Rishabh Profile
Rishabh

@rishabhj_says

Followers
391
Following
7K
Media
55
Statuses
3K

Co-founder @Localeai, Forbes 30U30

Joined March 2012
Don't wanna be here? Send us removal request.
@rishabhj_says
Rishabh
2 months
RT @arnav_kumar: To the young dreamer out there -. Being underestimated is exclusively the privilege of the young and ambitious. Wear it li….
0
2
0
@rishabhj_says
Rishabh
5 months
Is this guy crazy or what? We don’t have the ability to digest better AI models. There’s already a lot to absorb.
@sama
Sam Altman
5 months
change of plans: we are going to release o3 and o4-mini after all, probably in a couple of weeks, and then do GPT-5 in a few months. there are a bunch of reasons for this, but the most exciting one is that we are going to be able to make GPT-5 much better than we originally.
1
0
1
@grok
Grok
7 days
Join millions who have switched to Grok.
191
311
2K
@rishabhj_says
Rishabh
5 months
Is Claude down? . I don't know how to code anymore.
1
0
4
@rishabhj_says
Rishabh
5 months
has anyone built checkpointing in @cursor_ai? something minimal like git commits after each prompt based change that you can easily roll back to. Love how cleanly bolt dev handles this.
1
0
1
@rishabhj_says
Rishabh
6 months
Is passing a string to gpt to convert it into json vibe coding?.
0
0
0
@rishabhj_says
Rishabh
6 months
One of the coolest things about vibe coding is how I can now throw away features very quickly. I don't feel guilty about just deleting an entire module because I'm coding at least 10x faster.
0
0
0
@rishabhj_says
Rishabh
6 months
SaaS is not dead. Yes people are replacing $100k saas applications with in house AI tools. But that’s old saas. New age apps will do the work that large teams are doing today and it is going to grow faster than anything we have seen.
0
0
1
@rishabhj_says
Rishabh
6 months
The way we judge "good execution" is about to change completely.
0
0
0
@rishabhj_says
Rishabh
6 months
Second-order effect: MVP standards will skyrocket. My 2-week MVP now has more features than what I built in 2 months at my startup 6 years ago.
1
0
1
@rishabhj_says
Rishabh
6 months
Not sure if this is "vibe coding". But I'm super excited about how product building will look in another year.
1
0
0
@rishabhj_says
Rishabh
6 months
Has this messed up my data model? Absolutely. Will I fix it later? That's the plan.
1
0
0
@rishabhj_says
Rishabh
6 months
Database/infra = straight terminal work. Sometimes I copy snippets from Claude when thinking feels like too much work.
1
0
0
@rishabhj_says
Rishabh
6 months
Tests? Tried. Too laborious. They break often. Will bite me later but moving too fast to care now.
1
0
0
@rishabhj_says
Rishabh
6 months
For complex data scripts: function by function approach in cursor. AI either sucks at complex logic or I suck at prompting. Constant testing is key.
1
0
0
@rishabhj_says
Rishabh
6 months
3/ Backend = cursor. I describe an endpoint, request/response format, and it creates most of the code. Just works.
Tweet media one
1
0
0
@rishabhj_says
Rishabh
6 months
Frontend = bolt. I describe what I want at a high level. Then I go component by component fixing designs and adding API integrations. Not perfect but fast.
Tweet media one
1
0
0
@rishabhj_says
Rishabh
6 months
Built a product in 2 weeks that would have taken 6 weeks before. AI tools have completely changed how I code. Here's my current workflow:.
4
0
9
@rishabhj_says
Rishabh
6 months
Most people now understand why evals matter for AI. But who's actually writing them at companies? Will this fall to engineering teams, or are we seeing the emergence of a specialized QA function dedicated to creating these evals at scale?.
0
0
1
@rishabhj_says
Rishabh
6 months
The trough of disillusionment.
@svembu
Sridhar Vembu
6 months
One fact about the recent AI hype is that corporate customers and analysts are not terribly excited. They are in a "prove it to me" mode, as they should be. I am in the camp that overhyping anything is a bad idea. I am personally enthusiastic about some technologies but I will.
0
0
1
@rishabhj_says
Rishabh
6 months
Most LLMs are surprisingly bad at converting docs into working code. It's like they understand English and code separately, but struggle to map between them. Has anyone figured out a good workflow around this?.
1
0
2