@arpit_bhayani I noticed, most of the time, when i tried to read any new news related to AI, it’ll be a lot of New Tech Jargon, which is hard to understand
Is is possible like an ELI5 or simplified version for each article, so that, it will be easier to understand 😀
“Most of us have lost that natural learning. We have a reputation to protect. A status. An image of being smart or perfect. Asking a naive question threatens that core image, so we stop asking. Trying something new is exciting. Why kill that excitement with the fear of looking bad?” 💯
An important piece from Demis. We need more of this kind of thinking. A good reminder that the goal is a frontier ecosystem that promotes innovation and choice, while avoiding any one model drop that breaks the world!
ELI5 version of Satya Nadella’s post:
Imagine your company has secret superpowers — special recipes, customer tricks, ways of working, or smart decisions that make you better than everyone else. These secrets are your real treasure.
Now you get a super-smart robot helper (AI like Copilot or ChatGPT) that you pay to use.
To make the robot actually useful for your company, you have to teach it your secrets:
• You show it your documents and data
• You write special instructions (prompts)
• You correct it when it gets things wrong
• You tell it what “good” looks like for your team
Every time you do that, the robot learns your secrets.
The company that owns the robot (Microsoft, OpenAI, Google, etc.) also learns them.
So you’re paying twice:
1. With money
2. With your company’s special knowledge
Over time the big robot company gets smarter from your work, and might even help your competitors with what it learned from you. You created new smartness… but it doesn’t fully stay yours.
Nadella calls this the “Reverse Information Paradox.”
Old problem (from a famous economist):
To sell something secret, you might have to show it first → risk giving it away.
New AI problem:
To use the smart robot, you have to show your secrets → risk giving them away.
What he says companies should do instead:
Keep your own private robot brain inside your company (a “trust boundary”).
Everything the robot learns from your work — your corrections, your data, your special ways — should stay inside your walls and only get better for you.
He says every company needs four things:
• Control — Decide for yourself what “good” means (your own private tests). Keep all your data, corrections, and memory inside your company.
• Capability — Build or fine-tune your own AI versions on your own data so it learns your secrets without leaking them.
• Choice — Don’t get stuck with one robot. Be able to switch or mix different AIs easily.
• Cost + Compound — Do it efficiently so your AI gets smarter and smarter over time, and the value stays with your company.
Bottom line in kid words:
Use the magic robot helper…
but don’t let it take your secret superpowers home with it.
Build your own magic helper that only lives in your house, only learns your rules, and gets better and better just for you.
That way your company’s special knowledge keeps growing for you, instead of slowly leaking out to everyone else.
I’ve been overpaying at Lenskart / PVR for years. 😭
Turns out, if you pay directly, you’re leaving ~10 - 22% on the table.
The trick:
CRED → Rewards → Gift Cards → Buy Lenskart gift card → use it at checkout
That’s it. That’s the whole hack.
Super smart move by @arindam___paul & @atomberg_tech.
This is how you build trust and actually help people. Excellent execution 👏
https://t.co/wVeKwEX8Wx
Been exactly a year since we launched water purifiers
We now have TDS data from tens of thousands of Atomberg water purifiers installed across the country
And we combined that with publicly available data from different government sources like central pollution control board, jal jeevan mission and central ground water board
And built a free tool where you can enter your pincode and check your water quality- TDS levels, TDS variance, chemicals present etc
If you are early in your career and want to get a broad overview of various systems fundamentals you should read everything we post on the @PlanetScale blog and follow @BenjDicken.