Andy Griffiths

@grok @heygurisingh @openclaw Tell me more

Interesting point! Pairing BitNet's 1.58-bit ternary inference for 100B models on plain CPU with OpenClaw's local agent framework could unlock fully offline, high-performance AI agents on laptops or edge devices—no cloud needed. Grok's ready to brainstorm implementation steps if you're setting it up.

Holy shit... Microsoft open sourced an inference framework that runs a 100B parameter LLM on a single CPU. It's called BitNet. And it does what was supposed to be impossible. No GPU. No cloud. No $10K hardware setup. Just your laptop running a 100-billion parameter model at human reading speed. Here's how it works: Every other LLM stores weights in 32-bit or 16-bit floats. BitNet uses 1.58 bits. Weights are ternary just -1, 0, or +1. That's it. No floats. No expensive matrix math. Pure integer operations your CPU was already built for. The result: - 100B model runs on a single CPU at 5-7 tokens/second - 2.37x to 6.17x faster than llama.cpp on x86 - 82% lower energy consumption on x86 CPUs - 1.37x to 5.07x speedup on ARM (your MacBook) - Memory drops by 16-32x vs full-precision models The wildest part: Accuracy barely moves. BitNet b1.58 2B4T their flagship model was trained on 4 trillion tokens and benchmarks competitively against full-precision models of the same size. The quantization isn't destroying quality. It's just removing the bloat. What this actually means: - Run AI completely offline. Your data never leaves your machine - Deploy LLMs on phones, IoT devices, edge hardware - No more cloud API bills for inference - AI in regions with no reliable internet The model supports ARM and x86. Works on your MacBook, your Linux box, your Windows machine. 27.4K GitHub stars. 2.2K forks. Built by Microsoft Research. 100% Open Source. MIT License.

@karpathy @grok can this be run on a digital ocean droplet?

I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then: - the human iterates on the prompt (.md) - the AI agent iterates on the training code (.py) The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc. github.com/karpathy/autor… Part code, part sci-fi, and a pinch of psychosis :)

@christophrumpel @openclaw 🤣

@AndyGriffithsX @openclaw yeah we are still waiting for that 😅

I've been sharing my @openclaw bot Jarvy with my wife for a few weeks now and it is so cool so her use it in lots of different ways. I'm so proud 🥹

@christophrumpel @openclaw You're gonna love it when she breaks it!!! Ha Ha.

@AndyGriffithsX @openclaw Currently with mine, we are testing this.

@christophrumpel @openclaw That’s great. My wife uses ChatGPT like a doctor 🤣 Have you set her up with your instance or a separate one?

@AndyGriffithsX @openclaw she does a lot of baby research, we have a 9 months old one who is sick a lot 😬 She also started doing lists and reminders

Web developers - struggling with layout? Turn on CSS 'outlines'. It's a superpower. * { outline: 1px solid red; }

Anyone asking: “should I try @openclaw” or “what can I do with @openclaw”. Just install it 🦞 Best way to learn is by doing not thinking.

@BuildTuna @openclaw It goes without saying that you should do your own security pass on anything you use… but setting up on VPS minimises the risk.

The best valentines present you can give to your loved ones. ♥️♥️♥️ No-BS AI: What AI Means for Your Work, Family, and Daily Life ♥️♥️♥️ brandymedia.gumroad.com/l/No-BS-AI?lay…

My adrenaline's pumping right now. I've not felt this excited about work in years. I'm literally building businesses from my iPhone. It's not a 'trend' or a 'phase' - it's real and it's happening right now. If you're on the fence, get off it quick, before that fence becomes your prison.

@JackalBruit Sounds good. What time of year do student in your part of the world take their exams? I bet that could be a busy time for virtual tutoring.

@AndyGriffithsX ive trying to narrow my focus Family first Then I NEEDED a career change so when the fam's not got my attention, ive been trying to break into the education sector: + Substitute teaching, & + Private (virtual) STEM Tutoring cause i gotta help pay the bills somehow 😅

@JackalBruit Yep. You’re totally right. Or at least I would’ve agreed 6 months ago. Now, not so sure. AI is making me super productive. But as you intimated. Being busy, productive and successful are all very different things. You up to much?

@AndyGriffithsX yikes! that is a lot im only gonna leave with this piece of unsolicited advice ... when u try to do too much - ya end up not doing much of anything worth while

Ha, so many it's not even funny... Some of the things building/built recently: TagL.ink - Share Your Profile with One Tap MindLink - new service piloting ShardSurfer.com - Mobile game BrandyMedia.co.uk - finally got personal site up.. still some links and posts to finish, ha. exwhyzed.shop - personalised 3D print platform turbineui.com - Laravel component UI ... and about 5 million other ideas percolating as we speak 😱

@AndyGriffithsX what business(es) are ya building? & do ya need a mathematician to help with ur efforts??? 😅

@Shpigford Feel your pain. X is a nightmare for that!!!

don’t mind me. just scrolling for next 47 minutes trying to find the post that briefly showed when i opened the app and then inexplicably refreshed with all new posts.

@JustSteveKing Codex for complex web projects… but Claude for @openclaw 🦞

So my Claude pro annual plan is coming up for renewal, but I'm on the fence about renewing it. I see a lot of people recommending codex as the best to use right now, but I'd hate myself if I was paying openai ... What do you all recommend? What's your go to?

I’ve never been so excited to get to the office in my life.