DOGE kills first omnibus bill, Zuck joins Elon vs OpenAI, Google AI comeback

Real-time transparency can kill multi-hundred-billion-dollar bills in hours.

Elon and Vivek used Twitter to spotlight a 1,500-page, $340 B omnibus; within 12 hours public outrage forced Congress to pull it.

Why: Social media collapses the information asymmetry that normally lets rushed legislation pass unread.

"This was a multi-hundred billion dollar grift that was stopped on a dime over 12 hours of tweets."

Zero-based budgeting exposes hidden bloat.

At Twitter, roles and SaaS tools that hadn’t been used in months were eliminated; similar audits across federal agencies could yield massive savings.

Why: Legacy budgeting assumes last year’s spend is sacred; zero-based forces justification from first principles.

"When we went into Twitter, nobody was coming to the office... they were paying for software to route people to the right desk in office suite."

AI model quality is approaching an asymptote, shifting competition to UX and cost.

Experts like Ilya Sutskever note diminishing returns from scale; differentiation now comes from memory, routing, and user experience layers.

Why: Training data is becoming commoditized; marginal gains require exponentially more compute.

"there's this terminal asymptote that we're seeing right now in model quality... the data is kind of static and brittle."

Market share is fragmenting as enterprises become model-agnostic.

OpenAI’s share fell from ~50 % to ~34 % while Anthropic doubled and Google gained, showing rapid commoditization.

Why: Cost-quality trade-offs vary by task; enterprises route prompts among 30-50 models via LLM routers.

"we are completely promiscuous in how we use models... you're going to rely on 30 or 40 or 50"

Stablecoins can undercut Visa/MasterCard’s 3 % tax on global GDP.

USDC and Tether already move billions for companies like SpaceX; regulated stablecoins could replace card rails and remittances.

Why: Stablecoins settle in real time with near-zero fees, forcing legacy networks to compete.

"the idea that you can take that 300 basis points you pay to these companies and crush it to zero would be a boon to global GDP"

Regulate stablecoins first, then tackle broader crypto markets.

Chamath advises Sacks to push US-regulated stablecoins and cheaper payment rails before addressing Bitcoin or DeFi.

Why: Stablecoins enjoy bipartisan utility and avoid the political quagmire of speculative tokens.

"I would tell Sacks, go to stable coins, then disrupt the Visa rails and then go to the other stuff."

Over-regulating AI early risks freezing innovation.

California’s SB 1047 imposed liability on model developers, potentially deterring Meta from releasing Llama updates.

Why: Fear of litigation shifts risk-reward against open-source and rapid iteration.

"does Zuck then want to release the next version of Llama if you're taking on that much risk"

State-by-state AI rules create a compliance nightmare.

Aaron Levie warns that 50 different state regimes would strangle national AI deployment.

Why: Fragmented regulation multiplies legal costs and slows rollouts.

"you probably don't want state by state legislation on this topic... you're in a world of hurt"