I've been thinking about all of the ways to make a complicated product more elegant. ie: progressive disclosure, model selection, upselling to existing customers through customer marketing vs. an overwhelming number of products on a single page Then I think about @higgsfield_ai and sometimes chaotic can be good. The more rabbit holes the better.

I went to Milan for a few days and came back so inspired to be in SF. A little art, lots of good food, and a cute train ride was exactly what I needed to get some inspo for building the creative community here. https://t.co/rlQcIcVyXg

+2 more

In NYC they really got people 67ing at 6:07 on 6/7 https://t.co/crvCDmO6Ir

@b4bendetta Why spend millions on a launch video and paid ads with low recall, when you could build community IRL. As an investor, I’d much rather see a company that understands culture and connection. Ramp is only company doing this well. https://t.co/sx5kgDZzTt

It’s being heavily rumored that ‘OBSESSION’ lead star Inde Navarrette was only paid $20k for her role as Nikki Freeman. https://t.co/Wq0usslz6l

Hosting a night for builders & creators at the @contralabs_ai NYC office next week for a special discussion with @briannekimmel , Founder of @WorklifeVC and @pirroh, President & Head of AI of @Replit Founders, builders, designers, product leaders, developers, creators, and AI folks, this one’s for you https://t.co/CMklclV7ky

GX users, have you been thinking about making a Custom Expectation? Get started on the right foot: use this flowchart to ID the Expectation class that’ll get you the result you’re looking for! https://t.co/Ur7NPfAkLe

If you’ve been using GX and thinking about creating a Custom Expectation, now’s the perfect time to start: use this flowchart to ID the Expectation class that’ll deliver the kind of result you want! https://t.co/fZMY4PoZ1I

📣 Our March community meetup is tomorrow, 3/19 at 9am PT! We’ll be sharing the latest updates about GX 1.0, and more–join us at https://t.co/mPSBfD59s6

Finns! Swedes!! https://t.co/cITVVb7BvN

@hvnorris @futuredrwillis I assign this: https://t.co/AODaWeVu7R

We used it to find clusters of similar images, duplicates, and outliers in a subset of the LAION dataset Here's what we find. https://t.co/J9Ef5hJRS3

Try it on your own dataset: Colab notebook - https://t.co/TBMIKh5Mm9 GitHub repo - https://t.co/WqhB4lpQQ7

SDXL 1.0 is currently the best image model from @StabilityAI. With @huggingface transformers, accelerate and diffusers package, run the model in less than 10 lines. On free Colab (T4 GPU). It takes less than a minute to generate an image. Colab nb - https://t.co/dOdPwxtTnj https://t.co/VxnM2jjltj

TLDR; 1. Use ONNX and ONNX Runtime. 2. Use TensorRT + optimized parameters. 3. Bake preprocessing as part of ONNX model. Details -https://t.co/pnz1wCbKmC

Optimized model on @huggingface - https://t.co/glZYtOgsWJ Gradio demo using the optimized model - https://t.co/p0C2pHWhnJ

Code to reproduce the results - https://t.co/wIgmFZP9xb

I'm on the GitHub trending developer list today! Never thought this was ever possible. Thank you for supporting my open-source work on x.infer - Framework agnostic computer vision inference. https://t.co/7BPzI8Ggkc https://t.co/ppe1HueXCa

I spent the last 3 weeks working on my first open-source package - x.infer: Framework agnostic computer vision inference. Load a model from supported libraries and run inference with 2 lines of code. Supports 1000+ models from transformers, ultralytics, timm, vllm, ollama. https://t.co/96ThhC0wGE

active-vision includes a @Gradio labeling UI to label images sampled using active learning strategies. https://t.co/DQwJQJ27sM

I wanted to apply active learning to computer vision but couldn't find many resources. I spent the last month building `active-vision` a framework for AL focusing on computer vision. Framework heavily based on @fastdotai. Work inspired by @WWRob book Human in the Loop ML. 🧵 https://t.co/ZhDTBMWrkq

Help me RT if this has been useful to you https://t.co/AHyMndTi8Z

I hate configuring yml files, so I made a python wrapper - DEIMKit Repo - https://t.co/85B8girTc1 The motivation is simple - Start training on a custom dataset using only clean Python code. No yml configs. https://t.co/IHGQBrKlEz

Steps to become a senior programmer: 1. Install my /teach skill npx skills add mattpocock/skills --skill teach 2. Create a new working directory on your laptop mkdir junior-to-senior cd junior-to-senior 3. Kick off your coding agent in the directory claude 4. Copy this prompt /teach me how to be a great strategic programmer. My opinion is that AI is eating 'tactical, on-the-ground' programming. The day-to-day work of a developer involves not only coding, but also planning, QA, codebase design, and much more. I'm interested in learning the strategic skills - that, in a previous era, would take me from junior to senior - but in this era are table stakes. 5. Paste it into the coding agent Below is an example of what the first output will look like. I used Opus 4.8, medium effort. 6. Continue working with the agent until you're a senior

https://t.co/FJi2laxfok

https://t.co/FJi2laxfok

Nvidia is proposing a beast of a CPU system for Windows PCs. It has 128 GB of shared memory and comes with up to 6,144 state-of-the-art CUDA cores. CPU wise, the chip has 10 performance cores and 10 efficiency cores. The performance cores are based on the Cortex-X925. These chips appear to support six 128-bit SIMD execution units (SVE2), not as good as recent AMD chips, but better than Apple Silicon (on paper). The game changer is the unified 128 GB memory. That is the path Apple took years ago. Instead of separate memory for the CPU and GPU, everything shares a single pool. It is increasingly popular. The memory is not as fast as dedicated GPU memory, but it is cheap enough while delivering enough bandwidth to run AI models locally. I am not sure how many people will run AI models locally. It still seems like a niche application to me. However, it will make decent machines to play video games. It will be interesting to see how Intel and AMD respond. I think that the AVX-512 instructions supported by all recent AMD processors are far superior to the SVE2 instructions of the Cortex-X925. They can eat more data and they are more versatile. But Intel has been shy, thus far, in making it available on customer systems.

a reminder that branding matters. https://t.co/y9WOLdp3GM

a reminder that branding matters. https://t.co/y9WOLdp3GM

Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers. Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?

DeepSeek-V4 now runs 500K context with 90% less KV cache FlashMemory introduces Lookahead Sparse Attention: a tiny Neural Memory Indexer predicts which chunks your next tokens will need, keeping only 13.5% of cache in GPU memory—with zero backbone retraining and better accuracy. https://t.co/qxclBmFV3n

Why Fable 5 til the 22nd of June? Wtf?! @claudeai https://t.co/xjvxnT3F48