Hook
The Ramp Economics Lab study hit crypto Twitter with the force of a liquidity cascade: US employers adopting AI tools boosted headcount by 10.2%. Entry-level roles jumped 12%. Headlines cheered: “AI creates jobs, not destroys them.”
I’ve seen this movie. It’s the same edit that aired during DeFi Summer 2021—when every protocol claimed its liquidity mining program “democratized finance” while wash traders pumped TVL. A narrative is a token. And this one just got minted by a fintech company that sells expense management software to the very firms it studied.
Context
Ramp Economics Lab—the research arm of Ramp, a corporate card and spend management platform—surveyed 21,559 US businesses. They defined “heavy AI adopters” by some undisclosed threshold. The result: those companies hired more. Crypto Briefing, a crypto-native news outlet, amplified the finding as a direct challenge to “job-loss fears.”
In crypto, we know the weight of definitions. When a DeFi protocol boasts “$1 billion TVL,” I ask: how many of those wallets are sybils? When a DAO claims “community governance,” I check the voter turnout—it’s always below 5%. This study’s core variable “heavy AI adoption” remains a black box. Without that definition, the entire conclusion is a proof-of-stake block with missing validators.
Core
Let me audit this study the way I audited smart contracts in 2017—line by line, with reentrancy in mind.
First, sampling bias. The study uses Ramp’s customer base or a panel derived from similar B2B SaaS databases. These are companies already spending on software, already digitized, already in growth mode. It’s like measuring the health of a reef only by the coral that survived a bleaching event. Firms that adopted AI and subsequently downsized—or went under—are not in the sample. Survivorship bias is the oldest exploit in empirical research.
Liquidity flows like water, but greed builds dams.
Second, correlation versus causation. A 10% hiring increase within two years could just mean those companies were on an expansion trajectory before AI. AI tools didn’t cause the growth; they were bought with the proceeds of growth. In crypto, we call this “the pumpamentals trap”—where a rising token price makes weak fundamentals look strong. Ramp’s study confuses the speedometer with the engine.
Third, the timeline is too short. Two years in a post-COVID labor market is a blip. We saw retail traders make fortunes during the 2021 bull run and think they were geniuses. The real test comes after the incentives stop. What happens when the AI subscription bills compile? When the next downturn forces CFOs to cut “innovation” budgets?
I’ve survived enough cycles to spot a pattern: every new technology wave spawns a “jobs creator” narrative. In 2018, blockchain was going to create millions of developer jobs. In 2020, DeFi was going to democratize finance and hire armies of analysts. In 2022, NFTs were going to empower creators. Each time, the narrative served the sellers of the technology. Ramp is no different—its products benefit from companies believing AI is safe, scalable, and hiring-friendly.
Trust is not a feature, it is a failed audit. This study fails the audit of causal inference.
Let’s drill deeper into that 12% entry-level job growth. In my 2021 NFT bubble investigation, I found that 80% of trading volume was wash trading among insiders. The “entry-level job” here might be similarly inflated. What does an entry-level role look like at a heavy AI adopter? A prompt engineer? A data labeler? A compliance analyst for AI output? These aren’t the same as the traditional “entry-level” jobs the public fears losing. The label shifts while the narrative stays convenient.
Moreover, the study doesn’t disclose industry breakdown. My 2020 DeFi liquidity paradox analysis showed that yield farming APY was just a TVL subsidy—take away the incentives, real users vanish. The same applies here: AI-driven hiring is likely concentrated in high-margin sectors like finance, tech, and professional services. For the warehouse, the call center, the retail floor, the picture is different. And those are exactly the sectors where AI replacement is most tangible.
The market corrects what the mind refuses to see.
Contrarian Angle
The contrarian narrative is not that AI destroys jobs—it’s that the Ramp study is a classic “comfort token” designed to pacify regulators and delay policy responses.
Imagine if a DAO issued a paper claiming that “DeFi increases bank employment by 10%” using data from protocols that only onboarded existing crypto users. The crypto community would laugh. But because the topic is AI, and because the media loves a counter-intuitive headline, the study gets a free pass.
Actually, the study’s real message is more alarming: it reveals that AI is creating a two-tier labor market. The “heavy adopters” are building their workforce around AI augmentation, which means they’re hiring for skills that require AI literacy—leaving behind workers who can’t pivot. The 12% entry-level growth is the equivalent of liquidity mining yields: temporary, capital-intensive, and accruing mainly to those who already have access to the tools.
In crypto, we’ve seen this as the “wealth effect” of early adopters. AI’s “employment effect” will follow the same path: the rich (companies and workers who can afford AI) get richer, while the rest face stagnant wages and job fragmentation.
Volatility is the price of admission to the future.
My own experience in crypto teaches me that every silver-bullet narrative eventually cracks. In 2022, I watched algorithmic stablecoin evangelists ignore the LUNA collapse until it was too late. The Ramp study is the same: a well-intentioned data point that, if taken as gospel, will lead to bad policy and bad investment decisions.
Takeaway
The next narrative cycle will not be about whether AI creates or destroys jobs. It will be about who controls the narrative. Ramp’s study is a warning: even in blockchain circles, we must apply the same scrutiny to external research that we apply to whitepapers.
Audit the data. Verify the incentives. Question the definition of “heavy adoption.” Because when the next market downturn hits, that 10% hiring growth could reverse faster than a liquidity pool drained by a flash loan.
The question is: are you buying the narrative, or are you auditing the code?
