Leading edge fabs are coming online in Arizona for Intel and TSMC upcoming chip nodes for power efficient mobile SOC's used in smartphones and many other electronics.
With new Integrated Circuit chips we are talking about systems of billions of connected transistors where each transistor is smaller than a corona virus particle. Chips like this are made on large silicon wafers at Chip Fabs using leading edge DUV and EUV lithography by ASML.
More than 200 stages of processing go from melting purified sand in argon purged spiral spin extraction boule making furnaces to a finished cut bonded packaged chip that can be soldered or connected to the mainboard logic of an digital electronic device or system.
We are talking about the most complex manufacturing process making the most complicate products with the smaller transistors ever, at high volume 24/7 since the demand for chips increasing steadily as they are integrated into more types of products, often in groups of chips, sometimes even hundreds of chips in a new car or SUV as of 2023.
5nm node a well establish mature high volume commercialized chip making process in Taiwan by TSMC to make the A17 processor for Apple Silicon in the latest iPad, iPhone, iMac and other consumer electronics by Apple like the Series 9 or Series 10 Apple Watch or Apple Watch Ultra 2 and Ultra 3. In small battery powered electronics the chips have to have the best power efficiency, so the latest chip making node since smaller transistors use less energy.
In a smartphone, smartwatch or smart ring, sensor fusion combines with the SOC which contains all aspects of the computer into a single chipset. CPU, GPU, RAM, FLASH, Networking, IO, Power & more are also zones of the SOC layers and interconnects.
Modern smartphone SOC have more than 20 billion transistors in something smaller than a typical fingernail, thats able to operate on a 3.85vdc lithium polymer battery of about 10 watt hours for 18-40 hours of runtime when the battery is new & fully charged.
After a few years of warm operation in peoples hands & pockets, the fragile fade prone lithium ion polymer batteries lose 10-40% of their charge holding capacity, so runtime per charge decreases the same amount. 100% charging is what accelerate the damage of Li-Ion batteries in electronics & EV's
Loss of lithium in the carbon or silicon or nano-wire anode with each charge cycle means a slow progressive loss of runtime or capacity loss. This means that EV's lose range as they age, meaning they can't drive as far on a charge as the battery weakens with time and cycling. 30-80% charging makes common lithium ion batteries last 3-10X longer!
A modern automobile contains hundreds of chips, especially for autonomous modes & advanced driver assist features and functions. Radar, lidar, camera, AI, synthetic vision systems, collision prevention, lane keep assist, automatic high beams, and many other autopilot or one pedal driving modes with adaptive cruise control and other features like pedestrian detect and road sign reading information displayed persistently on the dash, so if you missed the latest speed limit sign you can just look at the dash to know.
I am not giving an exhaustive list of ADAS, but we need chips to make modern appliances, vehicles, internet and telecom hardware, consumer electronics that are wildly popular worldwide, and many other products like hearing aids, printers, routers, modems, gateways, smart lights, thermostats, the World Wide Web, credit cards, payment terminals, debit cards, banks, finance, accounting, business management, software, middleware, firmware, e-commerce, heat, light, fans, pumps, relays, controls, and more to improve energy efficiency in automated manufacturing, in security systems, in all sorts of systems and devices in hospitals, and in law enforcement. Chips are broadly important in most sectors of most economies and an essential part of the future even more important than oil or coffee to economic development.
Watch this YouTube video https://youtu.be/_TOCRjF9WuE?si=AEPeCDakgV3SWGD-
A transistor a simple electronic solid state switch to turn the flow of current on or off! Switched by the clock at 4 GHz or 4 billion on off cycles per second, these IC chips are very different than the 10 Hz 86 billion analog brain in the human skull.
Neuro-morphic Spiking Neural Networks are CMOS chips that aim to act as analog gate neurons in digital compute systems, for specialized AI tasks for imaging for smartphone photography and video capture for example. Consider IBM TrueNorth from 2014 and IBM Watson.
Synthetic computer vision needs flexible analog processing similar to the human brain so computers can intuitive learn about the world like a real human person does during life. Machines will not become conscious because they will lack a limbic system, are not born and do not die, are unable to love or be loved.
Advanced flexible learning AI will be able to accurately simulate super human intelligence in most domains of life as autonomous drivers, in manufacturing, to perform superlative control of HVAC systems with class leading energy efficiency that actually learn from the building and air handling systems and air quality sensors, how to best manage the facility with the least amount of energy, always iteratively improving by learning more.
Learning systems mean that we can have smarter lighting, variable control of pumps and fans with on-demand output scaling and eco-throttling to save energy when demands are low. AI can improve grid power stabilization for example, by matching power plant production capacity to varying loads created by homes and business electrical utility customers.
Today Google's (Alphabet Inc) bot crawlers have super human reading intelligence, as they fly around the World Wide Web collecting and ranking information from different websites. Amazon has similar AI that monitors Prime users on Amazon.com to make product suggestions, to work with Alexa to do the same, and to enhance the user experiences of Prime customers.
AI applications are in a broad array of industries to improve cyber and network security by understanding layered user permissions and who is trying to access the information. AI persona verification (PRISM) with audio sampling of the humans voice and video analysis of the person face in 3D from the smartphones front camera, utilized so that intelligence agents can access sensitive national security information over 4G LTE or 5G-NR or wireless internet connections while traveling abroad. Even commercial customers accesses AWS or other software as a service platforms have to have secure access, just as people shopping online need protection of their financial and personal information. I didn't even mention cryptographic currency or cryptography and encryption, but those are extremely important for home titles, vehicle titles, ownership verification, taxes, insurance, and many other aspects of business, accounting and finance. Banks in particular are main users of network security technology.
Imagine Smarter WIFI at home that can tell you about your body and healthy by using RF multi-pattern scanning of your body, brain, muscle, bones, with wireless non-destructive testing to give you health data and insights about your health, aging, fitness, diet and more. Imaging a whole home AI that controls smart lights, the furnace, air filters, the fridge, oven, microwave, windows, blinds, and even gives edifying suggestions about how to cost effectively improve the energy efficiency of your home as it studies and learns how you use the home, about the home, about the systems of the home.
Imagine the AI as the ultimate personal assistant that seamlessly follows you from your home into your vehicle and wherever you go, on your phone and smart clothes & smart accessors. It can even send this health data to your doctor who can video conference with you and even send you sampling gear to sample your stool, urine, sweat, blood, tears and other body fluids, that you send back to the lab or doctor for enhanced analysis. Thus the Health Data systems will given patients better useful information like a Dexcom G7 helps people with Type 1 diabetes to actively monitor their blood glucose today.
Imagine a computer system that helps you to be a better parent of your children, a better spouse, a better neighbor, a better friend. Imagine a thoughtful considerate kind & respectful AI that gives gentle thoughtful suggestions to help people go for a walk more often, to eat better for their health, to go to bed earlier by turning off the screens and reminding them about sleep health importance. Imagine that the AI helps a person with personal finance to achieve their goals in long term investing. There is so much potential to make a huge positive difference in peoples every day real world lives with emerging technologies.
No comments:
Post a Comment