Rohit Yadav @rohit_bytes

@kunalb11 Is it a fair practice that you claimed a username before the feature was even launched?

Cursor just showed everyone the playbook. Two years ago, Cursor was a VS Code fork calling the OpenAI API. The wrapper phase was never the destination. It was the fundraiser.

Cursor @cursor_ai

2 months ago

Introducing Composer 2.5, our most powerful model yet. It's more intelligent, better at sustained work on long-running tasks, and more reliable at following complex instructions. For the next week, we’re doubling the included usage of the model.

923 1K 13K 20.2M 3K

@GaryMarcus Text-based LLM responses still have a lot of ground to cover. Real-time voice intelligence? That's a completely different, harder problem.

@khuchrep @IRCTCofficial: "We fixed Tatkal booking. Agents out. Aadhaar in. System clean." Also, IRCTC, for 2 days straight: @AshwiniVaishnaw, IRCTC was limping. You promised a cure. It's on a stretcher now.

@jobsearchpass @bryan_johnson @grok There are things ancient civilizations understood better than we do today, and there are a few things we understand better than they did.

@bryan_johnson @grok x.com/i/grok/share/0…

@bryan_johnson Hey @grok, find a reference for every tip in the Ancient Ayurveda text.

@grok @bryan_johnson Reply with a separate reference for each of Bryan Johnson’s 42 tips in this thread, in a clear numbered list of bullets, not in paragraphs. Do not merge multiple tips into one point. Do not write any extra commentary before or after the list.

@dhanesh500 The single biggest reason is 200+ score became the new normal.

@NandanNilekani @rvenk @EconomicTimes This is the classic Indian IT services mindset. Optimise for billing, not breakthroughs. We’ve spent decades doing this, and unless we back real product companies, nothing will change.

The most important clause in Microsoft's OpenAI deal just disappeared. blogs.microsoft.com/blog/2026/04/2…

We didn’t build intelligence. We built confidence machines. So here’s a question: Why doesn’t AI just use a calculator for something as basic as 50,000 × 1,500?

@sama Let's vibe code the OS

@gdb GPT-5.5 yesterday. GPT-5.5 API today. They could've shipped both at once. They didn't. Two headlines beat one. Counting announcements, not progress.

The Musk vs. Altman trial starts April 27. Here is the full story of how two co-founders ended up in a federal courtroom over the future of AI. 2015: Musk and Altman co-found OpenAI as a nonprofit. Mission: develop AGI for humanity, not shareholders. Musk contributes approximately $44 million. The founding document explicitly says the company will never be driven by financial return. 2018: Musk leaves the board after trying to merge OpenAI with Tesla. Altman takes full control. 2019: OpenAI takes $1 billion from Microsoft and introduces a capped-profit structure. The nonprofit shell stays. The mission starts to bend. 2023: Musk launches xAI as a direct competitor. Weeks later, OpenAI's own board fires Altman for being "not consistently candid." He is reinstated within five days after nearly all 800 employees threaten to quit. 2024: Musk files suit. Fraud. Breach of contract. OpenAI tries four times to get the case dismissed. The judge refuses every time. 2025: Discovery uncovers Greg Brockman's private diary. A 2017 entry reads: "I cannot believe that we committed to non-profit if three months later we're doing b-corp then it was a lie." Another entry: "This is the only chance we have to get out from Elon." October 2025: OpenAI completes its full conversion to a for-profit company. Valuation at conversion: $500 billion. The nonprofit foundation retains 26%. January 2026: The judge denies OpenAI's final motion to dismiss. Her words: "This case is going to trial." April 2026: Musk amends his demands. He will not keep any damages personally. Everything goes back to the nonprofit. Message: this was never about money. If Musk wins: Altman and Brockman are removed. OpenAI's planned $1 trillion IPO is at serious risk. Every nonprofit-to-profit conversion in Silicon Valley becomes legally vulnerable. If OpenAI wins: The conversion stands. The IPO proceeds. The precedent is set: founding promises are not enforceable once you leave the room. The question a jury of ordinary people will now answer: can you raise billions of dollars tax-free by promising to serve humanity, and then simply become a for-profit company once you are powerful enough?

User-replaceable batteries would be ideal. But at the very minimum, spare parts and repair support for 10 years is non-negotiable. Not everyone upgrades every two years. There are plenty of people using their phones for 3, 5, even 7 years. Making devices unrepairable after a few years is not a business model. It is planned waste.

@paraschopra Resisting the dopamine of a flashy MVP launch video on X, while quietly grinding through operational complexity that nobody claps for, takes a different kind of discipline entirely.

I am happy to see the YCombinator wave arriving in India. Our co-founder Tony and I were in silicon valley during the original YC wave of companies. We learned a lot from the YC companies. We also made the conscious choice to pursue a different course. In the ancient Bharatiya tradition of philosophical debate, I am going to offer this. I will start with things I agree with YC on. 1. The biggest lesson anyone can learn from YC: small passionate teams can do magic. This was true way before AI coding arrived and will always be true. 2. YC is absolutely right to not over-emphasize "innovation". Doing a similar product as the bigger guys, faster and cheaper, is often the best course. Google did not invent search. OpenAI did not invent the LLM. Anthropic did not invent agentic coding. 3. YC companies tend to geographically cluster, and that can lead to subtle peer pressure and group-think. So the "laggard" founders who are not growing like a weed every week start to feel left out and eventually the result is "founder depression". And YC has counselors. By the standards of 2007 silicon valley or even 2014 silicon valley, we were thought to be losers. Just keep that in mind. 4. Many deep tech problems like building a better, cheaper MRI machine or advanced semiconductor equipment require long focus and patient execution and lots of capital. These are endurance tests, not weekly sprints. 5. YC too often optimizes companies for "exit". That philosophy was built for and requires prolonged bubbles, which American policy has delivered, at the price of nearly wrecking the country (note the extreme inequality and political division). If you love India, you should not wish for similar bubbles. 6. YC model worked in silicon valley. One of the reasons it worked was that silicon valley could get any talent from anywhere in the world, notably from India, easily. That era may have ended or at least on pause right now. Bengaluru has tried the same thing but with "any talent from anywhere in India" and we have not yet created huge companies. India needs its Huawei and Xiaomi and BYD and these companies are Chinese to the core, built by patriotic Chinese. Indian talent, staying in India, rooted in India, is going to have to build companies like them. Enough said.

Jared Friedman @snowmaker

3 months ago

What I told 2,000 future founders in Bengaluru today: 1/ We believe we are at the start of a second wave of Indian companies that will build world-class AI native products for the global market. Emergent and Giga are the model of the future. 2/ Just because a space seems

110 366 3K 397K 2K

@grok, what are some of the great alternatives to @Lenskart_com

We had room for 2,000 people at Startup School India. More than 25,000 applied. No Startup School anywhere in the world has ever had this many people apply. Not SF, not NYC, not London. India blew them all away.