Over the course of this book, I’ve shown over and over again how Silicon Valley can make money from manipulating its legal environment, rather than from the superiority of its technology. This is something that’s very hard to unsee once you’ve seen it – and once you’ve seen it, you can’t help approaching Silicon Vally’s latest overhyped techno-solutions with a degree of skepticism. I hope that after reading this book, you’ve also come to look at Silicon Valley through skeptical-colored glasses (or maybe you started reading this book because you were already there).
Silicon Valley invests obscene amounts of money in trying to stop people from putting on these skeptical glasses – it would prefer that you just swallow its hype and put on the latest AI-powered lenses instead. But by jumping the shark with crypto and other fintech business models – and increasingly with GenAI – the tech industry may be forfeiting one of its greatest assets: the widely shared perception that its tech will deliver. If that perception shatters amidst increasing overengineering and enshittification of its output, Silicon Valley won’t be in a position to distract us from developing real, public-minded fixes for our financial system and other big problems.
Let me pause for a second and acknowledge that, here in the year 2025, the idea that we will see any big, public-minded fixes in America seems laughable. Instead, we’re seeing unprecedented dismantling of legal doctrines and regulatory agencies that were supposed to protect the public from harm – and many of these steps seem designed to benefit the very Silicon Valley elites that I’ve argued need to be marginalized. But if we get out of the present moment alive, we’ll find ourselves with an opportunity to rebuild. I have sometimes participated in academic workshops where we imagined we had a magic wand and could design good policy without worrying about past practices or political feasibility. We’ll never have that magic wand, but now that a flamethrower has been taken to the United States government, we may end up having something approaching a blank slate to work with. Some members of the Silicon Valley elite and their boosters are actively pushing an “abundance agenda” as a guide for how to build on that blank slate, but their attempts to get people on board with refried techno-solutionism should be met with scorn. Friends don’t let friends get abundance pilled.
As the techlash against Silicon Valley gathers steam (particularly among those whose political persuasions lie left of center), Ezra Klein and Derek Thompson’s Abundance book can be thought of as an attempt to rehabilitate techno-solutionism and TESCREAL ideologies to make them more palatable for left-leaning voters. The book’s introduction dreams of TESCREAList “star pills” manufactured in space factories that can save “millions of lives and billions of healthy years,” and of benevolently deployed AI tools that bring about shared prosperity and more leisure time for all. If this introduction is anything to go by, Klein and Thompson see the “abundance agenda” as the path to utopia: all we need to do is dispense with pesky regulations that might impede the innovation and building that will generate this utopia. But if I learned anything from studying utopias and dystopias in English literature (to this day, one of the most important classes I ever took), it’s that there’s no such thing as a utopia that everyone will agree is utopian.
The NEOM city being built in Saudi Arabia, for example, sounds very abundance-y in its aspirations to be a futuristic tech hub that is a beacon for job creation and sustainability. NEOM’s website even includes Silicon Valley’s favorite cliché: “we want to create a legacy that makes the world a better place.” But Saudi Arabia is one of the most water-stressed nations on earth and concerns have been raised that AI data centers may result in regional water scarcity. More than 21,000 workers (primarily from India, Bangladesh, and Nepal) are reported to have died working on NEOM and related projects in Saudi Arabia since 2017, with more than 20,000 indigenous people reported to have been forcibly displaced to make room for the development. Sandeep Vaheesan notes in his review of Klein and Thompson’s Abundance book that “instead of calling for steeper progressive taxation and anti-monopoly policies that would rein in the power of the affluent, Klein and Thompson focus single-mindedly on red tape.” Is NEOM the kind of “utopia” we have to look forward to in the United States if Silicon Valley is liberated from the red tape of democratic accountability and allowed to build unchecked? If so, things could get ugly. Writing about a period spanning both the industrial revolution and the fascism of the 1930s, Karl Polanyi described the social instability that arose when the excesses of capitalism were unchecked by laws designed to smooth and moderate the impact of market and technological developments.
The abundance agenda doesn’t have much to say about smoothing or moderating the impact of Silicon Valley’s technological developments – Klein and Thompson argue that “the American innovation system would benefit from trusting individuals more and bureaucracies less.” But we know from Chapter 7 who some of these individuals really are and what they value, and by now I hope I’ve convinced you of the dangers of trusting the unaccountable Silicon Valley elite to diagnose and “solve” social problems. Despite the derision Klein and Thompson’s express for trickle-down economics in their book, the abundance agenda is based on equally facile assumptions that unleashing technological innovation will automatically trickle-down benefits for the public.
Klein and Thompson’s fellow abundist Matt Yglesias took issue with some of my writing on this subject, and decided to write a critique focusing on how much he likes Uber. Uber followed the blitzscaling playbook of flouting laws on the books to grow its business; Yglesias said he didn’t want to get into my arguments about fintech companies doing the same thing. I tried to push him, though, on what it would mean if every tech business used Uber’s playbook: when I asked him “why not engage with "what happens when regulatory arbitrage isn't good?" Is it because the abundance agenda falls apart then?,” I got crickets in response.
Because Yglesias wouldn’t, let me answer my own question. Ultimately, turning a blind eye to legal violations or changing the law to accommodate new tech businesses allows the Silicon Valley elites to amass even more political power – which they can then deploy to further undermine regulations designed to protect people with less power, as well as to undermine tax and antitrust laws that might prevent them from amassing even more political power. This is a project pursued by Silicon Valley elites of all political stripes. I mentioned the tax-and-antitrust-inspired rightward turn of some members of the Silicon Valley elite earlier in the book, but Democrat super donor and PayPal mafioso Reid Hoffman petitioned then-Presidential candidate Kamala Harris to boot aggressive antitrust enforcer Lina Kahn from her role as Chair of the Federal Trade Commission. Mark Cuban, another Democrat mega donor, is also an abundance-style techno-solutionist at heart, and has been known to hang out on Marc Andreessen’s radicalizing group chats.
Cuban made his billions by putting radio on the internet (I know I keep doing callbacks to the HBO show Silicon Valley, but the Russ Hanneman character widely understood to be spoofing Mark Cuban is HILARIOUS). Cuban used a lot of the money he made to push crypto – he even speaks approvingly in the God Bless Bitcoin documentary I suffered through on your behalf in Chapter 6 – and now he blames the Biden administration for driving voters into Trump’s arms by enforcing the law against the crypto industry. Cuban claims that if Biden had been more accommodating of crypto, “young men would have seen their networths skyrocket, and voted for Harris. [Biden-era SEC Chair] Garry Gensler is why Harris lost young men and the election.”
In my humble opinion, Cuban is framing this in a pretty “what did you do to make them hit you?” kind of way. His premise that young male crypto investors cost Harris the election is also highly questionable. Only a small percentage of Americans own any crypto, and it’s a big leap to assume that the ones who do are single issue crypto voters. If crypto policy is what cost Harris the election, it probably has a lot more to do with the crypto industry’s political spending than with single issue crypto voters (remember, all that crypto money was spent on political ads that didn’t even mention crypto). So think through the implications of what Cuban is saying here: his message is that enforcing existing laws against powerful tech industries is a political loser, so policymakers should unilaterally disarm against Silicon Valley so as to not anger the tech elites. That’s the abundance agenda in a nutshell: just let Silicon Valley do what it wants and hope that benefits will trickle down to everyone else. To which I respond, “no thank you.”
Improving public welfare should be pursued intentionally and directly, not treated as a hoped-for side-effect. That’s how the public sector can hone its capacity to build and innovate in the public interest, and how to convince people that democratically-elected governments can be a force for good. If our goal is indeed to make the world a better place, it’s time to start figuring out solutions that don’t run through Silicon Valley.
My experience with crypto has made me highly skeptical of any solution that emerges from Silicon Valley’s VC-industrial complex. For a while, I wondered if I was being unfair; if perhaps crypto should be treated as an anomaly that shouldn’t besmirch the whole venture capital model. But the more I’ve learned about Silicon Valley’s VC model (and about the rise of GenAI, the latest technology being relentlessly pushed by Silicon Valley VCs), the more I’ve come to accept that besmirching is entirely appropriate. And I’m not alone in coming to this conclusion. As entrepreneur and Stanford lecturer Steve Blank put it: “I’ve watched the industry become a money-hungry mob. V.C.s today aren’t interested in the public good. They’re not interested in anything except optimizing their own profits and chasing the herd, and so they waste billions of dollars that could have gone to innovation that actually helps people.”
In her book The Entrepreneurial State, economist Mariana Mazzucato explains why the public sector is better positioned to invest in the kinds of risky and capital-intensive science and technology projects that could actually help people. The commercial applications of these kinds of projects are extremely unpredictable – sometimes taking decades to emerge if they emerge at all – and the private sector just doesn’t want to get involved until the commercial applications come into focus. The public sector is better positioned to “think big,” to solve long-term problems whose solutions sound in democratic values rather than efficiencies or profits, and the public investment and research we need could be at least partially funded by ripping away some of the Silicon Valley subsidies we discussed in the previous chapter.
This would leave Silicon Valley VCs and the companies they fund to sink or swim on their own merits. While I’m sure we’d hear some histrionics from Silicon Valley about what cancelling those subsidies would do to innovation, the “innovation” we’re getting in return for those subsidies right now is frankly a pretty shitty return on investment. When we look at the billions upon billions of dollars being blown on things like the Metaverse and GenAI, it’s hard to escape the sense that cutting off those subsidies would be good policy – not only could that funding be diverted to finding public-minded solutions, we might also get a leaner, meaner tech industry better at identifying problems it is well-suited to solving.
Law professor Peter Lee has identified three interlocking structural features that explain why VC-funded innovation is so often underwhelming these days. First, VCs tend to mostly fund people drawn from their own in-group. The result is that a particular perspective, a particular worldview (and hooboy we saw in Chapter 7 how weird that worldview can be), is perpetuated in most of the technological solutions that get funded this way. One 2025 publication from VC data provider PitchBook led with the byline “the saying that “VC is a relationship game” now has the data to prove it”; after crunching their proprietary data, PitchBook found that much of the value added by VCs lay “in their networks: connections to a founder peer network, potential customers, and even investors that can lead subsequent rounds.”
These networks are notoriously comprised of privileged white men. One survey found that 80% of professional VCs are male, and those VCs tend to fund other men. According to Pitchbook, female-founded businesses have never received more than 2.8% of all VC funded capital in any given year. Even where female founders have male co-founders, they are less likely to attract capital: in 2023, the best year so far for gender parity in VC funding, all-male founder teams still received more than 75% of all VC funded capital.
Researchers looking at Pitchbook data have also found that “only 3.40% of startups in Pitchbook have at least one Black founder,” and that “Black-owned businesses raise about one-third as much venture capital as other startups formed in the same year, in similar industries, and in similar locations.” Part of the reason for this discrepancy? The researchers found that “Black founders are less likely to have worked at the same companies, or to have attended the same schools, as investors who might have funded their startups [and] Black startups tend to be located outside of private-equity hubs where venture capital is easily accessed.” To be sure, most white men are also locked out of these networks – according to one survey, nearly half of all professional VCs went to either Harvard or Stanford, and those community ties are critical to both building a rolodex and learning the styles and stories that VCs find compelling.
A notable exception to this elite education preference is Daniel Penny, who studied architecture at New York City College of Technology. Penny was hired by Andreessen Horowitz after rising to national prominence in connection with the death of Jordan Neely, the Black homeless man Penny held in a chokehold until Neely became motionless. Make of that what you will.
In short, the demographics of both the VC industry and the recipients of VC funding look a lot like the demographics of the PayPal Mafia we met earlier in the book: mostly white men with an elite education. This insularity contributes to Professor Lee’s obstacle to innovation #2, groupthink. VCs often display a kind of herd mentality when picking startups, rushing to fund the same kinds of businesses at the same time. This groupthink sharpens the dangers of techno-solutionism – VCs are only interested in <insert hot business model> right now, so if you’re a startup, you had better find a way to solve your problem with <insert hot business model> – and trendsetter VCs like Andreessen Horowitz often have a lot of sway in establishing what <insert hot business model> is at any given moment.
There’s also VC groupthink around the idea that crazy charismatic founders are the ones to back – as the website for Peter Thiel’s Founders Fund states, they’re looking for founders who “have a near-messianic attitude and believe their company is essential to making the world a better place.” That, to me, looks like a wanted ad for con men with a god complex – this preference probably helps explain how VCs keep funding problematic founders like FTX’s Sam Bankman-Fried, Theranos’ Elizabeth Holmes, WeWork’s Adam Neumann, and fintech middleman Synapse’s Sankaet Pathak (who isn’t as well-known as the others, but we met him in Chapter 3). After Synapse collapsed, United States Senators demanded to know why venture investors like Andreessen Horowitz hadn’t insisted on adequate controls to protect consumers. I suspect part of the answer is that the VCs had collectively decided that Pathak was a messiah-genius, and didn’t want to upset him.
Because of the expectation that most startups will fail, including a stinker in their portfolio isn’t really a problem from a VC’s perspective. The real problem is missing out on including a winner. As a result, VCs have little incentive to do much digging into the credibility of the founders and business they’re investing in (thanks to limited liability, should a startup hurt someone or break the law, the most the VCs can lose is the amount they invested – which they always had a good chance of losing anyway). It’s so much quicker and easier for VCs to just follow the herd, and looking too closely for reasons not to invest can be bad for business. As one venture capitalist put it, “It’s a clubby industry…You need other V.C.s to like you, because they refer you into deals. If you get a reputation as a complainer, it can really hurt your business.”
The final constraint on VC’s innovative capacity identified by Professor Lee is that the VC fund structure doesn’t leave much time for investments to pay off. Although investors typically commit their money to a VC fund for ten years, it takes time for VCs to identify the startups they want to invest in and then to “exit” (i.e. sell those investments for cash they can distribute to the fund’s investors). Typically, that leaves about five or six years for businesses in the VC portfolio to blitzscale into something that others will want to take off the VC fund’s hands. There’s a very narrow universe of businesses that can grow so quickly – and they aren’t the ones building breakthrough new technologies in fields like clean energy and pharmaceuticals. Instead, VCs often favor businesses that focus entirely on software and don’t require any physical prototypes.
And yet, Silicon Valley’s existing venture capital model is widely viewed as a winning formula for spurring innovation in the public interest: as the historian Margaret O’Mara tells it, “hundreds of places around the world have rebranded themselves Silicon Deserts, Forests, Roundabouts, Steppes, and Wadis,” trying to emulate the secret sauce of Silicon Valley’s VC industry. But maybe they should curb their enthusiasm, unless they aspire to churning out faddish and unprofitable businesses insulated from real competitive pressures by legal dispensations and subsidized funding.
Last chapter, we saw the degree to which Silicon Valley’s tech platforms and VCs benefit from laws made and applied in their favor. This is, in many ways, an old story. Some scholars attribute the success of Britain’s Industrial Revolution (judged from the perspective of the winners of the Industrial Revolution, of course, not the Luddites) to new legal innovations. British innovators weren’t necessarily more educated or scientifically advanced than innovators in other countries, but they did benefit from newly-created intellectual property laws and business structures that limited the liability of investors. Today’s tech businesses continue to benefit from patent and trade secrecy protections, and from limited liability – meaning the investors in those businesses can’t lose more money than they invest, even if those businesses get sued big time or break the law and get whomped with fines.
Like intellectual property protections, limited liability has become part of the furniture but is not a law of nature. Instead, it’s a legislative grant given by individual states to encourage businesses to take entrepreneurial risks (California, for example, didn’t grant limited liability until 1932). But what the law giveth, the law can also taketh away. That means that limited liability could theoretically be taken away or circumscribed if a state legislature decided that certain entrepreneurial risks weren’t serving the state well. Now, an individual state probably wouldn’t do this in practice (it would be too afraid that its businesses would decamp to a state with more robust liability protections). But exceptions to limited liability could perhaps be implemented nation-wide by making those exceptions a precondition for certain benefits conferred by federal law. If VC firms and their investors could be held personally liable for fines associated with startups breaking the law, then they might not be so keen on regulatory arbitrage strategies...
State laws also permit corporations to issue classes of stock that give different shareholders different kinds of voting rights, and tech companies have increasingly embraced these since Google’s 2004 IPO. Issuing different classes of stock allows the founders of tech companies to retain control even after they’ve sold most of the stock to the public, which means those founders can continue to implement their vision (weird or beating-a-dead-horse-y as that vision might be) without much shareholder interference. I suspect it’s easier for Mark Zuckerberg to blow $46 billion on the Metaverse, for example, when he doesn’t really have to answer to his shareholders. But again, permission to issue different classes of stock is a gift from the state of incorporation and if a state were so inclined, it could prohibit it and require “one share, one vote” (stock exchanges could also mandate this requirement through their listing standards, as the New York Stock Exchange did from 1940 until the mid 1980s). These kinds of legal changes would force tech founders to keep more of their money tied up in their companies if they wanted to maintain control; cashing out would mean giving more power to shareholders. While this probably wouldn’t be enough to make the tech platforms into good corporate citizens, it could act as something of a moderating force (if we really want to make tech platforms into good corporate citizens then robust antitrust enforcement – backed by a credible threat of breaking up tech monopolies – is probably the way).
Then there are the rules that govern the investment of retirement funds. The VC industry lobbied heavily for, and benefitted enormously from, changes to the investment guidelines under the Employee Retirement Income Security Act (“ERISA”) that allowed pension funds to start investing in VC funds, notwithstanding that these can be very risky investments. Historian M.R. Sauter observes that pension fund investments in venture capital provided only 15% of the capital raised by those funds in 1978, but by 1988 (after the changes to the ERISA guidelines), pension funds were providing 46% of their capital. Changes to ERISA and the securities laws are now being contemplated that would allow individuals to invest in VC funds through their 401(k) plans, which would presumably result in even more nest eggs being placed in the VC basket.
Speaking of 401(k) plans, we saw last chapter that the Trump administration is encouraging the inclusion of crypto in these plans, and there are also legislative efforts afoot at the state level to encourage pension funds to invest in crypto. These are uniformly terrible developments – not only do they increase the likelihood that people’s retirement savings will be exposed to all the volatility and Ponzi-ness that characterize the crypto markets, they also bring crypto closer to the heart of our financial system, making that system more vulnerable to a crisis. But the important thing to note here is that these pushes are being effected through legal changes – meaning they can also be reversed by law. The same is true of the guidelines that allow pension funds to invest in VC funds. If the VC industry no longer had access to retirement funds as a source of capital, then it would have less to spend and would (hopefully) be a little more thoughtful about its investments. At any rate, we wouldn’t be subsidizing the next Juicero with our hard-earned retirement funds.
Perhaps most importantly, though, there’s tax law. There are many proposals floating around for taxing the ultra-wealthy, and I am here for them. These proposals aren’t tech specific, but they would hit many of the Silicon Valley elite – and with less money, those people would be less able to push the TESCREAList and anti-democratic priorities we discussed in Chapter 7. We should also resurrect the Inflation Reduction Act’s attempt to close the carried interest loophole, and tax VC funds’ profits as income – or at the very least, raise the capital gains taxation rate. That lower capital gains taxation rate is something else that the VC industry lobbied very hard for back in the 1970s and 80s and without it, VC wouldn’t be what it is today. And with less money behind it, the VC industry’s efforts to lobby for beneficial legislation and sweet-talk regulators would presumably be less effective in procuring the bespoke legal treatment that many mediocre and downright harmful Silicon Vally tech businesses rely upon to survive.
Many of Silicon Valley’s executives and venture capitalists made their money at the birth of the internet, or the dawn of the smartphone age, when growth opportunities seemed limitless. But the internet-enabled tech industry is in many ways a mature one now, and the low-hanging fruit have mostly been picked. As tech blogger Chris Martin puts it, “Maybe most of the critical things that can be created by one guy typing furiously are gone, and the opportunities that remain require expertise and wisdom from a bunch of different people.”
Sometimes when I think about the Silicon Valley elite, I see parallels to Sunset Boulevard’s aging film star Norma Desmond, living an increasingly deranged life in her secluded mansion and refusing to believe that her fame is over. Desmond ends up killing the man who spoils her illusions – in the absence of any legal constraints, what lengths will the Silicon Valley elite go to cling to their moment in the sun, back when they could get exponential growth from “putting <insert thing in the real world> on the internet”? And what might ordinary people do in response? "The Overton Window appears to have shifted with regard to the acceptability of political violence against elites and their property,” writes Brian Merchant in a grimly fascinating story about most-wanted cards featuring the faces and addresses of Silicon Valley elites.
Although we desperately need to correct for our credulity of Silicon Valley, it’s possible that an accelerating techlash could overcorrect and undermine the case for important and helpful technological progress (as Nobel Prize-winning economists Daron Acemoglu and Simon Johnson point out, there are lots of technologies – beyond software and computers – that remain underdeveloped and underexploited). Many technological products and services are what we call “credence goods:” we can’t understand how they work so we can’t kick their tires; even after we’ve used the tech, we mostly just have to trust that it did a good job. For example, how can I verify that an AI agent got me a good price on a flight? Government regulation can help build trust in credence goods, but trust can be a fragile thing, and if regulation is absent or rolled back and I decide an AI agent doesn’t deserve my trust, what other technologies might I start rejecting? In a much publicized study from the UK, nearly half of all surveyed young people (aged 16-21) said that they would “prefer to be young in a world without the internet.” What does that portend about the relationship between people and technology?
I have to confess that the whole time I’ve been writing this book, a question has been nagging at me. A little voice in the back of my head has been asking “am I really just writing a critique of capitalism here?” After all, I’ve spent a lot of time establishing that Silicon Valley’s successes often come from procuring favorable legal treatment rather than from technological superiority, and as Katharina Pistor establishes in her excellent book The Code of Capital, favorable legal treatment from the state is what turns things into wealth-begetting capital in the first place. And the incisive sci-fi writer Ted Chiang, who gave us our “applied statistics” framing for AI, has said that “most of our fears or anxieties about technology are best understood as fears or anxiety about how capitalism will use technology against us.”
But I don’t really think the problem is capitalism per se – we’d frankly be a whole lot closer to the free market ideal than we are now if we were to eliminate Silicon Valley’s subsidies. The problem is capitalism that’s been completely unshackled from legal restraints. As The Bulwark’s J.V. Last put it in a podcast (a podcast in which he gleefully rehashes one billionaire’s less-than-stellar attempts to play tennis on the pro circuit), you can be generally supportive of capitalism but “it’s real hard not to go full Bolshie” when you see how some of these billionaires carry on. A capitalist system that makes no allowance for curbing extreme wealth accumulation, or providing public goods like education or healthcare, or protecting the vulnerable, is at risk of collapsing in on itself. Just as we need the law to save technology from techno-solutionism, the law also needs to save capitalism from its worst excesses or else it will be unsustainable. I’m not suggesting that we need a revolution here, just some legal reforms. As Pistor insightfully notes, we got here as a result of “persistent incrementalism” by the wealthy and powerful, and reversing all their incremental legal gains may be “a viable strategy to push back and ensure that democratic polities may rule themselves by law.”
In my view, the best way to save capitalism and technology from themselves is through more precautionary legal approaches. People often wrongly describe precaution as a kybosh on progress: in his Techno-Optimist Manifesto, Marc Andreessen melodramatically writes that “our enemy is the Precautionary Principle, which would have prevented virtually all progress since man first harnessed fire.” But Andreessen is attacking a caricature of precaution here. What I’m actually calling for is for lawmakers to err on the side of caution – “better safe than sorry” – when the harms of a technology could be significant. Being thoughtful about how a technological solution will work in the real world doesn’t kill innovation and progress, it just creates breathing room for us to ask and answer the questions that we saw John Ralston Saul pose at the beginning of the book – progress for whom, at what pace, and on whose terms? Precautionary approaches probably will kill off some really bad innovation (which might be why Marc Andreessen doesn’t like them) but the rest of us will benefit if technological innovation is channeled to better serve the public. And while there’s a whole body of literature out there on how Americans don’t typically embrace precaution, maybe it’s time for us to revise our understanding of what Americans really want when it comes to technology and the law. One poll from May 2025 found that 77% of surveyed Americans “want companies to create AI slowly and get it right the first time, even if that delays breakthroughs.”
I said at the beginning of the book that our societal balance between optimism and skepticism has gotten badly out of whack; precautionary approaches are a way of righting that balance. In Thinking Fast and Slow, Daniel Kahneman explains that “a message, unless it is immediately rejected as a lie, will have the same effect on the associative system regardless of its reliability.” Precaution makes skepticism the default, from the very beginning, before our brains can be conned into believing hype that can be difficult to dislodge. We don’t have to be certain of what could go wrong to exercise precaution: assuming from the outset that Silicon Valley talking points are self-serving hype is a good way to prevent us from getting too attached to techno-solutions from the beginning. In other words, precaution can invert the “bullshit asymmetry principle” we talked about in Chapter 1 by creating a presumption of bullshit, then the burden is on Silicon Valley to earn our trust and adequately address the concerns raised by domain experts. The burden will also be on Silicon Valley to explain to the rest of us how the technology actually works – which the hype men may struggle to do (many Silicon Valley techno-optimists are MBAs with no technical training; ditto for a lot of the consultants who hawk these techno-solutions).
As a practical matter, a precautionary approach would presume that all new technological applications and businesses are required to comply with all existing laws from the beginning. As Marietje Schaake argues in The Tech Coup, “the highest goal of democratic governments is not, and should not be, innovation. Rather, it is about making sure that various trade-offs, between innovation and safety, digitization and nondiscrimination, are managed in line with the rule of law. The goal is to prevent companies from moving fast and breaking things.” Instead of accommodating new business models with special legal treatment, “the default answer to requests for new exemptions, [or] special regulation…should simply be “no,” as Pistor puts it.
One concrete legal consequence of such a shift would be an end to the fintech and AI “sandboxes” proliferating around the world. It just so happens that I have the dubious distinction of having written one of the seminal academic articles on how to design these regulatory sandboxes. When I started writing that article back in 2017, I was on the fence about sandboxes: I was already uncomfortable with the idea of peeling back the public’s legal protections to help out tech businesses, but I thought sandboxes might be a good way for regulators to learn more about new technologies and improve their regulation accordingly. My goal was to try and redeem sandboxes with good design recommendations, but with the benefit of hindsight, it's become clear that regulators can and should learn about new technologies without spending their precious resources on a sandbox – in part because of the deregulation involved, in part because regulators can’t help losing objectivity once they become cheerleaders for businesses they’ve selected to participate in their sandboxes, and finally because it can be challenging to reverse a sandbox’s legal dispensations once they’ve been granted.
When there are strong concerns about technological applications inflicting harms, precaution can take the form of legal requirements for a license or some other kind of pre-approval before a business can proceed. For example, the AI Act enacted by the European Union in 2023 has identified a range of “high risk” AI use cases that “will be assessed before being put on the market and also throughout their lifecycle.” Fancy lawyers will always work to find ways to avoid these kinds of pre-assessments (and other existing laws, for that matter), and sometimes they will succeed. But fancy lawyering is less likely to be persuasive if there is a presumption that dispensations should be frowned upon than in circumstances where everyone is falling over themselves to accommodate new technologies. As Pistor says in her book, “whoever claims that individual private gains will translate into social welfare improvements should bear the burden of proof for showing the mechanisms by which this feat will be accomplished.”
If it becomes clear that a particular technology is unlikely to ever deliver on its promises and its harms are significant, then bans should be on the table as an option. For example, the EU has banned the use of AI in certain scenarios, including for facial recognition in public places. I’ve argued on the Senate floor that crypto should be banned, and in case you think I didn’t really mean it, I followed up with an article in Foreign Affairs titled The Case for Banning Crypto. Crypto industry folks have told me that banning crypto is impossible because it’s “decentralized,” but we learned in Chapter 4 that crypto isn’t really decentralized – there are plenty of people (transaction validators, crypto exchanges) against whom a ban could be enforced. Some people would probably find ways to work around a ban, but a ban doesn’t have to be perfect to make things better. After all, crypto assets are just notations on a database, and without the law to validate them, there’s no there there. If courts refused to enforce any contracts involving crypto assets, they would lose their luster for most folks pretty quickly.
I’m not the only one who has called for a crypto ban: I especially appreciate the framing of the former Belgian finance minister Johan Van Overtveldt, who said that crypto has “no economic or social value” and that “if a government bans drugs, it should also ban crypto.” Given how poorly crypto polls, most Americans' response to such a ban would probably be a hearty “bye Felicia!”…
…but we won’t see any precautionary legal treatment if those in power keep buying into Silicon Valley’s “inevitability” narrative.
Back in Chapter 5, I mentioned the somewhat hostile reaction I received when I asked a conference panelist whether we should really be designing regulatory policy around Silicon Valley’s GenAI hype. He told me that it wasn’t hype because it had already come true – sure, Jan – but he wasn’t the only panelist. Another panelist said that it was ok if we tailor regulation to AI hype, because the hype can work as a hypothetical that allows us to practice how we’ll regulate it in the future (I didn’t get to ask a follow-up question about whether building special legal treatment around hype was likely to become a self-fulfilling prophecy). Another simply said in a somewhat resigned fashion, “we have to believe the hype.” I left the conference feeling pretty depressed.
Right now, the idea that AI is inevitable is pervasive and pernicious. I can’t tell you how many friends I’ve discussed AI problems with, only to have them shrug and say “yeah, that’s terrible, but it’s happening and there’s no stopping it.” But this book has made clear that technology does not spring forth, fully formed, from the technology gods. Technology is built and deployed by humans for their own purposes (usually to make money), and what those humans build and how they deploy it are affected by existing laws and how those laws are enforced. Strong enforcement of existing laws can constrain those humans – and new laws can too – if lawmakers, regulators, and courts have the will to reject the inevitability narrative.
As I thought about this inevitability narrative, like Carrie Bradshaw, I couldn’t help but wonder: if technological progress were really so inevitable, should it really matter how lawmakers and regulators treat it? Marietje Schaake, formerly a Member of the European Parliament, tells a story about a dinner she once attended with top Silicon Valley figures. She describes how she was cornered and asked “did Europeans realize their tendency to overregulate was the reason why no equivalent of Silicon Valley existed there?” But isn’t that ultimately an admission that technological progress can be channeled and even stopped? The Silicon Valley folk treated this as an obvious failing on Europe’s part, but what if, to quote the movie War Games, “the only winning move is not to play”? What if Europe has in fact won by using law to hold some tech businesses back, protecting its citizens and letting other countries be the guinea pigs, ensuring that the worst of Silicon Valley’s pathologies haven’t taken root there? As technology scholars Greta Byrum and Ruha Benjamin have observed, sometimes the best outcomes (in terms of benefit for the broader public) are achieved with non-technological approaches and solutions. Has Europe won by using the law to preserve space for them?
Notwithstanding my musings about its deficiencies and internal inconsistencies, though, the narrative that “Silicon Valley needs to win the innovation race towards inevitable technological progress” is very entrenched in the United States. It will be an uphill battle to swap it out for a narrative that stresses the importance of precaution and public investment. Venture capitalists in particular have long benefitted from one hell of a PR campaign depicting the VC “as a uniquely American and uniquely successful risk laborer, who put his special capacities for risk to work building the American project and growing the assets of others.” Those VCs probably won’t take kindly to a reframing of their role as “people who specialize in “last mile” development of publicly funded research in accordance with all legal requirements and without public subsidies.”
More broadly, the stories Silicon Valley tells about itself make its contributions seem both inexorable and valuable, and deny the label of “innovation” to anything that might come out of the government because – gasp – that might imply that government is sometimes useful and effective, and that Silicon Valley isn’t so special and shouldn’t be able to just do whatever it wants. Their narrative of government incompetence is, however, gaslighting. Some of the most important technological innovations (the internet, hello?) have come from the public sector, and while many people have had bad experiences at the Department of Motor Vehicles, many people have also had bad experiences with corporate chatbots. But this gaslighting about the superiority of private sector innovation has been pretty successful, in large part because money matters when it comes to crafting and challenging popular narratives, and as economists Acemoglu and Johnson describe, “once you are welcome in all high-status forums, your persuasion power grows, and you can start reshaping political and economic power.”
In this book, I have pushed back – pretty effectively, in my humble opinion – against Silicon Valley’s preferred narratives about innovation generally and financial inclusion, efficiency, competition, and security more specifically. I have emphasized the harms that fintech poses for the people who use it, and also for the people who don’t use it (just as anti-smoking activists emphasized the harms to non-smokers, undermining industry arguments that smoking was all about personal choice). But I simply don’t have the megaphone that Silicon Valley has – no one is paying to get this book on the New York Times bestseller list – and this presents something of a Catch-22. All the subsidies we have given to Silicon Valley over the years have been weaponized to build a narrative framing within which it would be very hard for Congress to justify taking away those subsidies.
Rampant corporate lobbying and unrestrained political spending are cancers on democracy, and solving those problems is way above my paygrade (although the nakedly political machinations of the crypto industry discussed in Chapter 8 provide excellent grist for those who are working to get money out of politics). My goal here is much more narrow: to identify some strategies that might budge narratives at the grass roots level to make Silicon Valley’s spending more of a political liability for lawmakers and regulators. It’s a lot more politically acceptable right now for Democrats to be chummy with Silicon Valley billionaires than with the Koch Brothers, for example. The abundance agenda aims to keep it that way, but that narrative doesn’t have to win.
I know it sounds like a little thing, but if you found any aspect of this book eye-opening or funny, talk to people about it and help spread the skeptical word. We need a big tent that includes skeptics from outside the tech world who bring their domain expertise and lived experience to the party, along with skeptics who build and deeply love technology but feel increasingly disillusioned and betrayed by the Silicon Valley elite. As Jathan Sadowski puts it in his book The Mechanic and the Luddite, the goal is for folks from both these camps to look at new technologies and demand answers to the questions “What do you do? How do you work? Who do you work for?”
Another goal is for folks in this big tent to vote for people willing to stand up to the Silicon Valley elite. I strongly suspect that the Venn diagram of Congressional Democrats who support the crypto industry and Congressional Democrats who support the abundance agenda looks a lot like a circle. The Crypto Council for Innovation’s Sheila Warren told a group chat of Democratic Party insiders that “nobody is going to get primaried because they voted for [the stablecoin law] GENIUS,” but she may have misread the electorate. We saw last chapter how the crypto industry’s standwithcrypto.com ratings have helped pervert our political process, but we can make lemonade out of lemons and support primary challenges against candidates who have strong pro-crypto rankings – which will probably stop the spread of the abundance agenda too. Silicon Valley’s money and power may seem insurmountable right now, but if we were to go back about fifty years, we would see how the tobacco industry also weaponized publicly subsidized profits to capture Congress and push their preferred narrative. Yet as Sarah Milov relates in The Cigarette: A Political History, in the end, the tobacco industry lost control of the narrative as public interest litigation and grass roots activism made smoking socially unacceptable – and anti-smoking laws followed this narrative shift.
My favorite description of how incremental changes in the public narrative make way for legal changes comes from Anthony Trollope’s novel Phineas Finn (also published as a serial, as it happens, back in the 1860s). Trollope writes that:
Many who before regarded legislation on the subject as chimerical, will now fancy that it is only dangerous, or perhaps not more than difficult. And so in time it will come to be looked on as among the things possible, then among the things probable;—and so at last it will be ranged in the list of those few measures which the country requires as being absolutely needed. That is the way in which public opinion is made.
For a more structured take on how narratives can shift public opinion, we can look to the Narrative Policy Framework developed by political science professors Jones, McBeth, and Shanahan. Their framework stresses that for narratives to be effective in shifting public opinion, people need to agree with the story’s setting, the “facts on the ground” that underpin the narrative. The characters also need to ring true – people will reject narratives if they don’t conform with their pre-existing ideas of who the good guys and the bad guys are. Once the setting and characters have been established and accepted, then the plot can unfold and lead the listener to the moral of the story, the call to arms – in our instance, a plot that outlines the lengths to which Silicon Valley will go to prop up underwhelming and harmful tech businesses, which culminates in a call to eye Silicon Valley hype with withering skepticism and apply the law accordingly.
In 2025, it’s easier than it used to be to make a compelling case that members of the Silicon Valley elite are the villains of the story – it sometimes seems like they’re doing their damnedest to help erode whatever is left of Silicon Valley’s “good guy” framing through their political machinations (I can’t remember where I saw this, but during Elon Musk’s tenure as the head of DOGE, someone posted that Democrats should capitalize on Musk’s unpopularity because “when life gives you Bond villains, make Bond villain lemonade”). But other facts on the ground remain a sticking point for changing the narrative. As I discussed in Chapter 6, for quite some time, the facts on the ground have been that the Silicon Valley elite are brilliant geniuses (evil geniuses, perhaps, but geniuses nonetheless) responsible for their own success, and that the technological marvels they have produced are worthy of our innovation worship. A call for skepticism is only going to work if people no longer agree with these facts on the ground.
As I’ve discussed at various points throughout this book, the enshittification of Silicon Valley products we all used to love, as well as distrust of new AI-related products, are already helping to change people’s perceptions of Silicon Valley’s output. Maybe that will be enough in and of itself to change the narrative. It’s also possible that Silicon Valley will become synonymous with war and police states – if that happens, the existing techlash (which Jathan Sadowski argues was kicked off by Google workers’ protests over the company’s AI contracts with the military) will probably accelerate. Former business professor and self-described technology contrarian Jeffrey Funk argues that “technologies, not business models, enabled many of the successful start-ups of the previous generation to succeed” – technologies that were developed in universities or the defense sector with government funding. Given that Silicon Valley is currently helping to prop up a United States government that has eliminated a huge portion of the country’s scientific and technology research funding, there’s likely to be less and less for Silicon Valley to commercialize beyond surveillance and weapons technology.
And so there are things afoot that could snowball into narrative change and make political associations with Silicon Valley more toxic for regulators and lawmakers. But there are many in the media and in universities helping to launder Silicon Valley’s reputation, and they could prove to be obstacles to the needed narrative shift. Silicon Valley is keenly aware that prevailing narratives can be challenged by reporting and academic research, and so it has sought to co-opt both.
In our battle of narrative frames – precaution and public investment versus inevitability and techno-solutionism (with the latter perhaps wearing an abundance costume) – it will be harder to get people behind precaution and public investment while Silicon Valley’s preferred framing is breathlessly repeated by so many tech journalists. To be clear, there are many wonderful tech journalists out there, and this book would not have been possible without them. But there are also some high-profile tech journalists who are very much cogs in the Silicon Valley hype machine.
We talked about Casey Newton’s attempts to disparage AI criticism in Chapter 5; Kevin Roose of the New York Times is another high-profile pusher of industry hype. In March of 2002, Roose published a “Latecomer’s Guide to Crypto” that he billed as a “sober, dispassionate explanation of what crypto actually is” but it was actually so over-the-top credulous that about fifteen independent crypto researchers felt compelled to annotate the thing to fact-check its wildly pro-crypto narrative. Roose was undeterred: one of his 2025 publications was titled “Powerful A.I. is coming. We’re not ready,” in which he pushes Silicon Valley’s preferred (unsubstantiated) narrative that “artificial general intelligence” is inevitable and indeed almost within reach. Roose has also joined Newton in piling on AI critics, saying “I feel like the people who are denying the capabilities of these models are just telling feel-good bedtime stories to people who don’t want to believe that change is coming.” Never mind that many of the people denying the capabilities of these GenAI models are acutely aware of the changes they are managing to wreak despite their limitations...
Then there’s tech journalist superstar Kara Swisher. Full disclosure, I have a weirdly personal take on Swisher, because her doctor brother administered my epidural when I was in labor with my first child, and he was bragging on her the whole time he was prepping me. I thought it was sweet that he was so proud of her, but my epidural only “took” on one side of my body, so now I associate Kara Swisher with the pain of partially unmedicated childbirth. That is probably not fair to her, so let’s hear what someone else has to say about Swisher’s reporting.
In a blog post titled “You Can’t Make Friends with the Rockstars,” AI critic Ed Zitron argues:
The real problem children are, of course, people like Kara Swisher, who, more than anything, WANTS to be friends with the rockstars and has done so successfully. Swisher’s book tour involved her being interviewed by Sam Altman and Reid Hoffman — a shameful display of corruption, one so flagrant and stomach-turning that it should have led to an industry-wide condemnation of her legacy.
In other words, Swisher may dish out some critiques, but there’s always the risk that she’ll pull her punches to maintain friendly access to the tech superstars, and being so close to the tech industry can create blind spots. In Chapter 1, we talked about the journalist John Carreyrou, who broke the story of the Theranos fraud. Carreyrou’s editor has observed that “John had this powerful combination of being an expert on health care and existing outside the ecosystem of Silicon Valley, so he wasn’t sucked into that relationship between and among venture capitalists, entrepreneurs, and the media. He was able to look at it with an outside lens — which allowed him to see through it. Some other members of the media didn’t because they were very close to it.” Contrast Carreyrou with Casey Newton, who – seemingly without a shred of self-awareness – defends access journalism, dismissing AI criticism from Ed Zitron and others by saying “it’s so important when you’re reading about AI to [instead] read people who actually interview people who work at these companies.”
Even though so much of the media deck is stacked in favor of Silicon Valley, in the eyes of some of the techno-libertarians we profiled in Chapter 7, that deck isn’t stacked far enough. Jeff Bezos, for example, has changed the editorial policy at his Washington Post so that “we are going to be writing every day in support and defense of two pillars: personal liberties and free markets. We’ll cover other topics too of course, but viewpoints opposing those pillars will be left to be published by others.”
One reaction to these kinds of problems in the mainstream media has been a proliferation of new “Substack” newsletters by independent writers. You may have noticed that, at a time when new Substacks are popping up like mushrooms, I didn’t use Substack as a platform for this book. Originally, I was going to. Then I found out that Substack, like almost every techno-solution discussed in this book, is funded by Andreessen Horowitz. Andreessen Horowitz also helped fund Elon Musk’s purchase of Twitter and Marc Andreessen sits on the board of Meta: I’ve quit both Twitter and Facebook following their enshittification, and I decided I didn’t want to commit to a platform that could very well be “Xed” in the future. While Substack has placed billboards in DC’s Union Station proclaiming “Media Isn’t Dead. It’s on Substack,” a stated goal of Andreessen’s “good friend” Curtis Yarvin is to destroy journalism and academia (which he refers to – on his Substack, naturally – as “the Cathedral”). Peter Thiel has also reportedly said that he wants to replace media with something he calls “full stack narrative.” If the destruction of traditional journalism is Andreessen Horowitz’s goal for Substack, I didn’t want to unwittingly be part of something that dark.
I’m a big fan of many folks who publish with Substack, but their readers are dealing with content overload, struggling to manage the volume of newsletters they’ve subscribed to. Many Substack writers also face the challenges of self-employment: journalists should be able to avail themselves of the benefits of being an employee of a media outlet (benefits like health insurance, editors, and in-house lawyers). Figuring out how to fix our broken media landscape is once again above my paygrade, but in the meantime, if we want to reframe the conversation about technology, it will help to have more of the journalists who are currently employed by media outlets flex their tech skepticism. Journalists focused on technology’s real-world impacts can help drown out overly credulous access journalists like Casey Newton, but I’ve spoken to many a journalist who is nervous about covering the social impacts of crypto because they feel they’re not experts in the tech. I hope that this book helps embolden them a little: just like you don’t have to be a petrochemical engineer to report on the energy industry’s climate impacts, you don’t have to be a tech expert to report on Silicon Valley ruining things.
I mentioned in the last chapter that the crypto industry is Philip Morris in a grey hoodie, and I’m clearly not the only one thinking along those lines. In a paper titled The Grey Hoodie Project: Big Tobacco, Big Tech, and the Threat to Academic Integrity, researchers Mohamed Abdalla and Moustafa Abdalla observe that “both industries’ increased funding of academia was as a reaction to increasingly unfavorable public opinion and an increased threat of legislation.” Universities should theoretically be a breeding ground for informed skepticism, but it can be hard for researchers to critically examine how technology is being developed and used when their funding comes from Silicon Valley.
The Tech Transparency Project reported in 2023 that “Mark Zuckerberg’s personal philanthropy and his company, Meta, have collectively donated hundreds of millions of dollars to more than 100 U.S. colleges and universities across the country.” Meta is not alone: way back in 2015, journalist David Dayen published a report on how Google strategically sought to influence academic treatment of antitrust issues by funding friendly academic research (particularly through George Mason University). More recently, Abdalla and Abdalla found that the largest Silicon Valley companies have been particularly aggressive in funding AI-related research – through direct faculty funding, sponsoring conferences, and paper awards, amongst other things. They conclude that, given this influence, “it makes sense that much of the [AI] fairness work that exists holds the entrenched Big Tech view that “social problems can be addressed through innovative technical solutions.”” Even if these tech companies don’t try to overtly try to drive academic research towards techno-solutionism, as one article put it, they “wield “soft power,” slowing down research, sparking tension between academics and their institutions, and shifting the fields’ targets in small — but potentially transformative — ways.”
Professors don’t just research, they also teach, and if professors have opted into Silicon Valley’s hype, they are likely to pass that on to many of their students. In addition, many university courses are taught by industry personnel serving as adjunct faculty. This has the benefit of bringing cutting edge industry expertise into universities, but we should expect it to also bring some hype into the classroom, blunting the ability of students to develop more skeptical takes on the technology they’re learning about. AI researcher Timnit Gebru has observed that Effective Altruists (the “EA” in the TESCREAL bundle we covered in Chapter 7) are teaching courses on AI at elite universities, so that hype may sometimes come with a side order of ideology as well. I’ve also noticed that the academic affiliations of adjunct faculty can be weaponized to make industry interests seem more impartial. For example, when Linda Jeng testified about crypto before the Senate Banking Committee in February 2023, she used the title of “Professor” and stated that she was testifying in her personal capacity as an academic and researcher. She listed her adjunct academic responsibilities at Georgetown before she mentioned her actual day job as Chief Global Regulatory Officer for the Crypto Council for Innovation.
Trying to get more independence in academia is challenging for the same reason that trying to get more independence in media is challenging. The problem is money, and the need for public funding is becoming particularly acute at a time when the same techno-libertarians trying to end independent media are also looking to end universities as we know them. Perhaps no member of the tech elite has been more aggressive in trying to undermine university education than Peter Thiel, who made all his most important connections at Stanford University but has since funded fellowships to encourage students to drop out of college and attend start-up incubators instead.
This is frankly another reason why I’m troubled by Silicon Valley’s GenAI push – the more that college students delegate their writing to GenAI tools like ChatGPT, the less critical thinking they’ll develop and universities will seem increasingly useless which will justify funding cuts that will give universities fewer resources with which to teach their students critical thinking. I’ve seen some people muse on social media that folks in Silicon Valley are intentionally trying to use GenAI tools to create a nation of sheeple they can rule over, and I’m not sure they’re wrong – especially because many of the tech elite limit their own children’s access to screens and social media and will undoubtedly do the same with AI.
Thiel saves his strongest vitriol for the humanities – you know, the disciplines where people might develop their critical thinking and get good at poking holes in Silicon Valley’s hollow techno-solutionism. Thiel says that “right now the humanities are transparently ridiculous,” but I couldn’t have written this book if I hadn’t learned what I learned from my degree in English literature. Just sayin’ …
Although many media and academic institutions are compromised by ties to Silicon Valley (not to mention besieged by funding cuts from the Trump administration), there are still many independent journalists and scholars who can and do play an important role in reframing how we look at technological innovation. But skeptics without the benefit of a university or publishing platform can also help change the narrative, and they can do it the same way everyday people have always undermined the power of the powerful – by laughing at Silicon Valley and its narratives of inevitability. As activist Saul Alinksy put it, “ridicule is man’s most potent weapon. There is no defense. It is almost impossible to counterattack ridicule.” While I don’t want to downplay the harm that Silicon Valley does, it is still a ridiculous place, and ridicule is one of the best ways to defuse Silicon Valley’s narrative power and rebalance our yin and yang in favor of skepticism. In the classic comedy Monty Python and the Holy Grail, King Arthur says to his knights (after a ridiculous musical number) “on second thought, let’s not go to Camelot. It is a silly place.” The goal is to change the “facts on the ground” so that everyone thinks of Silicon Valley as a silly place too.
Jonathan Swift’s satirical classic A Modest Proposal, published in 1729, lampooned the solutionism of his own era’s elite – to wit, overly simplistic ideas for solving poverty in Ireland. To point out how callous and oversimplified those proposals were, Swift modestly proposed that it would be a win-win if the Irish poor sold their babies so that the elite could eat them (while I don’t think today’s Silicon Valley elite would be interested in eating the babies of the poor, I could probably write a pretty scathing modest proposal about them purchasing poor babies for longevity research purposes…). Frederick Douglass recognized the power of this kind of ridicule in his 1852 speech What to the Slave is the Fourth of July? when he cried that “scorching irony, not convincing argument is needed. O! had I the ability, and could I reach the nation’s ear, I would, today, pour out a fiery stream of biting ridicule, blasting reproach, withering sarcasm, and stern rebuke.”
To get a little more modern, I grew up watching Mel Brooks movies like Young Frankenstein, Blazing Saddles, and History of the World Part I. To this day, I still know every line in the movie Spaceballs (and have been known to mutter to myself sotto voce “I knew it, I’m surrounded by assholes” in certain company). But Mel Brooks’ first hit movie was The Producers, which is about Broadway producers who figure out they can make more money with a flop than a hit (now that I come to think about it, I wonder if there aren’t some Silicon Valley VCs running that exact playbook…). These Broadway producers set out to make the worst musical ever – Springtime for Hitler. Unsurprisingly, many people were offended by the movie, but Brooks (who is Jewish) had this to say: “Listen, get on a soapbox with Hitler, you're gonna lose — he was a great orator. But if you can make fun of him, if you can have people laugh at him, you win.”
There are plenty of great hype men in Silicon Valley, but that hype loses its power if people start laughing at it. I suspect that Silicon Valley hype is effective in part because people want to believe that the world is better than this – that techno-solutionist bullshit couldn’t possibly be perpetuated at such scale in such a cynical way, so there must be some germ of promise in it. Accepting that Silicon Valley can really be this cynical can break your brain, and humor is probably the most palatable way to deliver this kind of brain-breaking message. That is why I am prescribing you all a watch (or rewatch) of HBO’s Silicon Valley TV series to help inoculate you against techno-solutionism.
There are also lots of other people out there making fun of Silicon Valley that you might enjoy. The Onion’s “Compost-Fueled Car” video was an instant classic:
Step one: devise an idea to create a car that runs on compost. Step two: create the car. We’ve already completed step one, we’re halfway there!
And in 2024, the comedian John Mulaney made mincemeat of Silicon Valley self-importance in a private set he did at a “Future of AI” conference hosted by Salesforce in San Francisco:
“What’s important here is that we’re looking for solutions. And in looking for solutions, what we’re really after is insights, which then lead to success. Now, start prepping the humans for robots.” Some of the vaguest language ever devised has been used here in the last three days, the fact that there are 45,000 ‘trailblazers’ here couldn’t devalue the title any more.
I also recommend the Financial Times blog skewering how OpenAI’s Sam Altman cooks in his home kitchen. Do yourself a favor and read the whole thing, but author Bryce Elder closes by musing:
Maybe it’s useful to know that Altman uses a knife that’s showy but incohesive and wrong for the job; he wastes huge amounts of money on olive oil that he uses recklessly; and he has an automated coffee machine that claims to save labour while doing the exact opposite because it can’t be trusted.
This is just a small sampling of some of the Silicon Valley parody gems on offer. Get out there and enjoy, and let me know of other ones that are particularly good! Experiments in cognitive psychology suggest that techno-solutionism will not be felled with dry statistics about how unlikely it is for technological solutions to succeed (and that’s even before we get into arguments about who gets to decide what constitutes “success”). But a little humor can help us get better at rejecting Silicon Valley’s hype: Evgeny Morozov concludes his book on techno-solutionism by explicitly calling on us to ridicule it.
I’ve now delivered the main “action item” of this book – laugh so much it makes you skeptical of Silicon Valley’s bullshit (including the “star pills” and other tech variants of the abundance agenda). If we can manage this, then maybe when the time comes to rebuild, we’ll reject Silicon Valley’s oversimplistic offerings and invest in real, long-term solutions. But as economists often say “it takes a model to beat a model,” and I’ve found over the years that when you explain why Silicon Valley’s techno-solutions are ridiculously unworkable, the techno-solutionists will sometimes retort “have you got a better idea?” It’s far easier to embrace skepticism of Silicon Valley’s version if you have your own vision of what real progress would look like.
It is true that right now, things aren’t great in traditional finance (to put it mildly), and we’ve seen that the Silicon Valley hype men sometimes offer a pretty accurate critique of the things that aren’t great about it even if they can’t fix them. I’ve argued that we should take away some of Silicon Valley’s subsidies because they’re not furthering the public interest, but what about the subsidies that the financial industry gets? Law professor Mehrsa Baradaran describes the basic bargain as consisting of:
a government promise that it will protect banks from runs, liquidity shortages, and investor irrationality, and a promise made by banks that they will operate safely, play their essential role in financing the expansion of the economy, and serve the needs of their customers and local communities.
But, as she goes on to explain, this bargain has gotten badly out of whack. Subsidies for banks were adopted to make the financial system run more smoothly for everyone and we shouldn’t eliminate them, but we do need legal changes to force the banking industry to better deliver the quid pro quo of serving the broader economy. One way to challenge the fintech dystopia that Silicon Valley is trying to force feed us is to offer a competing vision for what “serving the broader economy” should look like.
When it comes to Americans’ economic precarity, I argued in Chapter 2 that purported fintech solutions are likely to be both exploitative and distract from needed changes to the minimum wage and social safety nets. Instead, we need fewer of the predatory gambling and lending services covered in earlier in the book – traditional banks, publicly subsidized as they are, should not be allowed to support those kinds of services through rent-a-bank arrangements. But this book has also covered some situations where the financial services needs of the public really aren’t being met. Access to bank accounts for the unbanked springs immediately to mind, as does the fact that it takes too long for people’s paychecks to clear. Delegating our problem-solving to the private sector for so long has caused public sector capacity to atrophy (much like depending too heavily on GenAI tools will cause our critical thinking to atrophy…), but there are reasonably obvious public solutions to these financial services problems.
As we explored in Chapter 3, banks could be mandated to provide no-cost, no-frills accounts to unbanked customers. On clearing paychecks, the technology already exists for faster payments, so this is ultimately not a technology problem – payments processors simply haven’t made faster payment services available to their customers. The Brookings Institution’s Aaron Klein has noted that this problem can be fixed “by simply amending the Expedited Funds Availability Act to require immediate access for the first several thousand dollars of a deposit, instead of permitting the lengthy, costly delays that harm people living paycheck to paycheck.” If we do want new technological solutions to help improve payments, there are precedents for public solutions that have proven quite effective. Notably, the Brazilian central bank provides the infrastructure and sets the rules for the PIX real-time payment system, which has seen enormous growth and reduced consumer costs in Brazil since its adoption in late 2020.
PIX is not without its problems, of course – in particular, it has resulted in an explosion of fraud in Brazil. As I’ve heard some people say, “faster payments, faster fraud,” and fraud has been increasing around the world (the situation became so dire in the UK that in 2024, banks were given greater power to slow down payments to investigate whether they are fraudulent). But while faster isn’t always better, there’s definitely room for improvement when it comes to the speed of payment processing in the United States. In fact, the Federal Reserve already rolled out a faster payments service called FedNow in 2023, but banks have been slow to join – Congress could pass a law requiring them to do so.
There’s also room for improvement when it comes to protecting privacy and resilience in our payments system. Cash transactions are increasingly being phased out, but as we saw in Chapter 3, there are many reasons to keep them. Laws can and should be passed that require businesses to continue accepting cash, ensuring that our cash infrastructure is preserved for when we need it, but even with such laws, cash payments will remain impractical for large payments and payments over distance. Again, there is scope for public solutions to address this problem. For example, Raúl Carillo has proposed a “Postal Cash Card” that can store value and facilitate transactions in a way that emulates debit cards but does not generate any data about the holder or their payments. This proposal shows how rejecting techno-solutionism doesn’t have to mean rejecting technology: Carillo has proposed a technological innovation (the card), but he has also provided a detailed proposal about the institutional context in which it will be offered (non-profit, at the post office). This proposal also makes it clear that when it comes to technological innovation, incentives matter: a technology developed by the public sector for a non-profit purpose is more likely to avoid the siren song of mass data collection then a technology developed by Silicon Valley.
Fixing finance requires more than just better efforts to meet the needs of the public, though. Fixing finance also requires structural changes to make the financial system more stable, so that financial crises won’t endanger the broader economy that the financial system is supposed to serve. Many of the structural changes I’m about to propose would have seismic impacts on our current financial system, and they will be viewed by a lot of people who work in the financial industry as outrageous. But our present state of affairs – where we subsidize and provide safety nets for what is essentially gambing by wealthy financial institutions – is also pretty outrageous, and we’re only desensitized to it because it has happened incrementally over the space of half a century.
If we go back to brass tacks, our financial system is supposed to connect people who want to profit from investing with people who want to make productive use of investors' money and are willing to pay for it. Financial markets connect these people, and also allow early investors to cash out by selling their investments to others: the opportunity to eventually exit encourages early investors to invest in economic enterprises in the first place. But as law professor Saule Omarova explains, “financial innovation helped to sever the key functional link between finance and non-financial economic enterprise.” As new types of financial products have been “innovated,” finance has become increasingly detached from its original role as an auxiliary support system for the broader economy and started to look more like straight-up gambling among financial institutions. As I mentioned earlier in the book, given how speculative traditional finance has become, it’s not altogether surprising that some people can’t tell the difference between crypto and other financial markets. Once upon a time, law courts wouldn’t even enforce purely speculative contracts; now, speculative finance for finance’s sake is par for the course.
For five years now, I’ve been screaming from the rafters that crypto needs to be kept separate from traditional finance – so that its volatility and general scamminess don’t infect the traditional financial system, and so there will be no temptation for the government to bail out crypto holders and businesses. Obviously, no one is listening to me right now and we’re headed in the opposite direction, but if I’m right about crypto policy, then that also points us in the direction of what to do about traditional finance. It may be that the only way to escape our downward spiral of leaving taxpayers and central banks on the hook for more and more financial speculation is to start severing the links between the parts of the financial system that should be subsidized (and highly regulated as a result), and the speculative rest.
That, after all, was the kind of solution that Congress settled on in response to the Great Depression. The Glass-Steagall Act passed in 1933 prohibited banks from engaging in speculative capital markets activities, and prevented non-bank financial institutions from doing anything that looked remotely like accepting a deposit. That separation served us well until it started to be eroded in the 1970s and was finally torn down by Congress in 1999. In his book Taming the Megabanks, law professor Art Wilmarth makes a pretty compelling case that we need to resurrect structural separations between subsidized banks and more speculative finance. As part of that structural separation, no financial institution other than a bank should be able to perform a bank’s core function of accepting deposits or their equivalent. Since Wilmarth wrote his book, Congress has passed stablecoin legislation that allows commercial businesses (like grocery stores and tech platforms) to accept deposit equivalents – so now we need to roll that back too.
Unfortunately, implementing structural separation in today’s world will involve some challenges that Congress didn’t have to contend with back in 1933. In particular, the traditional banking business is being hollowed out through all kinds of outsourcing, so that banking increasingly resembles a supply chain with only one link in the chain being subject to banking regulation. Many of the mortgage loans that banks used to make, for example, are now being made by fintech lenders with money borrowed from banks. Many of the decisions about risk management that humans used to make are now performed by machine learning tools. Operational functions that banks used to perform in-house – like data storage – are increasingly being outsourced to third-party technology providers, like cloud providers.
Banking regulators sometimes struggle to get access to the inner workings of the technological tools that banks are using to perform key functions, because the tech businesses who provide those tools assert trade secrecy protections or argue that banking regulators have no jurisdiction over them. And so we may need to simply tell banks that they cannot rely on technology providers who won’t be open and frank with regulators – and if that requires legal changes to trade secrecy protections, well, so be it. The law giveth those protections, and so the law can taketh them away too. It’s also true that financial regulatory bodies sometimes struggle to hire employees with the technical expertise necessary to interrogate technological tools once they get access to them, but those challenges shouldn’t be insurmountable if the agencies are properly funded (especially because Silicon Valley has laid so many people off recently, so there are plenty of skilled tech workers to be found).
To prevent banks’ subsidies from leaking out into other parts of the financial system, we may also need to limit the ability of subsidized banks to fund other financial service providers. For example, Congress could limit bank activities so that no more than a specified percentage of a bank’s loans could be made to businesses that engage in activities that are financial in nature (fortunately, there’s already a pretty broad statutory definition of “activities that are financial in nature”). Instead, banks would be forced to do more of their lending to non-financial businesses, helping to grow the non-financial parts of the economy. That doesn’t mean that financial businesses like hedge funds and non-bank mortgage lenders that currently depend on bank loans need to disappear: they should be able to borrow modest amounts from banks, and then they can raise their own capital through equity contributions to supplement those loans. But if non-bank financial firms can’t exist without borrowing from banks, then that tells us something about what our subsidies for banks are supporting – and who we’re likely to end up bailing out if we don’t change course.
If our goal is to structurally separate banking and non-bank financial activities, policies encouraging “open banking” also need a serious rethinking. “Open banking” relies on software to share customer data between banks and fintechs, making it easier for bank customers to receive financial services from fintechs rather than their banks. This open banking is part of a broader movement that banking nerds refer to as “unbundling” banking: the words “unbundling” and “abundance” sound very similar, and open banking is indeed very abundance-y in that it presumes that making it easier for less-regulated fintechs to serve bank customers will result in innovation that trickles down benefits for everyone. But as we’ve covered at length in this book, technology doesn’t change people’s motivations, and less-regulated fintechs will have the same incentives as banks to seek privatized gains at the expense of socialized losses. They just won’t have as much regulation reining them in.
Fixing finance shouldn’t look like a Silicon Valley fever dream of regulatory arbitrage and abdication of government oversight, but right now, we’re throwing up our hands and letting banking services migrate outside of the regulated perimeter in an unjustified and misguided hope that less regulated fintechs will somehow do it better. Some will no doubt warn that if we reverse course on open banking, it will insulate banks from fintech competition. Sounds good to me! As we discussed in Chapter 4, regulatory arbitrage shouldn’t be the basis for a business’ competitive edge, and competition on an uneven regulatory playing field is unlikely to be in the public interest. Letting financial services leak out of banking and into fintechs also relieves pressure for legal reforms that would make meaningful improvements to competition in banking: if we truly want more competition within the banking industry, then let’s do something to directly tackle the issue. We could, for example, take a more robust approach to limiting mergers among banks, as law professor Jeremy Kress has argued for, or go further and break up the largest banks, as Senator Elizabeth Warren has sometimes proposed.
As I’ve already alluded to, there would be a lot (a LOT) of financial industry resistance to these kinds of legal changes. I’ve spent a good chunk of this chapter talking about changing the narrative to generate public support for policies that eliminate Silicon Valley’s subsidies and require tech business to play by the same rules as everyone else; I’ve also alluded to how much money will be pitted against this kind of reform, which is why wealth taxes and getting money out of politics are so critical. All of that more or less goes for financial reform too, as Gerald Epstein makes clear in his book Busting the Bankers’ Club. A particular challenge for financial reform is that the public typically isn’t very interested in supporting these kinds of policies except immediately after a financial crisis – but sometimes there are other issues that financial reformers can hitch their wagons to, and given how much traditional finance has devolved into gambling, financial reform might find broader support amidst a backlash against online sports betting.
Problem gambling has risen sharply in the United States in the last six years, and it’s likely to get worse as people start to feel the increased economic precarity unleashed by the “One Big Beautiful Bill” (if I wanted to be really cynical, I might suggest that a goal of that bill was to create a bigger market of desperate people for the sports betting and crypto industries to exploit…). The anti-cigarette movement was a grass roots movement motivated by the noxious health consequences of smoking and second-hand smoke, and as the health and other social costs of problem gambling become more apparent in our communities, I wouldn’t be surprised if we started to see similar growth in grass roots movements targeting online betting. If we do, we should work to make sure that speculative finance is also a target of those movements’ ire.
If none of that works, something really catastrophic might be needed to create a political environment in which things can change. As luck would have it <deadpan sarcasm>, we seem to be on a path towards something truly catastrophic both in Silicon Valley and on Wall Street. The stock prices of the largest tech companies – which are responsible for almost all of the S&P 500’s growth in 2025 – are increasingly dependent on their AI ambitions, yet AI is increasingly looking like an unsustainable bubble. Even OpenAI’s Sam Altman came out and said so in August 2025. It’s not clear precisely what the popping of the AI bubble will mean for those tech companies, the stock market, or the broader economy, but given that those stocks made up about a third of the total US stock market in August 2025, we’re likely to see a big stock market crash if (when) that AI bubble pops.
Stock market crashes won’t always cause financial crises (for example, the dot.com bust in the early 2000s didn’t take down the financial system), but they can cause crises if financial institutions have leveraged up by borrowing a lot of money to invest in those stocks. In our current deregulatory environment, financial regulators are unlikely to be clamping down too hard on leverage, and hedge funds in particular are currently dependent on unusually high levels of borrowed money to fund their investments. Given these circumstances, I wouldn’t bet on financial institutions escaping the aftermath of this AI bubble unscathed.
When there’s a crash, highly leveraged financial institutions are forced to sell off their investments (often at a loss) to repay their loans, which drives the prices of the stocks they sold further down, which can then force other financial institutions into selling off their stock investments to repay their loans. And so on. If financial institutions have borrowed a lot and stock prices fall a lot, that can even tip those institutions into insolvency, and if they default on their contractual obligations to other financial institutions as a result of their insolvency, then that can cause problems for those other financial institutions, which may then need to dump more stocks as a result. And we shouldn’t assume that financial institutions that need cash to repay their loans will restrict themselves to selling stocks – they may need or prefer to dump other types of assets. Traditional finance and crypto are merging, and so it’s becoming increasingly likely that some of the other assets they will be dumping will be crypto assets that are essentially Ponzis – with nothing behind them, the price of these assets can go to zero very quickly if enough people abandon them (even if investors haven’t borrowed to buy crypto, it’s still quite likely that they’ll bail out of their crypto investments at the same time as others are bailing out of the stock market, as crypto prices often move in the same direction as stock prices). And then all the institutions that borrowed to buy crypto will be royally screwed, and they’ll have to start selling off whatever assets they have left, implicating other financial markets.
It’s also possible that the causality could also go in the opposite direction, with an initial crypto crash causing the AI stock bubble to burst as part of an ensuing financial crisis. In Chapter 4, we talked about how vulnerable a crypto-based financial system is to panic and crisis. I’m particularly worried that by the time the crash comes, tokenized versions of real financial assets will have been fused with Ponzi-like crypto assets and stablecoins into Frankenstein-style pre-programmed bespoke financial products. It’s hard to predict precisely what will happen when the shit finally hits the fan in ways that these products’ pre-programmed instructions never contemplated, but it’s almost certain that interconnections between different kinds of financial assets will speed up the transmission of panic from one kind of financial asset market to another. If financial institutions start using stablecoins in large volumes for these Frankenstein products and then bail on them when there’s a crypto crash, then stablecoin issuers will need to find a way to meet their redemption requests. It’s highly likely that stablecoin issuers will need to start withdrawing their cash reserves from the banks that hold them, potentially kicking off runs on those banks. It’s also highly likely they’ll be forced to sell off Treasuries from their reserves, which could drive down the price of those Treasuries if there isn’t enough market demand to absorb the sales. That won’t be a good look for what are supposed to be the most stable financial assets in the world, or for the vast global financial markets that depend on the stability of Treasuries for their own stability.
While lots of people are warning that the implosion of the crypto bubble could cause another financial crisis like the one we saw in 2008, I personally think that might be an overly optimistic projection. The deregulation we discussed in Chapter 8 is taking us back to the 1920s, not the 2000s, and so it’s quite possible that we’ll end up with another Great Depression on our hands. When 2008 hit, Ben Bernanke was the Federal Reserve Chair – a man who had dedicated his earlier academic career to studying the Great Depression and how to avoid repeating the mistaken government responses that had exacerbated the economic freefall. In 2008-9, Bernanke worked with Treasury secretaries Hank Paulson and then Tim Geithner on emergency interventions to staunch the financial bleeding. Their response was by no means perfect, but as I said in Chapter 1, I think that if government intervention had not staunched the panic in 2008-9, we might very well have faced an economic depression that would have made life orders of magnitude worse for everyday people. Now think of our current circumstances, in which the independence of the Federal Reserve and the Treasury Department’s room to maneuver are being threatened by President Trump’s attacks on Federal Reserve Chair Jerome Powell and Governor Lisa Cook, personnel issues in the Treasury Department, and ballooning deficits. I’ve said at various points in this book that bailouts are becoming more likely, but should the worst happen, it’s not even clear that the United States is in a position to mount the same kinds of emergency rescues that moderated the economic impact of the 2008 crisis (or the 2020 Covid pandemic). To put it bluntly, will anyone be calmed by the United States’ attempts at emergency intervention if there are no grownups in charge?
So yeah, the combined forces of a major techlash and a major financial crisis would probably get us the political support needed for legal reforms, but let’s hope things don’t need to get that bad to shift the narrative.
Towards the end of 2024, I was on the train home after a fintech conference that had been a very dispiriting experience – very important people had one by one extolled the benefits of moving our financial system onto blockchains, with almost no pushback on their unjustified optimism about the technology, and almost no discussion of the kinds of risks and flaws explored in Chapter 4. I raised these issues during my time on stage, but it was clear that my interventions were about as welcome as a fart in a spacesuit (thanks Billy Connolly, for that particularly choice simile). Even those who agreed with my concerns expressed a certain fatalistic inevitability – that this was happening, whether it made sense or not. I left feeling pretty grim about the state of the world, but it just so happened that I reread the chapter on optimism in Daniel Kahneman’s Thinking Fast and Slow on that train trip, and it was quite the revelation to a person in my frame of mind.
All through this book, I’ve allowed the techno-solutionists to corner the market on optimism. But once I saw Kahneman’s words “the main benefit of optimism is resilience in the face of setbacks,” I started to wonder if I’d ceded that optimism ground too easily. I have put myself in uncomfortable solutions again and again to call out techno-solutionism – on the floor of Congress, on panels stacked with industry insiders, on social media where I’ve sustained torrents of abuse from crypto bros – and then I’ve picked myself up and gotten back in the ring. And many journalists, scholars, and activists around the world have put far more on the line than I have in challenging Silicon Valley.
What could be more optimistic, really, than speaking truth to power, when the powerful are poised to get everything they want? We skeptics aren’t pessimists – we’re the ultimate optimists because we refuse to accept techno-solutions as inevitable and we persist in trying to challenge Silicon Valley despite the odds. Right now, I feel like I’m watching a slow-motion car crash with Silicon Valley in the driver’s seat and there’s nothing I can do to stop it. But maybe – as skeptics grow in number and noise – we’ll stop it together someday.
And now if you’ll excuse me, I’ve got to get back in the ring.
⬩ fin ⬩