Singapore’s Foreign Minister, Dr Balakrishnan casually explaining how he built his own AI agent (a 2nd brain for diplomacy) using Claude & WhatsApp integration etc. on a Raspberry Pi
“You cannot govern a technology you have only been briefed on.” 🇸🇬
Your body has one way to survive real heat: sweating. When the air is too humid, sweat can't evaporate, so your body stops cooling down. In those conditions, a healthy 25-year-old in shade with water can die in about 6 hours. India is now entering the part of the year where this actually happens.
The thermometer lies. A dry 45°C day is brutal but survivable if you find shade and keep drinking. A humid 32°C day can kill older people, sick people, or anyone working outside, because humid air can't hold more water. Your sweat just sits on your skin doing nothing, and your core temperature climbs until your organs start to shut down.
Scientists used to think the human limit was a mix of heat and humidity that felt like 35°C of pure humid air. New research has lowered that line. A 2022 Penn State study put young healthy adults in climate chambers and found they fail about 10°C below the old limit in dry sun. Sydney and Arizona State researchers published follow-up work in 2023 showing the line sits even lower for older adults, especially in dry sun. It moves with age, humidity, sun exposure, and activity level.
India's weather agency just issued its highest red alerts across Rajasthan, UP, Madhya Pradesh, and Maharashtra. Forecasts show 45 to 46°C this week, with some districts approaching 50°C (about 122°F) within two weeks. Chennai has already crossed the line where continuous outdoor work becomes unsafe. Coastal cities hit that line every summer.
India lost 181 billion working hours to heat in 2023, worth around $141 billion in lost pay, mostly for farmers. Consulting firm McKinsey estimates this could reach 4.5% of India's entire economy by 2030. Official death tolls look small compared to reality. The Health Ministry logged 360 heat deaths in 2024 and independent analysts counted over 700. Research using Lancet Countdown data estimates the real number is close to 150,000 excess deaths per summer, counting everyone whose heart, kidneys, or lungs gave out because of the heat.
When India glows red on a heat map, the color marks where the human body is being pushed past its physical limits. That zone expands every year.
dylan field shared on lenny's podcast how he got figma's first users:
he wrote a script to scrape twitter, built a network graph of the design community in Gephi, ran an analysis to find the most influential nodes, then reached out to every one of them
"can I buy you a coffee?"
this is basically what smart outbound looks like in 2026, just with better tools:
1) use data to find who's actively talking about your problem space
2) rank them by engagement and network reach
3) reach out referencing something specific they said
the playbook from 2012 is still the best GTM strategy :)
the only difference is now you don't have to write your own scraper or limit yourself to one platform, just use the Crustdata API
One of the highest ROI activities you can do in your life is to deeply internalize that building good habits is a short term investment that compounds to lifelong gains.
Any new good habit requires overcoming initial friction, but techniques like habit stacking and starting small help.
The trick is to realise that after a while, habit becomes effortless. So it’s just that initial dip you have to overcome. After that, all what you’re trying to do becomes automatic (that’s why it’s called a habit).
So if you’ve been sitting on reading, programming, exercising, dieting or anything else, know that mastering the meta-skill of habit building will probably change your life forever.
NEW research from NVIDIA.
Post-training agents with RL is powerful but expensive.
Every parameter update needs full multi-turn rollouts with environment interactions, making end-to-end RL prohibitively costly for long-horizon agentic tasks.
This research offers a practical middle ground.
The work introduces PivotRL, a framework that operates on existing SFT trajectories to combine the computational efficiency of SFT with the out-of-domain retention of end-to-end RL.
Instead of exhaustive full-trajectory rollouts, PivotRL identifies pivots, informative intermediate turns where sampled actions show mixed outcomes, and trains only on those high-signal moments.
Standard SFT degrades OOD performance by -9.83 points on average. PivotRL stays near zero (+0.21) while achieving +14.11 average in-domain gains over the base model versus +9.94 for SFT.
On SWE-Bench, PivotRL reaches competitive accuracy with E2E RL using 4x fewer rollout turns and 5.5x less wall-clock time.
The method is already deployed in production as the workhorse for NVIDIA's Nemotron-3-Super-120B agentic post-training.
Paper: https://t.co/sIBLUpyfMD
Learn to build effective AI agents in our academy: https://t.co/1e8RZKs4uX
Today I'm releasing my entire newsletter archive (350+ posts) and all podcast transcripts (300+ episodes) as AI-friendly Markdown files. Plus an MCP server and GitHub repo.
A few months ago I shared my podcast transcripts on a whim, and y'all built the most amazing things—an RPG game, a parenting wisdom site, infographics, a Twitter bot, and 50+ other projects.
Let's see what happens when I give you even more data.
Grab the data here: https://t.co/xEPCcPiZHO.
Paid subscribers get all of the data (some 350 posts and 300 transcripts). Free subscribers get a subset.
I don’t think anyone’s ever done anything like this before, and I’m excited to give you this excuse to play with that AI tool you've been meaning to try.
Here’s my challenge to you: build something, and let me know about it. I’ll pick my favorite and give you a free 1-year subscription to the newsletter. Just post a link to your project in the comments here: https://t.co/yJDmuCQfOx.
If you’ve already built something, slurp in this new data and submit it, too. I’ll pick a winner on April 15th.
Check out today's newsletter post for inspiration on what you could to build: https://t.co/yJDmuCQfOx
LFG.
In this episode of the @LightconePod, we sat down with @mukundjha and @madhavjha, the founders of @emergentlabs - an AI platform that lets anyone build and ship production-ready software.
In just eight months, users have created more than 7 million apps on Emergent, with the number doubling in just the last 45 days. We discuss how they built one of the most powerful AI coding agents, why they focused on non-technical users and what it's like building for a global audience from India.
00:00 - Intro
01:06 - What Is Emergent?
01:18 - Founder Backstory
02:09 - From AI Testing to General Coding Agents
02:52 - Getting Ahead of the Market
04:18 - The Pivot to Non-Technical Users
05:22 - Why Second Movers Can Win in AI
09:04 - Building for Production, Not Just Prototypes
18:21 - Live Demo: Building Apps with Emergent
24:40 - How Emergent Hires and Runs a Lean Team
29:04 - Is SaaS Dead? The Rise of Personalized Software
34:04 - The Future: Niche Apps, Solo Builders and AI Agency
@aravind Weak take from someone who's mostly pro-india. The jurisdiction under which the ship was sunk was still in the axis of influence. I wonder what would happen if a chinese or russian ship sailed close to the US. This is an outcome of toothless diplomacy whether you like it or not.
@tata_neu@jagograhakjago tagging for further help on this, they have let me complete a CIBIL inquiry on their platform and have stopped processing further while they claim to process it. I think they're just collecting users' personal information and misusing it. Please take action.
Worst customer support by @tata_neu. I raised a complaint with them on Feb 5th, and they haven't solved it as of today, i.e., Feb 28th. All they do is send AI-generated emails as a follow-up, beating around the bush instead of solving my problem.
Absolute disgrace.
@tata_neu Pure BS, it's been a week now and you haven't resolved my problem yet. There is absolutely no timeline to solve customer problems & I have wasted a CIBIL inquiry on your worthless platform. Will make sure this is escalated.