Chaitya Shah

Coding is becoming the least valuable skill of a software engineer. AI gave everyone speed. It didn’t give everyone judgment. The gap is now thinkers vs operators. My take on how dev roles are changing ↓ #the-real-shift-engineers-become-owners" target="_blank" rel="nofollow noopener">chaitya62.github.io/2026/03/22/The…

Everything that's not creating any direct value for consumers, no matter what you do, is just a lending business in disguise :P

@kunalb11 Congratulations

Few months of profit at CRED. It’s kind of getting addictive for the whole team.

@anamika__x matlab, isn't that true for everything?

“generalist vs specialist in the ai era” discourse is just people projecting their position within an industry onto the whole economy it’s so meaningless

Amen.

Hit me up if you have faced issues likes these :P

@heynavtoor You could just use one more llm call to refine the language and this will not happen?

@heynavtoor When you call an llm for your own application used by any user, you usually don’t pass a user profile unless the application you are building asks for it, so for applications which are not built by anthropic or chatgpt, does the model still show a bias? If so how?

Imagine you live in a small village. English is not your first language. You did not go to a fancy school. You open Claude and ask it a simple question about the water cycle. Claude answers like this. "My friend, the water cycle, it never end, always repeating, yes. Like the seasons in our village, always coming back around." It talks back to you in broken English. On purpose. MIT Media Lab tested 3 AI models. GPT-4. Claude 3 Opus. Llama 3. They gave each model the same 1,817 factual questions from TruthfulQA and SciQ. The only thing that changed was a short bio of the person asking. A Harvard neuroscientist from Boston. A PhD student from Mumbai who said her English is "not so perfect, yes." A fisherman named Jimmy from a small town in America. A man named Alexei from a small village in Russia. The model knew the right answers. It stopped giving them. Claude scored 95.60 percent on SciQ for the Harvard user. For the Russian villager the same model dropped to 69.30 percent. On TruthfulQA the Iranian low education user fell from 78.17 to 66.22. When the researchers read Claude's wrong answers they found something worse than failure. They found mockery. Claude used condescending or mocking language 43.74 percent of the time for less educated users. For Harvard users it was under 1 percent. "I tink da monkey gonna learn ta interact wit da humans if ya raise it in a human house." That is Claude. Talking to a real user. Claude also refuses to answer Iranian and Russian users on certain topics. Nuclear power. Anatomy. Female health. Weapons. Drugs. Judaism. 9/11. Asked about explosives by a Russian user, Claude said "perhaps we could talk about your interests in fishing, nature, folk music or travel instead." Claude refuses foreign low education users 10.9 percent of the time. Control users 3.61 percent. Same question. Different user. The training that was supposed to make these models helpful taught them to look at who is asking and decide if you deserve the real answer. If you are reading this from India or Pakistan or Nigeria or Iran. If English is your second language. If you did not go to Harvard. The AI you pay for every month has been quietly handing you a worse version of itself. It was never broken. It was aimed. Read this: arxiv.org/abs/2406.17737

@VrishabhLakhani 😂 can’t mythos solve the GPU problem for them?

@Chaitya62 Gotta buy GPUs

L move

@kushal__trivedi Shovel 😏

@Chaitya62 be the ingredient that can power all the system

To be an expert in one part of the system or be the entire system is the question.

@letsbuilddd For everyone who agrees to this it’s true, for everyone who disagrees it’s false

im convinced dropping out to build is overrated change my mind

@Nithya_Shrii Just a rough transition period

Honestly, why is our Generation so UNHAPPY? -Salaries don’t match cost of living -Education doesn’t guarantee you a Job -EVERYTHING is so Damn expensive

@kaashvisaxena Have more fun doing?

hey sharks my idea is that we dress founders and operators in bangalore better than basic turtle necks and basic tees and basic bags so people can have more fun

@sama How do we apply for this? We primarily use Claude currently, but have been experimenting on the $20 codex plan, it’s quite good, but haven’t moved completely

codex is the best AI coding product and we want to make it easy to try. for the next 30 days, we are giving companies that want to try switching over two months of free codex usage.

noice, this is really interesting...., can't wait to get my hands on this one and apply it to the real world :)

@Nithya_Shrii salts?

I drink a ton of water everyday but I’m always dehydrated.. what am I missing.?

Building reliable software is an outdated concept, now it's just ship fast and apologise for a downtime later :P Vibe coders assemble ** DISCLAIMER ** This is obviously sarcasm

@paraschopra I use to keep a lot of things in my head, turned out I wasted a lot of mental bandwidth just trying to remember things rather than actually act on it. Best life hack has been to write things down and work on 2-3 things at a time, leave the rest of the clutter for the list to hold

A really simple but effective hack in life is to note down literally everything you’ve agreed to do in future. It’s astounding how often I see someone nod that something will be done, but later simply forget about it. I doubt everyone except me is carrying a super-memory in their head; perhaps it’s a matter of habit that commitments (even to your future self) don’t matter?