Keep on to blur preview images; turn off to show them clearly
RT @nathanbarrydev: Added confidence-aware parallel decoding to my tiny text diffusion model! Before, we had “scheduled iterative refineme…
RL and efficient distributed pretraining • eXperiments lab • memes and training lores
Looking forward to chatting with @anthemos tomorrow, where we'll be diving into how @Zumper is leveraging AI as a leading housing marketplace. We'll also be discussing topics like AEO, so this is a group chat you won't want to miss. DM me if you'd like to join in!
Founder of Everything Marketplaces (@marketplaceshq). Always working with & investing in marketplaces at https://t.co/HgyZIpWIEQ
- “wen K3?” - “before sam's trillion-dollar data center is built” 😂 AMA link: https://t.co/6yZSsjQXvM
Co-founder & CTO @hyperbolic_labs cooking fun AI systems. Prev: OctoAI (acquired by @nvidia) building Apache TVM, PhD @ University of Washington.
Kimi AMA on K2 Thinking: 1. $4.6M training cost is not an official number 2. Trained on H800s (nerfed H100s) 3. KDA (Kimi Delta Attention) hybrids with NoPE MLA perform better than full MLA with RoPE 4. Muon scales well to 1T parameters. “there are tens of optimizers and architectures that do not survive the grill.” 5. Kimi K2 will have vision 6. K2 Thinking is natively INT4 to be friendlier to non-Blackwell GPUs while leveraging the existing int4 inference marlin kernels.
- “wen K3?” - “before sam's trillion-dollar data center is built” 😂 AMA link: https://t.co/6yZSsjQXvM
Assuming the model companies can touch every market and, more importantly, *do it well*, and that the advantage to being multi model is limited feels a lot like sitting out the last 20 years because Google/Amazon/Facebook could be the only winners
partner @a16z // saas + b2b fintech // strong opinions on 🍕
RT @jamisonfox: Big milestone for @GammaApp today: we’ve raised a $68M Series B led by Sarah Wang at @a16z. It’s been humbling to see this…
Growth investing @a16z