My wife tried to perfect sourdough — temps, flour, water, time. Stats background → Design of Experiments. We tracked it on paper, so I built an app. #ExperimentA: experiment with anything. Free, local, no data 🍞 Feedback welcome! https://t.co/8pjRoeZBzg #buildinpublic#firstapp
@jacksonhinkle China is a capitalist country with 1 party system. US is a capitalist country with a 2 party system, the difference is just one party, that's why their results economically are similar. US has more freedom and more chaos, China the opposite.
@harjassgambhir@karpathy Cars drive in 2D, planes in 3D. A huge difference from AI point of view. The other is training data. The only reason Waymo and Tesla can train is because they get their own data. Also error. 1% of car mistakes end in fatalities, 99% of planes mistakes will end in fatalities.
@elonmusk Solution is to flatten structure and hire third party contractors that you can fire if they under perform. Government is bad a executing and moderately good at oversight.
@elonmusk We have big problems adjusting the thermostat in a family of 4 house. Who is going to be in charge of the cyber thermostat? I assume that benefits would be uneven, just like a house: too cold/hot rooms/cities.
@elonmusk is there are way to decrease the number of Elon Musk posts? At least to hide reposts? I don't want to unfollow Elon but it is simply too much on my feed. (I'm aware than posting this increases the problem even more)
@AmericaPartyX Keep deporting hard working, crime-free, family-oriented Mexicans and the numbers will keep dropping. If fertility is top priority for humanity, I'm sure immigration paper work and a small fine should be an easy fix for this critical issue
@AmericaPartyX I only bet once against Elon and I was right: DOGE. I am 100% Elon will be put the first people safely on Mars but I knew DOGE was an impossible task not only for him, for anyone.
@karpathy Have you tried white noise? I bought a little cube with different sounds. I was waking up due to outside cats, shower water drops, ice cubes generation from fridge, etc. My problem was that my neighborhood was extremely quiet and your ears become extremely sensitive.
@ElonMuskAOC Learn. It's all we do. This also answers the question: what does our creator wants us to do in this reality? Learn. Pass the knowledge to our next generation and then learn more. Check or routing number is fine?
@musk_news13 I believe there is a huge potential conflict of interest but I like the initial transparency of receipts posted on https://t.co/ECnI4ksYTu. And honestly, only the richest man in the world can stand a chance of actually achieving something in waste reduction.
New course with @Intel! Multimodal RAG: Chat with Videos is available starting today!
Learn from @Vasudev_Lal to build an AI chat system that answers questions from video content!
Learn to:
🔄 Create embeddings from videos
🧩 Build a RAG pipeline for data retrieval
💬 Use Large Vision-Language Models (LVLMs) for Q&A using both text and image inputs.
In this course, you will make API calls to access multimodal models hosted by @PredictionGuard on Intel’s cloud.
Learn more and enroll for free: https://t.co/Ue7aYR80Qn
@elonmusk That's a lie. First, illegal immigrants don't vote and what people don't realize is that Hispanics which are "overwhelmingly" Catholic share a lot of the same beliefs as Republicans in terms of abortion, marriage, free market and even the right to bear arms.
Intel presents LLaVaOLMoBitnet1B
Ternary LLM goes Multimodal!
discuss: https://t.co/gjK5hs7ImQ
Multimodal Large Language Models (MM-LLMs) have seen significant advancements in the last year, demonstrating impressive performance across tasks. However, to truly democratize AI, models must exhibit strong capabilities and be able to run efficiently on small compute footprints accessible by most. Part of this quest, we introduce LLaVaOLMoBitnet1B - the first Ternary Multimodal LLM capable of accepting Image(s)+Text inputs to produce coherent textual responses. The model is fully open-sourced along with training scripts to encourage further research in this space. This accompanying technical report highlights the training process, evaluation details, challenges associated with ternary models and future opportunities.