read the same piece of news on three publications - MacRumours, slashdot and the verge. They are all source from the same Mark Gurman report.

I was not sure if I wanted to talk about it. But I guess I do.

This is not a good look on Apple, or the way it is being reported. Apple seems to be behind Meta on these things, till it comes out with a product (Vision Pro) and then Meta bashes it, but eventually creates its own take on the ideas (use hand gestures to control the UI).

The reason it is not a good look because this reads like - this company did something great, so let’s abandon our efforts and do the same thing. That is not a good place to be in. You seem like a laggard.

Meta Plans To Sell Targeted Ads Based On Data In Your AI Chats - Slashdot by

Meta will begin using data from AI chatbot conversations and other AI-powered products to fuel targeted advertising across Facebook and Instagram, with no way to opt out.

Well of course it will. As will others ChatGPT is next I guess.

What Researchers Suspect May Be Fueling Cancer Among Millennials - Slashdot by

A 150,000-person study presented at the American Association for Cancer Research meeting found millennials appear to be aging biologically faster than previous generations based on blood biomarkers.

This is a product of the lifestyles we have now - stress, ultra-processed food, etc.

Studies have linked early-onset cancers to medications taken during pregnancy, ultra-processed foods that now account for more than half of daily calorie intake in the United States, circadian rhythm disruption from artificial light and shift work, and chemical exposures.

Nothing’s new AI OS isn’t really an OS, or new by Robert Hart

Eventually — and here’s where the AI-native OS idea comes in — Pei says phones could be more proactive, changing apps’ placement, or even suggesting apps based on how we use the phone. But even then, this is not an OS. It is an interface. Pei admits as much, leaning on semantics to skirt the issue: “I guess the word or the noun ‘OS’ could be interpreted in different ways.”

It feels like Apple Shortcuts on steroids.

But the idea is exciting. The one thing this AI cycle has made possible is normal users building little tools to automatically do the arduous things they were doing manually.

Most of that has been on the desktop or the web.

This is exciting because it takes that same thing to mobile.

Designing agentic loops by Simon Willison

One way to think about coding agents is that they are brute force tools for finding solutions to coding problems. If you can reduce your problem to a clear goal and a set of tools that can iterate towards that goal a coding agent can often brute force its way to an effective solution.

This is an interesting idea. I need to try it out.

Claude Sonnet 4.5 is probably the “best coding model in the world” (at least for now) by Simon Willison

My initial impressions were that it felt like a better model for code than GPT-5-Codex, which has been my preferred coding model since it launched a few weeks ago. This space moves so fast - Gemini 3 is rumored to land soon so who knows how long Sonnet 4.5 will continue to hold the "best coding model" crown.

I wanted to try out chatGPT before committing to Claude for longer term. I did do that now. It is good at a few things, like, the web agent feature, which I’ve used extensively over the past couple of weeks.

OpenAI also released the codex model around that time. For coding tasks it has been better in my experience than Claude, hardly making any mistakes. I had to prompt it to change something a couple of times out of almost fifty or so code changes.

The 4.5 release is supposed to be better.

ChatGPT was worse at creating draft CVs though. I was looking to go back to Claude for this reason alone. I think I will now.