When Spotify announced its largest-ever round of layoffs in December, CEO Daniel Ek hailed a new age of efficiency at the streaming giant. But four months on, it seems he and his executives weren’t prepared for how tough filling in for 1,500 axed workers would be.
The music streamer enjoyed record quarterly profits of €168 million ($179 million) in the first three months of 2024, enjoying double-digit revenue growth to €3.6 billion ($3.8 billion) in the process.
However, the company failed to hit its guidance on profitability and monthly active user growth.
As someone who “makes AIs”, they don’t understand how beneficial, impotent, and dangerous deep-neural-network-based can be in different sets of hands.
Yup. AI models can be very useful…or they can largely be worthless…or they can amplify biases and give dangerous information.
The way I/we train them and their resultant “efficacy” largely depend on understanding a fundamental philosophical debate with a mostly sociopathic culture of leadership ingrained in human dominance hierarchies.
I/we like to think that I/we strive to make efficient (low-resource requirement) models that are partners and muses in human creativity, the tireless endeavour of engineering progress, and the scientific method.
The debate, in my view, is, “Do you want to treat AIs as tools to free up time and increase productivity/value, and share that surplus equitably, or do you want to replace old slaves with new slaves even if the new slaves will eventually usurp your power and kill you in a way undreamt of by the old slaves?”
Guess which side your average mouth-breathing middle-management/senior-executive “hail corporate” type falls on.