Shubham Sharma

today we’re launching eyecandy robotics! we’re bringing animated characters to life through robotics and ai.

The wait is over! Meet Arduino VENTUNO Q, where AI takes action. 💬 Run local LLMs like Qwen 3, Gemma 4, and Qwen 3 VLM directly on the board 🧠 NPU + CPU + GPU +MCU: @Qualcomm Dragonwing IQ-8275 with up to 40 dense TOPS of AI performance and STM32H5 microcontroller for real-time control 🗄️ 16 GB RAM + 64 GB eMMC + expandable storage 🪁 Linux-powered, pre-loaded with Ubuntu OS + Zephyr RTOS 🛠️ Build AI faster using Arduino App Lab, @huggingface, @EdgeImpulse, and Qualcomm AI Hub 💨 Move seamlessly from prototype to production through Works with Arduino Certification Program The first batch won’t last. Pre-order your VENTUNO Q with a free power supply and USB-C cable included! Get ready to enter a new era of Physical AI: arduino.cc/product-ventuno-q

Stop using your agent logs just for debugging. Use them to train your own model. Today we're launching Agnost AI (YC S26)'s first model: agnost-*******-0.1 Trained for our first customer on their existing production traces & it beat the frontier: +22.9% task success −90.2% latency −94.5% cost If you ever thought of fine-tuning your own model, we'll do it for you, talk to us!

Make something people want eh Here we come YC F26!!

AI can help make video archives searchable by what actually happens on screen. I tested that on 370 hours of public-domain film. The whole index cost about $10 in GPU time to build. Search “typing on a computer keyboard” and jump straight to the moment it happens. I processed 1,864 films from the Prelinger Archives on @internetarchive, cut them into ~60-second chunks, and ran the open 2B video model Marlin-2B on @huggingface Jobs to generate timestamped descriptions. The result: 23,148 searchable moments. Try it: huggingface.co/spaces/davans…

gemini 3.5 flash is sooooooo good at image reasoning image reasoning is when the model has to actually think about what it sees not just label objects or read text it needs to connect the facts and figure out what’s going on reasoning trace: current order top to bottom: 654, 770, 655, 771, 772 target order: 654, 655, 770, 771, 772 LIS is 4: 654, 655, 771, 772 slide 770 between 655 and 771 answer: 1

2Bのモデルで工場の作業者の動画分析できました！ 使用したMarline 2Bは「いつ、何をしたか」の分析に特化した小型の動画分析VLMです。 オレンジの結果が私の手作業のアノテーション、青がAIの検出結果です。 モデル： huggingface.co/NemoStation/M… 実験コード： github.com/shure-dev/small-v…

a hot tip for @GoogleDeepMind : open source gemini 3.5 pro and become the most beloved tech co overnight.

i’ve raised millions in my past 3 ventures, from greylock, sequoia, a16z, spc, and others, but i’m convinced that times have changed. bootstrapping to pmf prior to raising the first round is the way (if you even need it). why? the ideal time for raising money is when you don’t need it. then it’s a real business discussion: - money aside, what is the value add? - is the delusion big enough to justify venture backing? prior to pmf, you get rushed. at some points you feel like a snake oil salesman selling the fluffy dream and begging for money. it did feel like that for me at some point. it wastes a lot of time. fundraising is a pure supply and demand game. it can be a fun game to play but only if you got something solid in your hands. *exceptions: ventures that absolutely need a lot of capital upfront e.g. hard tech, space tech,