by Joshua Walker, CodeX Co-Founder and Fellow

If one is looking for a self-echoing diatribe against a single group or party: look elsewhere.  This article  may be a critique—but a critique of all of us.  A collective challenge—and thus the opposite of an “echo”.

This is rather a technical analysis of why the contemporary governmental attempt to make America governance more efficient has failed.  (Like analyzing data from a rocket that exploded.  To improve our next launch, we must analyze empirically.  And optimized governance is many orders of magnitude more complex than rocket science.^[1])

Has it? 

Has it failed?  Well, DOGE’s inceptional dare was to shave two trillion dollars off of the U.S. government’s budget.  Then, as it got organized: one trillion.  Now, the claimed outcome was reducing the budget in a few months by $180 billion dollars (presumably, an annual savings)—though that is credibly challenged.

But even if it had, or ultimately would, attain its inceptional goal of cutting two trillion dollars off of the federal budget, massively reduced or eliminated the annual US budget deficit (which in a vacuum, is quite a laudatory goal^[2]), could it really have been called “a technical success”?

No.

Not at all.

No one could call it a technical success because DOGE only handles a single dimension: money.  (To quote a sci fi author beloved of Silicon Valley: “42.”)  DOGE is an “extremis”, or limits analysis applied to human governance.  Now, money matters but it’s just one piece of that particular puzzle.  The humans, the users, matter more ultimately, and immediately—as we will see.

And thus DOGE is an airplane designed in one dimension.  It does not merely not fly at all.  DOGE doesn’t begin to conceive of a world in which one could fly between two different points within a finite time horizon.  Because DOGE doesn’t conceive of the vector (magnitude + direction across several dimensions).  It doesn’t acknowledge the multiple dimensions comprising human governance.

By DOGE’s definition, a government which very abruptly cut its budget to zero but just as abruptly doomed its entire populace would be success.  (And so efficient at both!!  ❤️❤️)  Yet “dead country” is probably not the optimization DOGE is aspiring towards; nor what it was conceived for, what it was formally commissioned to do.

So what is the right question to be asking?

Well, what is governance trying to maximize?

Like an airplane, “governance” is an attempt to optimize a machine for flight across four dimensions.  See “On Legal AI II: Transactions” (Walker, vLex Full Court Press; forthcoming, 2025).^[3]  What’s missing?  For starters: Law.  Law is needed so people don’t kill each other.  And DOGE ignored it.

DOGE failed because of its hubris and concomitantly stupid assumptions about law.  For DOGE, there was a baseline assumption that all law is stupid, redundant, and inefficient. . . .

Some of it is.  Some law is primarily pretextual (a status ladder to give someone power over you under the pretext of accomplishing some social / contractual good).  That’s a risk for every legal right.  See id.  Some law is cancerous.  Id.  But the alternative is that people beat each other to death with clubs.  In reality, Law is to society as DNA is to known life.  See, e.g., “On Legal AI”, v.I (2019).

That’s what rights are, public and private.  You only hate them when they are someone else’s.

DOGE is like a doctor trying to cure cancer who doesn’t believe in genetics.  So he just chops off ill-appearing human limbs, randomly, like a medieval chirurgeon seeing something outside of his purview.^[4]

DOGE’s limits analysis does not acknowledge legal effects or structures, nor their complex interactions with other data dimensions.  Nor does DOGE’s rubric acknowledge or weight human suffering, kindness, or even the very human utility which both governance, and money, are ultimately designed to serve.

In this, the technology, the servant (here: limits analysis), has become the master.

That analytic distortion or wrong emphasis reduces human utility writ large, and actually kills people.  There are credible allegations that abruptly pulling anti-retroviral drugs from children born with Aids has now killed them.  (Overseas or not, they are humans and require our consideration.  Even the most selfish, isolationist PR person in the world would cringe at the negative feedback caused to the US by this perception.)  Whether those allegations are true or not, whether similar allegations and hypotheses are true are not (and public health is its own science—presently ignored in some cases), both deserve investigation of likely effects BEFORE action.  (Cutting life-saving drugs for babies is what Jeff Bezos calls “a one-way door”.)  And yes, Mr. Musk, they require diligence and test long before “launch”.  Maybe even prayer—but several billion times more fervent than with your first human flight launch, because: (a) scale exists, human-wise too; and (b) America is a reflection of the world that it influences and attracts.

In other words, an ultra physics or science-based approach to “making government more efficient” failed because it was commensurately ignorant + arrogant about the scientific environment it was trying to optimize in.  To be successful, DOGE needs to translate “law”, and other things, into its proprietary “physics”.  Otherwise, it will optimize a machine for certain failure across all data dimensions.  Instead of an airplane, one-dimensional flight will yield high-speed “space confetti”:  Destructive particles moving at vast speeds into human and societal life.  Like radioactive particles hitting strands of DNA: You will trigger social oncogenes instead of curtailing them.

WHAT TO LEARN

The above has been about teaching law as information object and algorithm to the zeitgeist—Silicon Valley in particular.  But, geography aside, “Silicon Valley” agile engineering style is core to DOGE’s modus operandi.

And the truth is, both SV and DOGE folks have something to teach US too.

America does need fiscal conservatism.  It does need efficiency gains (otherwise, Air America will just sit on the tarmac . . . or fly ten feet at infinite cost . . . or fly off a cliff and into a ditch).  See, e.g., Buffet’s comment above.  Limits analysis is useful.  But it has to be done IN ALL FOUR DIMENSIONS.^[5]  It has to consider such objective-subjective objects as human life and suffering, at all costs.  It must be empathetic, if still mathematically rigorous.

The truth is—as Americans and generally—we don’t grapple with the truth.  America forgot how to debate.  Instead, all points on the political spectrum prefer propaganda and intellectual auto-erotica to real time debate—because that is what makes media wallets and consumer brains fat, indolent, and fixed on idols instead of outcomes.

What America needs is empiricism, facts.  And collective debate about those facts, and where it wants to go: From Baylor to Berkeley.  (BTW Deporting one or both of those institutions, or any of their peers, will not, in fact, make America greater.  It will be a gift to its adversaries—who will embrace these valuable brands and creative boluses.)

We need all points of America’s compass.  We need you.  At your best, your most open-minded, your most empirical and courageous.  These three things are peas in a pod: Courage, empiricism, a willingness  + toughness to boldly encounter, and levelly analyze, wildly divergent hypotheses, evidence, and human data (which include legal data).

This may sound political; but it is really a note to neutralize politics in our governance engineering.

HOW TO SUCCEED AT DOGE AND INFLUENCE PEOPLE

To make a DOGE that is “successful” for all Americans—or at least a critical and vibrant weight thereof—it must first acknowledge that law is both the thing that enables societal life (and thus the life of the nation) as well as a potentially lethal font of inefficiency and dysfunction.  Whether spawned or distorted by sociopathy, entropy, or extreme asymmetry between parties (see, e.g, OLAI II), law may embody cancer too.  Indeed.  But it is nonetheless a necessary dimension that DOGE ignored.  And that’s why DOGE has thus far failed.  It has and continues to ignore the information structure and transfer that allows for societal life: again, just like cancer doctors denying the existence and role of DNA; or computer scientists ignoring speed and computability in algorithm design.  Results: Predictable.  Failure: Certain.

DOGE also failed due to its inability to consider the effects of its “Delta V” (or “rate of change”) on human life / well-being—like an autonomous taxi that (a) arrives early but (b) gives all of its passengers whiplash in the process.

But don’t just be a negative ninny.  It’s easy to critique.  And self-satisfying.  In contrast: Building is hard—also, about a thousand times more rewarding for you and those around you.  We must design a DOGE with the aim of making one that actually works.  We can learn from the “can do” of Silicon Valley without drinking its suicidal cool-aid (hubris, law-blindness, etc.).

For example: Silicon Valley’s attitude towards failure.  The above “fail state analysis” for DOGE is NOT built for the purpose of dumping on any person who assails this hard problem, nor to assume or pillory anyone for presumed mal-intent.  No.  It is the same as gathering data from a failed rocket launch, a “minimum lovable product” launched at dawn, etc.  Our analysis is empirical, yes; but bluntness aimed at progress.

Now, fast forward to: Us—and U.S.—design engineers: We grapple now with two dimensions.  Not just one.  To make OUR DOGE, we need to up the data dimensions a couple more times.  That model is next.  (Maybe call it “DOGUS”—with the “U” for (i) “utility”, (ii) “U.S.”, and (iii) “us”.)

___

 

COMING UP NEXT: ENGINEERING DHARMA

This note will synopsize a functional design for a working DOGE.  It is apolitical.

___

^[1] We can debate / discuss why in a subsequent article.

^[2] That font of random right wing rhetoric (no) Warren Buffet, stated that to run DOGE is a “job I don’t want, but it’s a job I think should be done”.  See, “Warren Buffett says cutting the US deficit is a difficult job — but it ‘should be done’”, “Business Insider”, John L. Dorman (May 3, 2025)(available online as of June 1, 2025 at https://www.businessinsider.com/warren-buffett-federal-deficit-unsustainable-doge-spending-2025-5)(“‘I think that bureaucracy is dangerously contagious, and it really doesn’t have any checks on it,’ he said. ‘I wouldn’t want the job of trying to correct what’s going on in revenue and expenditures of the United States with roughly a 7% gap, when probably a 3% gap is sustainable.’”).

^[3] “Technology” is not a dimension within that construct.  That’s because “technology” is just a tool for accomplishing “Work” efficiently over Time.  Thus, both technology and Work are subconstructs under “Time”.  Work and technology are critical constructs, but they are not dimensions we are trying to optimize for.

^[4] See generally, “The Emperor of All Maladies”, by Dr. Siddhartha Mukherjee, and citations therein, for background on the genetic nature of cancer.

^[5] Also: We have to define a destination or a destination set for the airplane, too (and periodically update it through pilots, turbulence, net and individual passenger preference, comfort, and storms).  To mix metaphors: Our design must further contemplate and leverage “legal genetic engineering” to cure legal cancer (or cancer caused by technical involutions, or human nature).  These are the kinds of dimensions upon which governance success abides and lives—relatively and in absolute terms.