Microsoft is pushing Copilot onto the market and the user community. There's even "Copilot for teachers" in schools. On the other hand, I have heard critical statements that question the whole AI hype and see – at least what Microsoft is doing – as a mistake. The boss of Salesforce sees Copilot as the "new Clippy", with which Satya Nadella and his staff are burning a lot of money. A nice topic that I'll take a look at.
Advertising
What if AI fails?
In the spirit of the article plot, I would like to return to some of the thoughts on AI here in the blog. You can look at the current development of large language models (LLMs) and be enthusiastic, uncritical or have reservations. However, there is another question that concerns me from an investor's perspective.
No, I don't hold any shares in OpenAI – just a few shares in Microsoft. But I have been asking myself for some time whether the money that is being burned can really ever be earned? And this raises the question of whether the protagonists of AI solutions are not fueling a gigantic bad investment. Because the typical hype cycle is unmistakable in development.
Over a year ago, in August 2023, I asked in the German article Was, wenn die generative KI sich als Blindgänger entpuppt?, the question "what if AI will fail?". As long as AI solutions solve real problems (e.g. in the evaluation of images of tissue in cancer screening in medicine, in data evaluation in science, etc.), there would be nothing wrong with them.
But the current situation seems to me to be "hello, we have a solution and are now frantically looking for a problem tosove". When I listen to what Microsoft's management praises as the advantages of their Copilot, I often think "what they are talking about, I never have had this problem".
There is serious talk of "booking your next trip with AI and ordering a table in a restaurant". Creating even better emails and documents, automatically taking meeting notes and so on are further arguments. Unfortunately, this is as far away from the reality of millions of people's lives. Sure, AI could do the above-mentioned tasks, but the cleaner, plumber, etc. don't deal with these problems on a daily basis.
A user once summed it up on reddit as "those in Microsoft management are always trying to solve problems that none of us in the field of IT or production have – a bunch of absolute garbage with no idea of anything".
Advertising
Are the AI investments paying off?
And that brings us to "what if the whole thing flops". In this context, I found it interesting what analysts from investment banks think about this topic. At the end of June 2024, Goldman-Sachs published the analysis Gen AI: too much spend, too little benefit?
The analysts' message: the promise of generative AI technology to transform businesses, industries and societies is prompting tech giants and other companies to spend an estimated USD 1 billion on investments in the coming years, including significant investments in data centers, chips, other AI infrastructure and the power grid.
The conclusion is sobering: this expenditure has hardly paid off so far. And the question arises as to whether this high expenditure will ever pay off in the form of AI benefits and returns. And there is also the question of what impact this will have on economies, companies and markets – or not. The Goldman Sachs report is hard work for financial analysts – but quite revealing. On medium, someone summarized it under the title Goldman Sachs: AI Is Overhyped, Wildly Expensive, and Unreliable.
And Institutional Investor soberly quotes the question from the analysis of Jim Covello, Head of Global Equity Research at Goldman, "whether the $1 trillion that is likely to be spent on AI over the next few years will yield an adequate return". "What trillion-dollar problem will AI solve?" the analyst asked, noting that "replacing low-wage jobs with hugely expensive technology is basically the exact opposite of the previous technological shifts he's seen in the thirty years he's followed the tech industry closely."
Covello's conclusion: to justify the extraordinarily high costs, "AI needs to be able to solve complex problems, which it is not designed to do". So it could be that the beautiful house of cards that is AI will soon collapse and the big disillusionment will come.
Copilot: Microsofts Clippy 2.0
What does all this mean for Microsoft and its co-pilot? Sometimes the pieces of the puzzle fall into place and suddenly fit into the picture. Recently is stimbled on the statements made by Salesforce CEO Mark Benioff. Salesforce itself is active in the field of artificial intelligence and is therefore Microsoft's competitor.
But Benioff should know what the context is in this context. And Benioff speaks plainly in the Rapid Response Podcast (YouTube). It's a long 25 minutes, but in addition Windows Central has compiled the content.
The Salesforce CEO speaks of the "enormous disservice" that Satya Nadella and Microsoft have done to the AI industry with their Copilot approach. His harsh statement: "Copilot is just the new Microsoft Clippy. It doesn't work and delivers no value."
Benioff assumes that Microsoft's Copilot AI will not last long because the functionality is disappointing and offers no added value. One of his statements reads: "If you look at how Copilot has been sold to customers, it's disappointing. It doesn't work. It spits data at our feet, but it doesn't deliver value. I have yet to find a customer who has worked successfully with Copilot. Copilot is just the new Microsoft Clippy".
You can be of the opinion that Copilot helps in your personal environment – I don't deny that. However, my personal experience is rather sobering and negative.
Advertising
