Degenerativ

„Hey team! Quick one before the weekend: leadership's super excited about AI.“

Sonst ein beinharter Gegner jeglicher Videoproduktionen mittels sogenannter generativer — doch tatsächlich eben nur extraktiver — K'I' und das natürlich insbesondere in der Darstellung von Lebewesen oder auch nur einer natürlichen Umgebung unseres Heimatplaneten, habe ich mich von dieser herrlichen Verballhornung des neuesten Aufgusses von Business-Bullshit all dieser geschäftstüchtigen Blutegel in der Softwareentwicklung nur allzu gerne begeistern lassen:

„The ratio of "supernumerary" developers over "exponential" ones has remarkably plateaued to 3x since 2010 (last chart). Since 2010 the software industry has been desperate to find a solution to break the ceiling. It was never to be found. To eat the World faster, the software industry needs more developers, and no technique or process has been able to deliver them so far. It is not in spite of many promises (e.g., remember low/no code, programming with little drawings?). We have reached the peak of all possible sacrifices in 2010.

Now 16 years after, AI is what people think is the promise that the solution is in sight. It isn't. The AI boom is the measure of 16 years of growing desperation, not the measure of AI success. Imagine what will happen if investors realize it is just more smoke and mirrors... Irrational hopes and dreams shattered.

Astonishingly you won't see any of these data presented anywhere else. It's the charts of the shame. In itself it tells a powerful story.“

Vincent Lextrait

The ratio of "supernumerary" developers over "exponential" ones has remarkably plateaued to 3x since 2010 (last chart). Since 2010 the software industry has been desperate to find a solution to break… | Vincent Lextrait

The ratio of "supernumerary" developers over "exponential" ones has remarkably plateaued to 3x since 2010 (last chart). Since 2010 the software industry has been desperate to find a solution to break the ceiling. It was never to be found. To eat the World faster, the software industry needs more developers, and no technique or process has been able to deliver them so far. It is not in spite of many promises (e.g., remember low/no code, programming with little drawings?). We have reached the peak of all possible sacrifices in 2010. Now 16 years after, AI is what people think is the promise that the solution is in sight. It isn't. The AI boom is the measure of 16 years of growing desperation, not the measure of AI success. Imagine what will happen if investors realize it is just more smoke and mirrors... Irrational hopes and dreams shattered. Astonishingly you won't see any of these data presented anywhere else. It's the charts of the shame. In itself it tells a powerful story. Medium articles in comment.

LinkedInVincent Lextrait

„These hot takes all think what these companies are doing is “building” and “engineering”. But that’s not what Big Tech has been for 20 years.

The Stock Is The Product

And these “growth companies” are shareholder value maximization constructs. Technology isn’t even secondary (that’s attention economy).

Wondering why technology hasn’t really improved anyone’s life despite gobbling up all investment? More money than ever spent on a single thing in the history of mankind? Your money, your pension funds, your future.

Trust me, I’m a techbro.“

Georg Zoeller

„At first, AI was a GPU problem.

Then it turned out that AI was a memory problem — we had too little memory supply for the available compute.

After the memory market, the next realization was that power and energy are the ultimate constraints on AI.

And now the market thinks AI is a capital problem.

Building AI factories (data centers) requires trillions of dollars in capital. But the founders and engineers of AI labs and AI infrastructure businesses thought they would be able to secure trillions of dollars in cheap debt from the market.

But borrowing costs for AI companies have risen, and now they are in double digits in the secondary market.

So the market thinks AI data center businesses won't be able to build data centers worth tens of trillions of dollars with such a high cost of capital.

They have left 2 choices:

They can dilute their shareholders, which will reduce earnings per share.
Or they can cut AI investment, which will reduce future returns for shareholders.

That's why the market is punishing AI businesses.

But read it again — this is what the market thinks.

And the market is run by finance people, not engineers and builders.
I trust builders more than finance guys.

AI will stand strong and fund itself through its revenue.
The demand for compute is enormous and rising rapidly.

The demand is there. The only thing left is to build.
And banks, financial institutions, and governments will come up with new financing plans.
The AI buildout will continue.

At best, they can spread fear, but that won't break the fundamental strength of AI.
See you in a couple of months.“

Marjanul Islam

These hot takes all think what these companies are doing is “building” and “engineering”. But that’s not what Big Tech has been for 20 years. The Stock Is The Product And these “growth companies”… | Georg Zoeller

These hot takes all think what these companies are doing is “building” and “engineering”. But that’s not what Big Tech has been for 20 years. The Stock Is The Product And these “growth companies” are shareholder value maximization constructs. Technology isn’t even secondary (that’s attention economy). Wondering why technology hasn’t really improved anyone’s life despite gobbling up all investment? More money than ever spent on a single thing in the history of mankind? Your money, your pension funds, your future. Trust me, I’m a techbro.

LinkedInGeorg Zoeller