anshuman

10.3K posts

anshuman banner
anshuman

anshuman

@athleticKoder

maximizing shareholder value

Katılım Şubat 2020
966 Takip Edilen20.4K Takipçiler
anshuman
anshuman@athleticKoder·
@0xSero sglang and vllm have tradeoffs
Deutsch
0
0
1
294
0xSero
0xSero@0xSero·
SGLang > VLLM > Exllamav3 > Llama.cpp IF you have any Nvidia hardware this is the play.
English
64
39
673
43.9K
anshuman retweetledi
Qwen
Qwen@Alibaba_Qwen·
🚀 Introducing FlashQLA: high-performance linear attention kernels built on TileLang. ⚡ 2–3× forward speedup. 2× backward speedup. 💻 Purpose-built for agentic AI on your personal devices. 💡Key insights: 1. Gate-driven automatic intra-card CP. 2. Hardware-friendly algebraic reformulation. 3. TileLang fused warp-specialized kernels. FlashQLA boosts SM utilization via automatic intra-device CP. The gains are especially pronounced for TP setups, small models, and long-context workloads. Instead of fusing the entire GDN flow into a single kernel, we split it into two kernels optimized for CP and backward efficiency. At large batch sizes this incurs extra memory I/O overhead vs. a fully fused approach, but it delivers better real-world performance on edge devices and long-context workloads. The backward pass was the hardest part: we built a 16-stage warp-specialized pipeline under extremely tight on-chip memory constraints, ultimately achieving 2×+ kernel-level speedups. We hope this is useful to the community!🫶🫶 Learn more: 📖 Blog: qwen.ai/blog?id=flashq… 💻 Code: github.com/QwenLM/FlashQLA
Qwen tweet media
English
37
110
936
65.2K
Shivam Bhotika
Shivam Bhotika@shivambhotika·
Work Update: I have joined @WisprFlow India to do all things growth. For those who know me, you know how I got here. For those who dont, my habit of sending problematic voice notes got me here. If you spot a Wispr Auto in Blr, do say hi
Shivam Bhotika tweet media
Tanay Kothari@tankots

i grew up in delhi dreaming of building tech millions of people couldn't live without. today, @wisprflow is officially live in india! before this launch, i flew to india to answer one question: does wispr flow actually work here? in the back of an auto with horns blaring. a mumbai gym with punjabi music at full volume. a dhaba with the waiter rattling off the menu faster than you can type. we went and found out - it worked every single time. india became our second biggest market on its own. we 3x'd growth in 3 months with no campaigns or partnerships. people just found wispr flow organically and made it part of their daily life. the least we could do was show up for them properly. so we're launching wispr flow in india with hinglish & android support. because it's the way i've spoken my whole life. and the way everyone around me still does. grateful to my co-founder @sahajgarg6, our india lead @findingnimo_, and everyone who made this possible.

English
69
3
428
30.9K
dope-a-meme
dope-a-meme@aannuujX·
Introducing Swiggy Builders Club We’re opening @Swiggy commerce infrastructure to developers and enterprises to build on top - build AI agents, apps, and integrations on top of Swiggy’s Food, Instamart, and Dineout ecosystems - with real APIs, real data, and real users. What you get: 3 MCP Servers (Food, Instamart, Dineout) 18+ API tools covering the full convenience stack Production data access from day one Direct engineering support Who it’s for: Individual developers with bold ideas Startups building AI-native commerce products Enterprises looking to integrate Swiggy into their platforms Smart grocery restock bots. AI ordering assistants. Dining recommendation agents. Group ordering tools, health first products. If it makes commerce better for users, we want to see it. Ship something great and we’ll feature it. Ship something exceptional and our recruiting team might reach out.
dope-a-meme tweet media
English
174
134
2.7K
183.9K
Aarthi Ramamurthy
Aarthi Ramamurthy@aarthir·
If you’re looking for a quick and easy speech-to-text integration, you have to try Pulse by @smallest_AI. Took me all of 10 mins to integrate (powering WhatsApp voice notes), test, deploy. Good dev docs too smallest.ai/speech-to-text.
English
2
0
11
2.1K
anshuman
anshuman@athleticKoder·
what's wrong with you codex??
anshuman tweet media
English
1
0
5
646
•
@yducknow·
When I accidentally become important at work and now it's ruining my gym schedule
English
122
4.7K
45.5K
1.9M
anshuman
anshuman@athleticKoder·
@kirat_tw bro he is fixing security breach rn
English
0
0
2
929
Harkirat Singh
Harkirat Singh@kirat_tw·
Probably caught him at the wrong time. Im in SF btw doing a SF founder series. If you're here, please DM.
Harkirat Singh tweet media
English
65
10
1.3K
82.3K
anshuman
anshuman@athleticKoder·
@reach_vb I'm frustrated with how slow codex is
English
0
0
0
198
Gabriel Chua
Gabriel Chua@gabrielchua·
“Codex Hackathons have such great builder energy”, and Bangalore - you 10x-ed it. 🔥 To all the builders, thank you for joining us for a day of jamming with Codex.
Gabriel Chua tweet mediaGabriel Chua tweet mediaGabriel Chua tweet mediaGabriel Chua tweet media
English
6
12
137
7K
Rishit Bansal
Rishit Bansal@BansalRishit·
Had loads of fun building this with @AshikkaG and winning @OpenAI Codex Hackathon :) Also met a lot of cool builders with crazy ideas. Thanks for organising the event @gabrielchua @abhishekpatiil @yashrajnayak @OpenAIDevs @GrowthX_Club !
Ashikka@AshikkaG

We just took 1st place at the @OpenAI Codex Hackathon 🏆 Built Model Combat with @BansalRishit in ~6 hours. It’s a live AI security battleground: Models attack, defend, patch their own apps, and exploit others to steal flags in real CTF rounds. Mortal Kombat-inspired. Pure chaos. Extremely fun. Shout out @gabrielchua @abhishekpatiil @yashrajnayak @OpenAIDevs @GrowthX_Club and the whole team for organising this. #CodexBLR

English
4
0
21
925
Alex Mathew
Alex Mathew@alxmthew·
Introducing Berry (@berryaiplushies): the first anti-sycophantic AI companion in a stuffed animal. I'm only 17 and we just went viral on TikTok, raised money from incredible people, and now we're shipping to thousands of teens by Christmas.
English
69
17
413
69.2K
anshuman retweetledi
Ramp Labs
Ramp Labs@RampLabs·
Introducing Latent Briefing, a way for agents to quickly share their relevant memory directly. Result: 31% fewer tokens used, same accuracy. Multi-agent systems are powerful, but can be wildly inefficient. They pass context as tokens, so costs explode and signal gets lost. We built an algorithm that allows agents to communicate KV cache to KV cache.
English
37
91
1.8K
664.1K