Bridgewater's estimate of 18% US job displacement within five years is not a forecast—it's a code embedded in the current economic architecture. The signal is clear: labor markets are being recompiled. Andrew Yang's renewed AI tax push, aired on CNBC's Power Lunch, is a symptom of this shift, but the code doesn't lie about the underlying incentives. Yang wants to tax artificial intelligence instead of payroll, sending the revenue directly to workers as checks. He argues that firms skip payroll taxes and healthcare costs by choosing AI over new hires. The proposal is economically elegant, politically naive, and technically flawed. Signal over noise. Always.
Yang built his political brand on automation warnings during his 2020 campaign, proposing a universal basic income called the Freedom Dividend. He also backed cryptocurrency adoption and clearer digital asset rules. Now, as CEO of Noble Mobile, he's reviving the same thesis with a new twist: tax the machine, not the human. His comments echo a March appearance on Squawk Box, where he argued the government should stop taxing labor. The debate has drawn similar concerns from sitting US senators, and Yang explicitly referenced Anthropic CEO Dario Amodei's 2025 proposal for a 3% AI revenue tax. Amodei said the levy would apply each time a model generates revenue. Yang says the same logic should apply broadly, forcing firms to weigh AI costs against payroll costs.
The chart is a symptom, not the cause. The data backing Yang's push is stark. A CNBC and Generation Lab survey from August 13 polled Americans aged 18 to 34: 45% expect AI to hurt their careers, while only 10% expect it to help. Bridgewater Associates executives Greg Jensen and Nir Bar Dea published a New York Times opinion piece estimating 18% of current US jobs could be displaced within five years. They used that estimate to back their own AI token tax proposal. Meanwhile, the shift is already visible in customer service, which employs roughly 2.9 million Americans according to the Bureau of Labor Statistics. Yang argues retraining programs rarely work, pointing to failed efforts for coal miners and warehouse staff. He proposes sending tax revenue directly as checks.
But here's where the code gets interesting. Based on my experience auditing the 0x protocol in 2017—where I found a re-entrancy vulnerability in their token swap logic before launch—I learned that code execution is borderless. An AI tax on centralized corporate revenue is a fragile construct. The 3% revenue tax Amodei floated assumes models generate revenue through centralized APIs, like OpenAI's ChatGPT or Anthropic's Claude. But what about decentralized AI agents running on smart contracts? In the 2021 NFT cultural signal decryption that I published, I showed how digital assets attach to cultural signaling, not utility. Similarly, AI models will migrate to the most tax-efficient jurisdiction: the blockchain. Code doesn't lie.
Let's run the numbers. The US payroll tax rate is roughly 15.3% (employer and employee share). Yang's implicit AI tax, if modeled as a revenue tax, would need to be significantly lower to avoid killing innovation. A 3% tax on AI-generated revenue is a fraction of the payroll tax burden. But the displacement math is brutal: 18% of 160 million US jobs is 28.8 million workers. If each displaced worker requires $20,000 annually in UBI, that's $576 billion—roughly 2% of US GDP. A 3% AI revenue tax on, say, $1 trillion in AI-generated corporate revenue yields only $30 billion. The math doesn't close. Yang's proposal is a political signal, not a fiscal solution.
Sleep is for those who can afford to ignore the signal. The contrarian angle here is that Yang's push, while well-intentioned, assumes a centralized, trackable AI economy. But the future of AI is decentralized, open-source models running on blockchain networks like Bittensor or Gensyn. Taxing AI at the corporate level is like taxing the internet in 1995—it will miss the real disruption. The real risk is not job displacement but a new form of digital feudalism where AI owners capture value. A revenue tax might legitimize that. Instead, we should consider taxing the compute inputs (GPUs, energy) or the data used to train models. Those are measurable, borderless, and harder to evade. During the LUNA/UST crash in 2022, I traced the algorithmic failure in real-time, publishing a minute-by-minute forensic timeline. That experience taught me that fragile tethers break under stress. An AI revenue tax is a fragile tether.
Yang's retraining failure argument is valid. But sending checks without reforming the economic structure is like applying a band-aid to a code bug. The real innovation lies in creating new economic primitives: tokenized labor markets, decentralized autonomous organizations that distribute AI-generated value directly to contributors, and programmable income streams tied to compute ownership. The AI tax debate is a distraction from the structural shift. The next 12 months will see regulatory sandboxes where AI and crypto intersect. Watch for proposals that tax AI compute rather than revenue—that's the true signal. When the code runs on a decentralized network, who collects the tax?