IIT Bombay Director resigned but this convo will shook the education system 🚨
Girl student - Sir, you don't even know how much pressure is on a student
Shireesh Kedare - Okay, should we make cheating legal for you guys? So that you don't take pressure? We are asked to design papers for students who have the brains to study. For students who came into IIT through 80 percentile. Just because some of you students came through 30 percentile, we can't make the paper that much easy for other students.
Girl student - You are still being shameless sir. This was all your fault
Shireesh Kedare - I am resigning because there was some of my fault. But each one of you children who has ever done cheating should leave this IIT now. I will also ask the watchman to resign just because he couldn't check someone's pant for a mobile phone. All the parents around the world should also resign from their parental duties if they can't raise a studious child. That student feared them, not us. Your parents should be ashamed to send you to IITs when you are not even capable of passing a single exam of IIT.
Sir literally cooked this girl student. He resigned but he made sure to show the world how bad this education system is. Respect +++
Jev AI is a game changer for all the Agentic AI Applications out there. Sitting in front of LLM calls, it can cut down the pricing and latency by over 4x
A large crowd of iPhone 18 Pro & iPhone 18 Pro Max enthusiasts gathered outside DLF Mall, with people lining up since early morning to buy the new phones despite the MASSIVE surge in prices.
And yes, some will still criticize the government over the 0.4% MDR on UPI payments that merchants will have to pay.
"For China & India, U better improve Ur acts. Buy Ur oil & gas from somewhere else"-Senator Blumenthal after bill allowing 100% tariffs for buying Russian energy.
Dear PM .@narendramodi Ji, pls show the menu card Bharat holds for Counter Tariffs to these arrogant Americans.
This Menu Card is based on Bharat's biggest strength—Bharat's largest & fastest-growing market.
See the Menu which will give loose motions to Trump if fed by Modi Ji.
👉US Tech (Massive Leverage):
Target- US companies like Apple, Google, Microsoft & Amazon, which rely on Bharatiya talent, data & markets.
Current policies include Digital Personal Data Protection (DPDP) Act 2023, which mandates data localization for certain categories & proposed smartphone security rules requiring source code reviews by government labs. There's also a digital services tax (equalization levy) at 2% on non-resident e-commerce operators & H-1B visa dependencies that encourage offshoring to Bharat
To hurt US tech firms:
-Mandate full data localization for all personal and sensitive data
-Amend DPDP Rules 2025 to prohibit cross-border transfers without explicit government approval, forcing US cloud providers (e.g., AWS, Azure) to build expensive local data centers or lose market access. This would increase operational costs by 20-30% & disrupt global data flows.
-Impose higher digital taxes and source code mandates: Raise equalization levy to 5-10% & enforce source code sharing for all apps & OS in Bharatiya market, as hinted in 2026 proposals. This would expose proprietary tech, deterring innovation & forcing US firms to either comply or exit.
-Restrict outsourcing & talent export: Introduce 10-15% tax on remittances from Bharatiya subsidiaries of US firms or cap H-1B-like internal transfers, reducing reliance on Bharatiya engineers for AI training & development.
👉US FMCGUS:
FMCG giants like Procter & Gamble, Coca-Cola & Unilever dominate Bharat's consumer goods market through imports & local manufacturing.
Current policies include GST at 18% for many categories, import duties of 5-20% & FDI allowing 100% automatic route for manufacturing.
To hurt US FMCG firms:
-Increase import duties & require local sourcing: Raise duties to 30-50% on non-essential FMCG imports (e.g., packaged foods, personal care) & mandate 70-80% local content in products sold in Bharat, up from voluntary guidelines. This would force costly supply chain shifts or price hikes, eroding market share.
-Rationalize GST slabs unfavorably: Shift premium US-branded products to higher 28% GST slab while lowering it for local alternatives, making US goods less competitive in price-sensitive segments.
-Enforce stricter environmental & packaging regulations: Amend packaging rules to ban single-use plastics for US-style products, requiring biodegradable alternatives, which would hit import-dependent lines hard.
👉US Food Franchises:
US chains like McDonald's, KFC, Starbucks & Domino's operate via franchise models with local partners. Regulations include FSSAI standards for labeling & safety, with new 2026 rules requiring scientific backing for claims & annual labeling updates starting July 1.
To hurt US food franchises:
-Mandate higher local sourcing & menu localization: Require 80-90% ingredients to be sourced locally (up from 30-50% voluntary) & enforce "Bharatiya-origin" labeling for 50% of menu items, disrupting global supply chains.
-Impose franchise fees & taxes: Introduce 5-10% royalty tax on foreign franchisors & limit foreign ownership in franchises to 49%, forcing joint ventures with Bharatiya firms.
-Stricter health & advertising regulations: Ban marketing of high-sugar/fat items (common in US menus) to children under new FSSAI guidelines & require calorie disclosures, reducing appeal in urban markets.
Continued👇
India mein koi 5 minutes number Google kar le to doodh ka doodh paani ka paani ho jaye.
RBI surplus to Govt : ₹2.87 Lakh Crores (US$30 Bn)
Total listed bank profits : ₹4.11 Lakh Crores (US$42 Bn)
NPCI - which runs UPI pre tax ‘surplus’ : ₹ 1,888 crores (US$200 Mn)
To nuksaan kis ka ho raha hai UPI se aur kaunsi subsidy de rahi hai Govt UPI pe jo chubh rahi hai?
Cost of running ATMs and cash logistics in India is ₹30,500 crores (US$ 3.0 Bn). If you want to optimise shut ATMs and promote UPI instead.
Any levy on UPI is just tax collection. UPI is the one scientific achievement of India everyone acknowledges par ab tax ki bali chadhegi.
You can research the market.
You can analyze competitors.
You can build the product.
You can reason about positioning.
But eventually you have to say:
"Here's what I made. Give me your money."
That's psychologically different from everything before it.
If you work in software and want a way into robotics, start with a joint.
My actuator primer explains how motors, gears and electronics turn an AI command into movement, and what that costs.
@Medium : “We have great writers.”
Also @Medium : “Now pay to read them.”
Can’t we find a better revenue model than locking articles behind a paywall?
At this rate, @X articles might just win by being more accessible.
As LLMs are getting better and moving towards AGI, it still not there to take over the end to end complete economical tasks.
Can it start & run a business on it's own?
Why killing instantly is jail, but killing slowly is just penalty and temporary closure?
Food Safety- packaged and outside food is at all time low. @fssaiindia@PMOIndia
When I use Claude Code or Codex, I’m never fully relaxed.
I can let go of the constant scrutiny of the code and logic, but I can’t stop watching the agent.
There’s always this feeling that, if I look away for too long, it might go completely off course.
LLM Inference Engineering - Problem and Solution
Problem: LLMs are slow
Solution: KV Cache - Avoid recomputing previous tokens.
Problem: KV Cache consumes huge memory
Solution: PagedAttention - Manage KV memory efficiently.
Problem: GPU is underutilized
Solution: Continuous Batching - Dynamically add/remove requests.
Problem: Token generation is sequential
Solution: Speculative Decoding - Generate and verify multiple tokens together.
Problem: Attention requires large memory usage and expensive memory movement
Solution: FlashAttention - Reduce memory usage and memory movement during attention.
Problem: KV Cache is still large
Solution: MQA / GQA - Share K/V across multiple query heads.
Problem: Many optimizations need to work together
Solution: vLLM - Build an efficient inference engine.
Problem: Applications repeatedly process the same prefixes
Solution: RadixAttention / Prefix Caching - Reuse previously computed work.
Problem: Models are too large
Solution: Quantization - Reduce model memory and computation.
Problem: Large models are expensive
Solution: Distillation / SLMs / MoE - Make inference itself cheaper.
Keep Learning, Keep Sharing, and Keep Growing.
If lifes gives you an IIT seat utilise it like her.
- She joined IIT Madras in 2022
- While studying she attended a Congress rally, where she heared a lot of fake facts from Pappu to fool people
So she made a website, and listed all the fake talks of Pappu
https://t.co/JbwKt4bxMl