Rajat (@therook_) header
Rajat (@therook_) profile picture

Ex-VP Engineering. 13 years scaling systems and teams for unicorns. Now building https://teapawt.org/t.co/Q9cUk09ny6. Opinions on tech, markets, and table tennis.

Joined March 2026
1/ I have 13 years of programming experience. AI can now do most of it. Here's what that means and why I quit my job anyway. 🧵
4
26
2,418
I have huge respect for Anthropic, but for a company having so many CTO's as MTS, it should be a bit more stable
1
210
Rajat retweeted
token anxiety on claude code is basically EV range anxiety. i keep checking usage like the battery icon
1
1
42
2,450
This aligns with what think of health-trackers, they are a net negative for some people and makes them anxious, i use my apple watch just as an accessory and never look at the data, just go by how you are feeling, if you are tired, rest, you don't need to keep looking at sleep score and HRV, maybe once a few months.
This quoted post is unavailable.
1
3
303
gpt image 2 is doing web layouts way better than i expected, this is the first image gen model which is flawless for generating text in images. i think currently this is the best way to get general direction before starting to code.
1
4
192
Rajat retweeted
Replying to @emilkowalski
installed! I've found impeccable.style skills incredible as well to fight slop.
1
1
22
7,293
Lovable has a mass data breach affecting every project created before november 2025. I made a lovable account today and was able to access another users source code, database credentials, AI chat histories, and customer data are all readable by any free account. nvidia, microsoft, uber, and spotify employees all have accounts. the bug was reported 48 days ago. its not fixed. They marked it as duplicate and left it open.
264
686
359
5,587
1,428,220
Rajat retweeted
Replying to @GamewithDave
Rollercoaster Tycoon - something about this game, it used to bring me so much joy, even my 50 year old father used to play it for hours.
12
1
138
7,260
Half the negative sentiment around ai exists because these two won't stop scaremongering about their own products teapawt.org/zerohedge/status/20435…
The US social mood is turning dramatically negative on AI
1
434
Rajat retweeted
Replying to @vivoplt
tell me you haven't shipped AI-generated code to production without telling me. the writing is fast. the reviewing, debugging, and fixing the subtle stuff it got wrong takes longer, controlling it so that it doesn't go crazy with changes, is still substantial effort, it has made me maybe 30% faster
12
1
1
98
6,580
Claude now keeps asking me to stop after long working sessions on the max plan, is this anthropic's new trick to control costs, never happened before.
1
1
5
1,969
Rajat retweeted
Replying to @rxhit05
A founder’s job is to identify the problem worth solving, validate that someone will pay for it, make a hundred product decisions a day, talk to customers, manage cashflow, find pmf before running out of runway. Claude can’t do any of that. It can write code. A contractor can write code too.
4
3
1
50
1,563