Загрузка видео...

Не удалось загрузить видео

На главную

A simple visualization of Fluid efficiency Like a server, the same function instance can deal with multiple in-flight requests. Great for '↻ Thinking' LLMs. This is not a 'demo'. It's backed by real Fluid functions that report back their UUIDs (at 75% less cost.)

36,096 просмотров • 1 год назад •via X (Twitter)

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

This looks like a simple transparent shock absorber filled with oil. But what you are seeing is one of the most destructive phenomena in fluid engineering. This is cavitation in its true form. The white cloud forming beneath the piston is not foam and it is not air. The oil is literally changing from liquid to vapour at room temperature. When the piston moves rapidly, the oil is forced through tiny passages inside the damper. The fluid velocity increases, the local pressure drops, and if it falls below the oil's vapour pressure, the liquid begins to boil without any increase in temperature. The moment the pressure recovers, those microscopic vapour bubbles collapse almost instantly. And that is where the real damage begins. The destructive forces of cavitation is really not understood well by most. A collapsing cavitation bubble creates shockwaves and high-speed microjets that strike nearby surfaces with enormous local forces. Repeated millions of times, these tiny implosions can slowly eat away hardened metals, destroy precision components and reduce the lifespan of expensive machinery across industries. This same invisible phenomenon is one of the biggest challenges in naval engineering. Ship propellers operating under enormous loads can suffer cavitation erosion, losing efficiency while creating underwater noise. For advanced stealth submarines, that noise can become a major problem because cavitation can reveal their position. Decades of research have gone into specialised propeller designs, pump-jets, surface finishes and hydrodynamic optimisation to delay its formation. The same issues affects hydroelectric turbines that convert the energy of entire rivers into electricity, and industrial pumps that move oil, chemicals and water through critical infrastructure around the world. Perhaps the most remarkable part is that after 4-5 decades of advances in metallurgy, coatings and manufacturing, engineers still cannot simply build a material that is immune to cavitation. The solution is not to make stronger metals forever. It is to understand the fluid dynamics so precisely that cavitation is prevented before in those destructive bubbles ever form.

Ammanichanda

1,188,577 просмотров • 1 месяц назад

THAT $70 "RUN YOUR OWN LLMS" PI KIT CAN'T RUN A SINGLE LLM. IT'S A VISION CHIP WITH NO RAM. that clip sells a raspberry pi 5 in a slick case with an ai accelerator and the caption "your own llms." clean build, fun kit. the claim is where it breaks. the fine print: the popular $70 pi ai kit uses a hailo-8l, 13 tops. it's built for vision, object detection and image processing, and it has no memory of its own. so it cannot run large language models. full stop the board that actually can is a different one: the newer ai hat+ 2, hailo-10h, 40 tops, with 8gb of dedicated ram. that's $130, not $70 and even that runs only tiny models. llama 3.2 at 1b, qwen 2.5 at 1.5b, deepseek r1 at 1.5b. edge llms live in the 1-7b range, against cloud models at 500b to 2 trillion so the honest pitch: for $130 you can run a very small language model on a pi, slowly, as a fun learning project. that's real and it's cool. "your own llms" on a $70 vision kit is not. why this keeps happening: "ai kit" and a big "tops" number sell. tops sounds like intelligence. but tops measures vision-style math, not whether the chip has the memory to hold a language model. the spec that matters for llms is ram, and the cheap kit has none. the honest caveats, both ways: the $70 kit is genuinely great, just at vision. cameras, object detection, that's its job the $130 hat really does run small llms locally, which a pi couldn't do at all two years ago. that's progress "small" is the load-bearing word. don't expect gpt at home on a pi the takeaway: before you buy a kit because the caption says llm, check two numbers. not the tops. the ram, and the size of the model it can actually load. no 70-dollar miracle, no gpt in a pi case, no tops number that means what you think. save this before you buy the wrong kit for the word on the box.

RetroChainer

11,100 просмотров • 1 месяц назад