Brian Armstrong’s 14% workforce reduction at Coinbase came wrapped in an ‘AI-native’ vision, but Nathaniel Whittemore notes Robinhood just posted a 47% year-over-year crypto revenue crash. Axios was the only outlet that questioned the narrative - the rest swallowed it, letting AI gloss over bleak fundamentals.
Jason Calacanis sees a new playbook emerging. After Block cut 40% of its staff and mandated 100% AI adoption, it reported a 26% earnings beat. Code changes per engineer jumped 2.5x. The mandate is blunt: automate your own tasks or lose your seat.
"The current wave of tech layoffs is driven by AI adoption, creating a prisoner's dilemma where companies must cut costs to boost earnings and stay competitive."
- Jason Calacanis, This Week in Startups
Regulated sectors are opting out of the cloud entirely. David Moscatelli’s Go Abacus has 1,600 orders for its Go One, a $250,000 on-premise GPU box for banks and hospitals. The hardware sits behind the firewall, updated annually. Clients share model weights but never raw data.
Former CFPB Director Rohit Chopra warns the financial structure underpinning this boom is circular. Microsoft and Nvidia fund their own customers through equity and cloud deals, inflating valuations without traditional cash flow. If Anthropic or OpenAI miss revenue targets, the whole chain faces a correction.
"AI is undoubtedly changing work structures, but in this case, it serves as a useful shield against bleak fundamental performance."
- Nathaniel Whittemore, The AI Daily Brief
The entry-level path has snapped. A computer science graduate with a 4.0 GPA told Breaking Points he applied to over 100 jobs and landed one unpaid internship with 250 applicants. The work junior developers trained for - debugging, basic scripting - is now handled by the tools. Networking is the only reliable way in.
Meta bets $145 billion this year on consumer AI, but Jamie Dimon doubts the paid model for personal use. The industry’ real focus is enterprise: Anthropic’s revenue run rate surged to $45 billion, and it just committed $200 billion to Google Cloud over five years. Compute backlogs now total $2 trillion. Larry Fink says it will be traded like oil.
The human verification layer is the next scramble. Jose Caldera built a proof-of-human system on BitTensor, where miners attack the detection models to improve them. Calacanis predicts social feeds will soon offer a ‘human only’ filter. Without it, automated slop will overrun digital platforms.
