Pinned Tweet
I'm Vineet, a software engineer in Delhi NCR. About 10 years in, mostly Python backends, and still at it full-time.
On the side, I build AI-assisted tools for Indian shops and lawyers.
All of it: vin8003.com
1
148
The pressure cooker is the best progress bar ever built.
No percentage, no ETA, just whistles, and everyone in the house knows exactly when it's done.
4
Two kids taught me more about race conditions than any concurrency course.
Same toy, same moment, and no lock.
5
Being confidently wrong on X pays twice: once for the take, once for the apology.
Being quietly right pays nothing.
That's why the timeline is full of predictions and nobody ever posts the follow-up.
when scrolling social media i think it is worth bearing in mind that you can say “odds of reaching SOTA have dropped to zero” and get alot of likes and then 3 months later when they have reached SOTA you post apology form and get even more likes and the subject of your writing prints it out and thanks you for motivation
theres no consequences at all for inaccuracy and every incentive for hyperbole and the x algorithm is tuned for this
5
The real win of AI coding tools isn't writing code faster.
It's that throwing code away got cheap.
When a rewrite costs ten minutes, you stop defending a bad first design in review. Sunk cost was always the bottleneck, not typing speed.
8
The kitchen masala dabba is the best cache I've ever seen.
Seven hot keys, one lid, and nobody touches main storage.
6
Gurugram commutes aren't measured in kilometres. They're measured in variance.
The same short drive can be 20 minutes or over an hour, depending on one flyover, one shower, one waterlogged underpass.
You don't plan for the average day here. You plan for the worst one and call everything else a bonus.
7
Every society WhatsApp group runs on the same cast: one person who reports the problem, everyone else replying "same here", and nobody calling the plumber.
It's the most honest incident channel I've ever seen.
12
Auto-updates feel risky until you compare them to the real alternative: "I'll patch it this weekend."
Every weekend has a reason.
The scary vulnerability is rarely the clever one. It's the old version someone meant to update.
Scary NGINX vulnerability, but I checked and my VPS servers already had auto upgraded to the new one!
Very very very very very important to enable auto upgrades now, I even have auto reboot, which is a bit risky if you're not present, but I care more about my sites being up to date and safe than offline for a few hours
And generally security updates/upgrades are tested and safe
12
"Distribution is everything" is said loudest by people selling distribution.
Distribution gets you the first try. It can't get you the second.
If people don't come back, more reach just means more people finding out faster.
5
Nobody tells juniors this: read the whole error message. Slowly.
Most bugs that look mysterious were already explained in a stack trace someone scrolled past.
Debugging is mostly reading. We just call it debugging so it sounds harder.
10
"Quick call?" is the most expensive message in Indian office chat.
It's never quick. And it usually exists because someone didn't want to put the decision in writing.
Typing leaves a record. That's exactly why people avoid it.
9
Every office has one person who knows how the old system actually works.
They aren't in the docs. They are the docs.
The real migration risk isn't the code. It's that person going on leave.
8
AI can do the niche's task. It doesn't know the niche's paperwork.
The form the regulator changed last quarter. The export only one bank still uses. The field everyone fills in wrong.
That boring knowledge was always the moat. AI just made it the only one.
💯
There is no real way to survive for most SaaS companies here because AI is so completely different than any technological advancement before it
Before you could just niche down and specialize on one thing and you'd know for certain the big tech companies would never be able to dedicate enough resources to fill your tiny little niche market of $300M where you could take 1% of the market and make $3M/year
Riches in the niches!
And so you'd avoid going too general, because "you'd spread yourself too thin" and hit "death by a thousand paper cuts"
And general was where big tech companies were great at, they could dedicate huge resources in solving these big problems, like building Google Sheets, or Salesforce (whatever it is they do)
But that was the old days, and it's completely different now
With AI, big tech companies can easily fill any niche, or be general at the same time, they can do the whole range actually, without even spending time on consciously doing that, because they're not offering a particular SaaS offering at all
They're offering you artificial general intelligence (AGI) that can do it all
And as long as they improve their models every time and make them bigger and better, it slowly can do literally anything (okay not yet anything, for anything they'll need physical robots, but they're working on that too of course)
This leaves SaaS companies with a bitter ending:
- they either do nothing and stay in their niche like they economically should, but that doesn't work anymore now
- they try go general, and offer EVERYTHING and ANYTHING, and that works for a little bit until customers realize they might as well use whoever makes the models that the SaaS itself uses anyway, which is the big tech and big AI companies, who by the way either offer it free or for much cheaper
Everyone's been checkmated by big tech and big AI
This will be the creative destruction of the century
But there'll be some survivors and then a lot of new companies, but how they'll look nobody knows yet!
6
GRAP-3 got lifted across NCR because Delhi's AQI improved. Gurugram's AQI went up anyway.
The rules for the whole region run on one city's number.
The smoke never stops at Sirhaul to check which side of the border it's on.
9
The best CRM in India is a shopkeeper's memory.
He knows your name, your usual, and who pays on credit.
Every "customer data platform" is trying to rebuild that, with more dashboards and less trust.
7
A chess engine can tell you the best move. It can't make you understand it.
Play a move you can't explain and you're lost on the very next turn.
Understanding is the part you can't download.
7
Home automation rule I keep relearning:
If my family can't use it without me, it isn't automation. It's my hobby, installed in their house.
The real test is a guest flipping the light switch.
7
Being rude to AI doesn't make it work harder. It makes me specific.
Polite me: "could you take a look at this?"
Annoyed me: "the total on line 40 is wrong. Fix that. Touch nothing else."
The annoyed one was always the better prompt.
7
Open-sourcing your SaaS only works if the code was never the product.
Anyone can clone a repo on a Sunday. Almost nobody wants to own the upgrades, the backups and the weird bug in month four.
That's the part people actually pay for.
Reminder to
1. Find popular saas people kinda hate paying for
2. Make them free and open source
3. Build an insane amount of goodwill
4. Upsell the power users to a hosted version, team features and support
7



