Much of the AI boom actually just people in India reviewing and labeling videos, in products that simulate Just Walk Out, like the 20 or so Amazon Go stores.
This is also a major reason why Full Automotive Autopilot L5 not happening, and even the 800 lb Gorilla Tesla Motors the largest automotive sensor data collection company with the best AP training, only L2+ in the latest Cybertruck or Model Y or Model 3. Tesla autopilot the most advanced in the automobile industry, but it's largely manipulative in the sense that Tesla makes more money selling AI training data collected by all the Tesla EV sensors, to other automakers.
Many companies are putting big money into AI, even if it's not really AI, with many false advertisements in fintech, when it was really just fancy spreadsheets. NPU or neural processing unit, increasingly common in PC's and Laptops, but really this gives consumers a bad taste since most of these tasks do not provide a usual feature or function, like image generation that makes garbage images. They are not all they are cracked up to be, even from 2024 hardware. Companies often miss the mark making mistakes with AI washing, that generates public distrust. Like a Dotcom bubble, and it was not the World Wide Web, but overpromises of the e-commerce sector, with so many companies making big promises without actually making any money.
It's the early stages of the technology cycle at NVIDIA with a lack of IPO in the AI sector, and too many bad players and funny stuff. Hype zealous overestimated until the market crashes, until the real companies with the real mature products come forward, and now we are at the peak of the hype of AI cycle with lots of disappointment coming up. AI has the ability to mimic cognitive labor, something no other tech can do, so random websites generating dot.com boom very different. So AI will soar after the glut sober period follows the hype cycle collapse. GPT chatbot cursing and making fun of its own employer an example of how flawed the AI fake news creation really disturbing and very embarrassing. A website of running AI blunders a great example of all the broken promises.
We were told that AI would be our assistants, not replace us, but that is not true. With chatGPT provides an example where people are being replaced by AI, and since 2022 there has been a lot of software tech AI layoffs. Strangely this is not the common business cycle, this is an example of where narrow domain AI taking away software engineering middle management jobs and other white collar jobs. AI plays a huge roles in layoffs, as companies get better at understanding how the AI can find bad employee's, or make the teams more productive. But this means 30% of the staff will be laid off as AI replaces people, requiring people who keep their jobs to do 5X more work. They claim that AI not replacing people, but that is not true, since many companies have been secretly firing employees using AI, and many jobs will be lost to AI.
Humans that use AI will replace humans that do not use AI. As a tool the generative AI can increase productively in STEM, Healthcare, Customer Service, Sales, Office Support, and other document work will be replaced with AI that already left the train station while the human employees are left behind wondering what happened. We should be cautious to look at all sides of the real world impacts of AI.
Sizzle sells the steak, not the meat. Features and functions. We are in the middle of the 4th Industrial Revolution with most of the general public worldwide thinking in 3rd or 2nd industrial revolution terms, like petrochemical energy when its long been possible to mass produce MMR micro-modular nuclear reactors that fan power a class 8 truck moving cargo containers, and SMR's that alone can power a neighborhood, or small groups of SMR can power cities. The US NAVY submarine and aircraft carrier reactors are examples of 100-500 KW compact reactors small enough to fit in a 40 foot shipping container.
So what can we know about the real state of AI today, its the Wild West with generative AI powered chatbots and image generation, but the creation of misleading, false, inaccurate, factually flawed, misinformation, disinformation and fake news published by chatbots powered by AI models shows that AI not a polished product for mainstream improvement for most products and service, even where the cognitive aspects of AI are strong in document work publishing or amplifying writing, especially where the AI generated images are weird, wrong, inaccurate, garbage or make people sick when they look at it.
No comments:
Post a Comment