Vaibhav Gupta

753 posts

Vaibhav Gupta

@vaibcode

Making a programming language - BAML. @boundaryml, 🦄 aitw podcast: https://t.co/g4CJWgOsrj prev YC, google, msft, deshaw, and other things

Seattle, WA Присоединился Eylül 2012

440 Подписки1.6K Подписчики

Закреплённый твит

Vaibhav Gupta@vaibcode·17 Kas

I don’t enjoy the syntax and tooling we use to building AI pipelines. Langchain/ai sdk/et all have never really felt “right”. No type safety, single language support, prompt is hidden until I buy some observability suite. So we made a thing youtu.be/wD3zieaV0Yc?si…

YouTube

English

5.1K

Vaibhav Gupta@vaibcode·14h

@pranavcmadhukar Yep. Equivalent of saying “haha you’re dumb cause you can’t read hieroglyphics” I may still be dumb, but it’s not because of hieroglyphics abilities 😂

English

Pranav@pranavcmadhukar·15h

@vaibcode this paper is a great filter for those who understand what llms do well and can't do. i suppose it was somewhat helpful to illustrate the extent of generalization LLMs are capable of but it was a really high ceiling to test

English

Vaibhav Gupta@vaibcode·16h

This is so dumb. No shit LLMs can write brainf*ck or a language that uses whitespace as syntax. There is virtually no training data that says whitespace is semantic. Why would someone expect this? This is completely different than tranfer learning python -> java. It would’ve been surprising if LM said this, and that would’ve been a very interesting research paper, but you proved nothing here

Lossfunk@lossfunk

🚨 Shocking: Frontier LLMs score 85-95% on standard coding benchmarks. We gave them equivalent problems in languages they couldn't have memorized. They collapsed to 0-11%. Presenting EsoLang-Bench. Accepted to the Logical Reasoning and ICBINB workshops at ICLR 2026 🧵

English

514

Vaibhav Gupta@vaibcode·17h

new ci check for slop incomming →

English

178

Vaibhav Gupta@vaibcode·19h

@alonsosilva Why go on a path when you can pave one 🐑 🐑🐑🐑

English

Alonso Silva@alonsosilva·1d

@vaibcode So where is BAML going? OpenAI or Anthropic? 😉

English

Vaibhav Gupta@vaibcode·1d

seems like the models can't build everything and sometimes you just need damn good engineering. uv and bun are clear standouts here. both are incredibly amazing engineering teams that produced beautiful systems.

Charlie Marsh@charliermarsh

We've entered into an agreement to join OpenAI as part of the Codex team. I'm incredibly proud of the work we've done so far, incredibly grateful to everyone that's supported us, and incredibly excited to keep building tools that make programming feel different.

English

447

Vaibhav Gupta@vaibcode·1d

@GeoffreyHuntley “Make it better” My model likes it when I’m assertive.

English

513

geoff@GeoffreyHuntley·1d

this is my favourite prompt of all time: “how could this be better?” reply with yours and why it rocks!

English

121

14.1K

Vaibhav Gupta@vaibcode·2d

If you want the truth, read the code.

gabby@GabriellaG439

New blog post: "A sufficiently detailed spec is code" I wrote this because I was tired of people claiming that the future of agentic coding is thoughtful specification work. As I show in the post, the reality devolves into slop pseudocode haskellforall.com/2026/03/a-suff…

English

315

Vaibhav Gupta ретвитнул

Anish Palakurthi@anishpalakurT·2d

Looking for designers who have Blender experience! Will pay ~$500 for a single asset

English

Vaibhav Gupta@vaibcode·2d

@boundaryML What is Result and ? in rust.

English

Boundary@boundaryML·2d

Whats your favorite language feature and why is it match?

English

230

Vaibhav Gupta@vaibcode·2d

@GeoffreyHuntley @latentpatterns i'm hyped we should catch up soon!

English

geoff@GeoffreyHuntley·2d

@vaibcode this month - launching @latentpatterns and explorations in unikernel land. 😎

English

Vaibhav Gupta@vaibcode·6d

this man cooks ;) next moth is gonna be fun.

geoff@GeoffreyHuntley

@yenkel @chamath @theallinpod i don’t think software factory will be a thing until we get better languages

English

571

Vaibhav Gupta@vaibcode·2d

@philippberner i think i'm a clean shaven man (who is sometimes too lazy) :')

English

Philipp Berner@philippberner·2d

@vaibcode You need a mustache ;)

English

Vaibhav Gupta@vaibcode·2d

best way to spend every tuesday. yappin

dex@dexhorthy

shared this in a couple places - here's our simple tactic for using claude across many repos in a complex codebase - fun chat w/ @vaibcode

English

458

Vaibhav Gupta@vaibcode·3d

@GeoffreyHuntley oh no geoff...

English

118

geoff@GeoffreyHuntley·3d

tbh. i’m starting to fall in love with ocaml again

English

5.1K

Vaibhav Gupta@vaibcode·3d

@AndersonAndrue @aaronvi we have a bunch of other docs around correctness (typesafe exceptions, match, type-systems etc). this is purely about various syntactical forms when writing code.

English

Storm@AndersonAndrue·3d

@vaibcode @aaronvi I don't see code correctness listed in the concerns?

English

Vaibhav Gupta@vaibcode·3d

syntax makes a huge difference to how good coding agents are, and languages should be rethought. some great learnings from rust and go! great post by @aaronvi

English

1.1K

Vaibhav Gupta@vaibcode·3d

my sister worked on this :) api keys for gemini are way easier to get now (and have been for a bit!) @minimish7

Logan Kilpatrick@OfficialLoganK

@coffee go to ai.studio, we make them automatically for new users, then 3 clicks : )

English

1.5K

Vaibhav Gupta@vaibcode·3d

"The agent writes code, runs it, reads errors, fixes them, and repeats. Every syntax error is a wasted step. Every wasted step costs tokens, latency, and money." damn this post is just full of bangers.

Vaibhav Gupta@vaibcode

syntax makes a huge difference to how good coding agents are, and languages should be rethought. some great learnings from rust and go! great post by @aaronvi

English

382

Vaibhav Gupta@vaibcode·3d

@zeeg the most useful things the best engineers do is delete code and remove complexity. folks on the everything app just dont get this yet. 😂

English

777