AI Top-of-Mind for 5.13.24 — Oligopoly
Top-of-mind is industry consolidation. Looking first at OpenAI, ‘The Information’ reports on the company’s new voice assistant, designed to compete with Google, Apple, and Meta. And it could be previewed as soon as today (May 13). From the article:
OpenAI sees assistants with visual and audio capabilities as potentially as transformative as the smartphone. The assistant could theoretically do a range of things not possible today, such as acting as a tutor for a student working on a paper or on math problems, or giving people information about their surroundings when they ask for it, like translating signs or explaining how to fix car troubles.
But even with this level of innovation, do OpenAI and other smaller AI specialists stand a chance against Alphabet, Amazon, Apple, Google, Meta, and Microsoft? Not to mention Nvidia. AL Anany looks at the evolving landscape and partnerships, and one could imagine the companies listed under ‘FM partnerships/agreements’ absorbed at some point, dependent upon prevailing views on monopolies.
A Linkedin post from me in 2017 that foreshadowed this same trend.
He also notes that, taking a page out of history, banks and financial institutions were the main source of capital for the era’s industrial giants, and their unethical collusion was what led to the birth of Antitrust legislation in the first place!
· Today, Big Tech’s influence on the economy resembles that of banks way back when, but on a much higher level:
· They have phenomenal volumes of capital, as well as more data about consumers than banks ever did.
· Big Tech firms also supply vital products and services to businesses and consumers across every part of our economy.
· “[Tech Giants’] control over AI’s ‘upstream’ infrastructure means they can easily identify any serious potential rival in its earliest stages and then move swiftly to crush, sidetrack, co-opt, or simply acquire the upstart.
A brief follow-up on my recent Boston Dynamics ‘Spot’ and ‘Atlas’ postings. Writing in “Ai-Ai-OH,’ Jeff Hayward looks at current trends, especially initiatives by the DoD, and links to a ‘Defense.gov’ article that sheds more light on AI swarms. From the post:
· It explains that the U.S. Marine Forces Special Operations Command are testing “four-legged unmanned ground vehicles (UGVs) armed with artificial intelligence-enabled rifles.”
· According to the article, these robots can automatically detect targets, but require “authorization” from humans to finish the job.
Still on dystopia, CNN looks again at Apple’s iPad ‘Crush’ ad and its backlash. If you remember Jurassic Park, from the article:
“Don’t you see the danger, John, inherent in what you’re doing here? Genetic power is the most awesome force the planet’s ever seen, but you wield it like a kid that’s found his dad’s gun … You stood on the shoulders of geniuses to accomplish something as fast as you could, and before you even knew what you had, you patented it, and packaged it, and slapped it on a plastic lunchbox, and now you’re selling it.”
And a few other updates from the model front. The first from The PyCoach writing in ‘Artificial Corner’ on how to better leverage your paid ChatGPT subscription. These include voice interactions, custom GPTs, formula generation, and even analysis of what you’ve drawn on a whiteboard. The second is a wakeup call from Daniel Warfield in ‘Level Up Coding’ on the lack of real failsafes. He covers copyright and the growing number of lawsuits, then homes in on how the more unscrupulous are ever-more creative in ‘jailbreaking’ LLMs and uses ASCII art as an example.
And the third looks at Meta, with Al Anany concluding that they are winning the LLM race. Part of his focus is on the open-source nature of Llama, the ability for Meta to leverage it across their platforms, and the fact that it can run locally due to its lightweight variants. As a note, I plan to base my planned industry fine-tuning on Llama.
Then onto creative, with a new video tool, Vidu, out of China. Jim Clyde Monge writing in ‘Generative AI’ details its capabilities and whether it holds its own against OpenAI’s Sora. The published demo reels look much the same, and only time will tell. Vidu is not yet available, but you can join the waitlist here.
But do these AI videos seem real, or do they only offer a distorted view of reality? ‘From Narrow to General AI’ offers a perspective that many may share:
· After a while, and once the novelty wears off, you may begin to notice something else. The videos have an odd, hypnotic vibe. Watch them long enough and your eyes — and perhaps your mind — may start to glaze over.
· The clips are eye-candy to be sure, and their enjoyment comes from the visual delight of their crisp, colourful imagery. And just like candy, they can ultimately feel unsatisfying. They always seem like they are about to “deliver” something, but never do. Or to put it another way, they have no intention of delivering anything, the momentary experience is the point.
Lastly, more on AI at work. ‘Wired’ looks at worker burnout and how many are turning to AI for help. From the article:
· White-collar workers are so overwhelmed with emails, web chats, and meetings that they are using AI tools to get their jobs done — even if their companies haven’t trained them to do so, according to a work trend index published Wednesday by Microsoft and LinkedIn.
· Nearly 70 percent of people surveyed said they struggle with the pace and volume of their work, and nearly half feel burned out. Those using Microsoft 365 spend the majority of their workday communicating with other people in their company, and less time working in Word and PowerPoint — a larger problem that some AI tools seek to solve.
· The report also found that 46 percent of people want to quit their jobs this year.
