Pranav Wadhwa (@pranavwadhwa24) 's Twitter Profile
Pranav Wadhwa

@pranavwadhwa24

NYC | VT β€˜23 | tech/sports/misc

ID: 940395391887036418

linkhttps://pranavwadhwa.com calendar_today12-12-2017 01:38:27

150 Tweet

145 Followers

188 Following

Ali Partovi (@apartovi) 's Twitter Profile Photo

Sway: a plan to reimagine democracy I’m starting the most ambitious project of my life: a tech platform to reimagine democracy. I know you've probably got opinions on this. Please give feedback and maybe join us. (🧡 A thread...)

Leroy Kerry (@oelry) 's Twitter Profile Photo

We decided to help tax pros actually enjoy tax season. Not survive it. Not dread it. Enjoy it A crazy idea? Maybe Until it worked πŸ‘‡

Wispr Flow (@wisprflow) 's Twitter Profile Photo

It's official β€” Wispr Flow is now live on the iPhone App Store! We built the first immersive voice keyboard that lets you dictate with incredible accuracy anywhere β€” 5x faster than typing. Delightful. Effortless. Intelligent. Our mission is to change how people interact with

Bubble (@bubble) 's Twitter Profile Photo

It’s time! We’re opening native mobile up to public beta. Check us out on Product Hunt, grab a 1-month Starter promo, and drop feedback β€” we’ll be in the comments all day! bble.io/ph

It’s time! We’re opening native mobile up to public beta. Check us out on Product Hunt, grab a 1-month Starter promo, and drop feedback β€” we’ll be in the comments all day! bble.io/ph
Girish (@googrish) 's Twitter Profile Photo

1/ Can codebase-specific RL push the frontier for code LLMs? At CGFT, we helped a client RL-tune Qwen-2.5-7B on their internal codebase for unit test creation, with coverage-guided GRPO. The result? It beats o4-mini & o3. Here’s how it works (link to full blog in bio) 🧡

1/ Can codebase-specific RL push the frontier for code LLMs?

At <a href="/cgftlabs/">CGFT</a>, we helped a client RL-tune Qwen-2.5-7B on their internal codebase for unit test creation, with coverage-guided GRPO. The result? It beats o4-mini &amp; o3. Here’s how it works (link to full blog in bio) 🧡
Ali Partovi (@apartovi) 's Twitter Profile Photo

Today former presidential candidate Andrew YangπŸ§’β¬†οΈπŸ‡ΊπŸ‡Έ, pioneer of the Forward Party, started mobilizing an army of voters. If you believe in ranked-choice voting, open primaries, & more, vote with Andrew here: sway.co/andrewyang. (Powered by @VoteWithSway)

Pranav Wadhwa (@pranavwadhwa24) 's Twitter Profile Photo

I joined Cassidy back in May and the energy has been absolutely electric. Excited to continue building with this talented team

John Kim (@john_sungjin) 's Twitter Profile Photo

Today, we're launching Village to let investors orchestrate teams of agents to scale their judgment. Three years ago, we bet that both humans and agents would need new tools to fulfill the promise of LLMs to transform research. Village is that tool, the first IDE for research

Houjun Liu (@houjun_liu) 's Twitter Profile Photo

Introducing π˜π—΅π—Όπ˜‚π—΄π—΅π˜π—―π˜‚π—―π—―π—Ήπ—²π˜€: a *fully unsupervised* LM for input-adaptive parallel latent reasoning βœ… Learn yourself a reasoning model with normal pretraining βœ… Better perplexity compared to fixed thinking tokens No fancy loss, no chain of thought labels πŸš€

Introducing π˜π—΅π—Όπ˜‚π—΄π—΅π˜π—―π˜‚π—―π—―π—Ήπ—²π˜€: a *fully unsupervised* LM for input-adaptive parallel latent reasoning

βœ… Learn yourself a reasoning model with normal pretraining
βœ… Better perplexity compared to fixed thinking tokens

No fancy loss, no chain of thought labels πŸš€
Michelle Lim (@michlimlim) 's Twitter Profile Photo

Your website should build itself. It should adapt to your visitors and competition in real time. Introducing Flint: autonomous websites. We’re already powering pages for @Cognition and @Modal. Today we announce our $5M seed from Accel and break out of stealth. More below.

AC&E (@appliedcompute) 's Twitter Profile Photo

Generalists are useful, but it’s not enough to be smart. Advances come from specialists, whether human or machine. To have an edge, agents need specific expertise, within specific companies, built on models trained on specific data. We call this Specific Intelligence. It's

Generalists are useful, but it’s not enough to be smart.

Advances come from specialists, whether human or machine.

To have an edge, agents need specific expertise, within specific companies, built on models trained on specific data.

We call this Specific Intelligence.

It's
rajan agarwal (@_rajanagarwal) 's Twitter Profile Photo

recently, i spent some time working on cross-lingual alignment for LLMs via encoder injection! treating languages as modalities is a compute-efficient way to extend understanding of low-resource languages without extending pre-training or the tokenizer

recently, i spent some time working on cross-lingual alignment for LLMs via encoder injection! 

treating languages as modalities is a compute-efficient way to extend understanding of low-resource languages without extending pre-training or the tokenizer
Leonie Freisinger (@lcfreisi) 's Twitter Profile Photo

My cofounder Anna-Julia Storch and I grew up in manufacturing towns while spending our days chasing speed - on the mountains and the racetrack. With Dryft , we bring our love for both worlds to life. That’s why we couldn’t be more excited to finally come out of stealth and

Jamie Turner (@jamwt) 's Twitter Profile Photo

We just raised a bunch of money to reinvent backend engineering for developers. We have huge plans for Convex, and we need your help to realize them. news.convex.dev/convex-raises-…

Tanay Kothari (@tankots) 's Twitter Profile Photo

we just raised another $25M after 10x'ing our ARR in 5 months. the crazy part is this almost never happened. 17 years ago, I watched Iron Man as a 10-year-old kid in Delhi. that night, I pulled my first all-nighter teaching myself to code. not because I wanted to build apps or