@IRCTCofficial@AshwiniVaishnaw
Today i tried booking a tatkal ticket for my parents, Even tried with Premium tatkal by paying double. Went to last step, and then boom website crashed. And this complete flow i retried 4-5 times. Get good company to build website.
An AI Summit that sidelines its own builders?
• 7 AM queues
• 9 AM entry
• 12 PM full evacuation
• Hours of sanitization
• PM visit at 5 PM
. Day 1 Ends here
Meanwhile — exhibitors, delegates, startup founders left outside. No water. No clarity.
Media shows celebration. Ground reality was chaos.
If access was limited to select high-value guests, just say it upfront.
Don’t mobilize the ecosystem and then displace them.
This is not how we build India’s AI future. @abhish18@GoI_MeitY@PMOIndia #AiSummit #aiImpactSummit
I was taking an interview,
suddenly the candidate dropped off the call,
we were in the middle of solving a coding question.
I could anticipate that the internet broke on his end,
and after around 5 minutes, he joined again.
He was all in panic mode. 😢
"Sorry, sorry. I'm extremely sorry", he said.
He started explaning,
"I just lost my internet connection due to a power cut."
I could hear him breathing heavily,
I assume, he must have ran fast to fix the issue.
Because of this disturbance of 5-10 minutes,
he just lost where he was in the coding problem.
He almost started thinking everything once again from scratch, because of this small unforeseen issue he just lost a lot of precious time in the interview.
So what am I trying to explain is,
0. Whenever you're giving an interview, just keep a backup internet option available at your end.
So, keep your hotspot running on your phone. ✅
1. I've seen a similar situation happening to a candidate when he had to run and get his charger from another room.
So, keep your charger plugged in. ✅
2. Also, keep a notebook and pen next to you, you never know. ✅
Lastly
Please don't panic,
If anything unforeseen happens, just keep calm.
Remember that on the other side of the call is your friend.
It's okay if something messes up,
interviewers understand that,
It can happen at their end too. 😌
So,
Just try to keep calm,
don't over-explain,
take a deep breath,
focus on where you were,
and continue with the interview like nothing happened. ✌️
And remember,
`Every interviewer was once a candidate.` 🤷♂️
@Hiteshdotcom From way back I've been telling my friends that you're not earning anything from this courses, you could make way more in consulting and other businesses you already have. Also from that 0.8$ you'll be paying office rents, TDS, CA expense to manage this expenses. Thanks man
@aahiknsv And I guess that's why parents says. Beta paisa kamao lekin itna pagal mat hojao uske peeche ki health kharab kro. Health hai to sab hai.
Bounce back 🧿
Optimized API res time from ^1200ms to under 100ms.
12x performance upgrade.
Issue: Just don't send the data you don't need. Let UI request data they need.
Learning: monitor APIs from time to time, Initially it might not be taking this long, but over time it grew on data
@aidenybai There's always a proper way, make a type of interface. But if you are in a hurry and your manager says ship tonight. Maybe this is the way to go. 😂 Sometimes for some functions you don't have to create an interface and a whole separate file and all for it
DeepSeek makes obvious something we've been saying for the past years at StackAI...
1. LLMs are commodities, much as crude oil is...
2. The value of LLMs accrues significantly at the application & orchestration layer, where the refining into useful automations is happening (<- yes, that's where StackAI sits).
3. OpenAI/Anthropic/etc have NO moat; transformers were never a secret, and there is only so much data out there you can mine.
And this is what markets are pricing right now, the lack of defensability from foundational model providers.
But there is one thing that truly creates a Moat... a deep understanding of how Enterprises operate and the different use cases that bring outsized value and results (we have some of these documented on our website for awareness of the outsized results, the rest, I'll keep them private until they become well-known ;).
That's why Accenture is booking $2B in revenue... Ok, you may argue, that OpenAI’s revenue today is about $3.7B; but have you considered their losses? Some reports suggest up to $5B this year... and you guessed right, these margins are not getting better anytime soon, and we all know it. If something, since there is not much room for optimization yet (unless we find a revolutionary new training/inference architecture and chips to support it).
The truth is there is a HUGE gold mine at the application layer, and specifically, the stakes are really high for enterprise-level applications.
@Samuelklam Hey Samuel,
Page doesn't exist and ashby link is also not working. Really interested to showcase my skills and to work with Permitflow as a Full Stack Web developer.
Optimised a data heavy table on UI which was taking around 7-8 secs to load 10yrs of data (excel like table). Brought it down to around 200-400ms. All this by just using hash map instead of array for some data storage.
@RosinolToni FE is not just simply building great UI, it's handling data smoothly without any jitters. Thinking of UX, user should be able to find the CTA they're looking for instantly, there eyes should be dragged there. Things like this makes frontend engineers. Else it's a UI developer