AI4U: Getting AI from Kindergarten to University

What's Up in Artificial Intelligence—and What Are The Real Risks? For Investors and Users.

Thanks to 惠普企业服务 and 英伟达 , I had the great pleasure to learn first-hand about "Building Self-Hosted Generative AI Solutions at Scale".

Their team, including Rebecca Dal Canton Cyrill Hug Jordan Nanos et. al., did a great job in presenting key essentials, providing very practical insights and hosting the #Swiss AI community. Apropos community: Did you sign-up already with "Connect AI" chief-edited by Matthias Zwingli ? A very professional newsletter about this topic!

Some stunning facts upfront about 惠普企业服务 (HPE US Equity): Largest supercomputer operations worldwide; top 1% globally in supplying supercomputing data centers, within one week; HPE can build a prototype AI application. One overseen fact about 英伟达 (US Equity): 60% of its employees are software engineers, only 40% of the staff is focused on hardware–for what Nvidia is very well-known for.

?

1.???? WHERE STANDS AI TODAY? THE FACTS AT A GLANCE:

• AI training [models, data] are still the biggest demand element in the whole AI space. An estimated 80% of all AI sales globally goes to training.
• Data sovereignty as key topic amongst all corporate clients.
• Public cloud AI applications will drive data storage and server business drastically.
• Energy efficiency and #sustainable data centers are for the whole sector incl. HPE an important topic; HPE has one of the world's most energy [and water] efficient data centers installed.
• The AI hype is still at full speed; exchange-listed companies, which CFOs and CEOs are introducing “AI-powered solutions” during analyst/investor calls, may see their stocks trading 6% higher the next day.
• AI solution provider will see significant new business over the next 2-4 years; The compound annual growth rates (CAGR) in AI will remain on double-digits levels.
• Big opportunities to embedd smart AI solutions in corporate #sustainabilty and #ESG reporting–but way to early for large roll-outs due to lack of model training.

The sober fact now: Despite all buzz about AI worldwide, no significant (!) AI use-case is live as of today. Most companies are still experimenting in various ways. Honestly, creating artificial pictures (such as https://pixlr.com etc.) for all sorts of genre is not really the next big thing but a funny element. More helpful are AI applications in professional media (for artists and producers looking to take their creativity to the next level), such as https://www.beatoven.ai/ , https://soundful.com/ or etc.

It will take about two to three years, until AI applications from today's Kindergarten level arrive at "high school", not saying yet at university. However, right now the speed is at highest level to reach advanced learning levels.

2.???? WHAT ARE THE REAL RISKS – FOR INVESTORS AND ALL OF US AS SOCIETY?

As I’m grown-up with the 1999 Dot-Com Bubble and the “Neue Markt” euphoria, it is relatively easy to identify some parallels from a business and investment perspective. Without doubts, large players like Alphabet, Facebook and Microsoft will set certain unreversible trends in AI and create a mainstream feeling for using AI-powered tools in our daily [office] live.

Also, AI infrastructure players like Nvidia, ASML, Hewelt Package Enterprise Co., Cisco etc. will benefit from the need of the whole world to develop, host and run [somewhere] their AI adoptions.

“An AI assistant for everything? What ever limits our ability and willingness to use of brainpower permanently, paves the way to end as stupid, uneducated follower of everything.”

However, the biggest risks are way bigger than just facing dropping stock prices in overhyped AI companies.

AI is by no means new. Visionary computer scientists at Dartmouth College in New Hampshire laid the foundations back in 1956. Recently, it has been refined further and further. The advantages of computer systems taking over repetitive tasks - instead of moderately trained humans - cannot be denied. However, it is also true that AI does not work as if by magic but has to be painstakingly programmed by hundreds of thousands of IT employees between California, Krakow and Calcutta. Nothing works alone, not even with AI. The idea that entire books will appear out of nowhere overnight or that a car will take off on its own and drive its drunken owner home safely is frankly still fiction.

The more time passes, the more obvious the limitations become. These are unchic and are therefore deliberately concealed by management consultants and AI entrepreneurs. AI now delivers first-class content for simple simulations of text completion or "draw me a picture of the Matterhorn", but as soon as things get complex and strategic, intuitive elements are required, AI quickly runs out of steam. This is not a shortcoming, but an important stop sign so far. Humanity is not a bad thing, on the contrary. The total automation of almost all activities by computer systems quickly tips over into a kind of AI fascism: only a central management system decides what is right and wrong, individuality in solutions and humanity are considered obscure and degenerate.

What quickly becomes annoying today as a chatbot at insurance or telecommunications companies will become the new standard for customer inquiries of all kinds in the future? In the future, will your own voice no longer be a guarantee for flawless identification with banks or authorities because it can be falsified by AI? Even today, it is no longer possible to distinguish between fake news and serious journalism without any doubt - thanks in part to the misuse of AI. And social media channels can no longer really distinguish between robo-users and real people, let alone between real images in posts and artificial Photoshop fakes. The gigantic list of AI risks is getting longer every day. In a most recent survey amongst about 3,000 adults across Switzerland, 40% stated that they associate high or very risks with artificial intelligence and their applications. That's a word.

For investors, this results in a clear strategy for action: invest in

a) companies that actively monitor, minimize or even eliminate AI risks.

b) all those companies that position themselves as highly profitable AI infrastructure providers.

c) stay away from XXL valued AI [later stage] scale-up and carefully select AI start-ups.

Finally: For the next “AI afterwork party”, remember this name: NVIDIA GH200 Grace Hopper Superchip — it’s the World’s most powerful and versatile AI chip currently. Name it, and you are dubbed “AI pro”. ??

PS: This article has not been written the help of AI. By special purpose.

All information provided by this article is for information purposes only and is not, and does not constitute or intend to constitute, investment advice or any investment service as referred to in the Act on Financial Supervision. Such information also is not and should not be deemed to be an offer to purchase or sell or a solicitation of an offer to purchase or sell, or a recommendation to purchase or sell any securities or other financial instruments. The content on this website is based on sources that are considered reliable. No guarantee is provided on its accuracy, correctness or completeness either express or implied.