L–The Bias I Carry
How the 2025 Nobel Prize in Physics forced me to recalibrate my own detector
Context
In October 2025, the Royal Swedish Academy of Sciences awarded the Nobel Prize in Physics jointly to John Clarke, Michel H. Devoret, and John M. Martinis
“for the discovery of macroscopic quantum mechanical tunnelling and energy quantisation in an electric circuit”.
The work being honored dates back to the mid-1980s. The laureates showed that a carefully engineered superconducting circuit could exhibit macroscopic quantum tunnelling and discrete energy levels: a designed electrical circuit behaving as a quantum object in its own right.
This essay is not about the technical details of those experiments. The Nobel scientific background, original Physical Review Letters papers, and modern lecture notes cover them far better than I could.
This essay is about the bias I noticed in myself when I first read the announcement—and about the work required to correct it.
Why this essay matters
For me, this text is an accountability exercise. I want my students—and my children—to see a version of me who can notice bias, name it publicly, and change course.
For readers, I hope it functions as a case study. Most scientists recognize tribal instincts in themselves, but rarely see someone walk through a correction process in detail.
For the field, the stakes are larger. Scientific credibility depends not only on technical rigor, but on our ability to evaluate claims by content rather than clan. Expertise does not immunize us against bias. If anything, it equips bias with better arguments. The question is not whether bias exists, but what counter-systems we build.
My core thesis is simple:
Tribal bias in science rarely forms through dramatic conversion experiences. It accumulates through hundreds of small, ambiguous encounters—each forgettable on its own, collectively load-bearing. Correction requires systems, not willpower.
Three anchors eventually pulled me back:
Technical respect – Was the physics actually good?
Institutional trust – Did the Committee make a defensible choice within its constraints?
Moral responsibility – What am I modeling for students and children?
1. The Measurement
On an ordinary October afternoon—between an email about lab procedures and a student asking why the laser would not lock—I saw the headline:
“Nobel Prize in Physics 2025: John Clarke, Michel H. Devoret, John M. Martinis.”
Before I finished reading the citation, a thought surfaced that had nothing to do with physics: “Really? The community that already raised billions for the ‘platform of the future’ now also gets a Nobel Prize?”
My jaw tightened. I did not think about Josephson junctions or escape rates. I thought about startups, glossy keynotes, funding asymmetries. I thought about our corner of our fundamental research field—trapped ions—where progress often felt slower, less spectacular, and somehow less rewarded.
Only a moment later did a quieter realization arrive: this reaction did not fit the scientist I like to imagine myself to be. I tell students to read carefully before judging. I tell my children to judge people by what they do, not by where they come from. I had done neither. I had reacted to a symbol.
The Nobel announcement functioned like a sharp measurement pulse on a system I usually prefer not to observe too closely: my own pattern of tribal thinking.
2. The Formation
I cannot point to a single moment when the tribal map in my head formed. That is precisely the problem.
Bias does not usually announce itself with dramatic conversion experiences. It accumulates—through hundreds of small, ambiguous encounters, each individually forgettable, none individually sufficient to warrant a change in self-image. Over time, they crystallize into something solid enough that we call it “intuition”.
2.1 Ion-trap socialisation
I was trained in atomic physics, specifically in the ion-trapping community. That meant nights spent worrying about dark counts, micromotion minimization, heating rates and secular frequency drifts. It meant valuing:
explicit noise budgets,
conservative interpretation,
slow accumulation of control,
coherence demonstrated through long, careful sequences rather than a single spectacular plot.
These instincts are not wrong. They made me a better experimentalist.
But there is a trap hidden inside this carefulness:
When you are trained in one particular way of being careful, it becomes easy to mistake other ways of being careful for being careless.
With some distance, I also see how comforting the ion-trap self-image can be. We tell ourselves we are conservative because we care. Sometimes that is true. Sometimes it is a shield—used to dismiss faster-moving work, to excuse our own forms of hype (“ultimate fidelity,” “metrological relevance”), or to maintain citation and authority boundaries that feel justified from the inside. Tribal narratives rarely feel like tribalism when they flatter us.
2.2 The accumulation
Bias did not arrive fully formed. It accumulated.
I remember one moment precisely because it felt trivial at the time. A late-day conference talk. Slides advancing faster than I could parse the error analysis. An impressive improvement summarized in a single plot; uncertainty handled in a sentence. A question about a control I would have considered mandatory was answered quickly—“standard”—with mild impatience.
Nothing about this justified condemnation. What I actually had was a feeling and a missing plot. Yet later, in the hallway, I caught myself retelling the talk as “typical overselling.”
What made this moment matter was not the moment itself. Over the following years, variations of it repeated: a paper I skimmed because the abstract already irritated me; a funding announcement that felt disproportionate to what I had personally verified; a keynote that blended careful physics with a narrative of inevitability. None of these encounters would have justified a verdict on their own. Taken separately, each could have ended with “I don’t know enough to judge.”
But they did not remain separate. Over five years, a dozen such moments—none decisive, none dramatic—compressed into something that felt like intuition. By the time I noticed it, the intuition no longer felt provisional. It felt earned. That is how accumulation works: duration substitutes for evidence, and familiarity masquerades as certainty.
This is the cognitive structure of confirmation bias: interpreting new information in ways that are partial to existing beliefs, expectations, or hypotheses. The bias does not require ill will. It only requires asymmetric weighting.
2.3 Compression into symbols
Over time, entire communities collapsed into symbols. A name on a slide stood in for a platform. Work I had not read became “probably overstated.” People I had never met became “one of them.”
Evidence that fit the pattern was weighted heavily; counter-examples became exceptions. The fundamental attribution error did the rest: similar behavior received different explanations depending on group membership.
Once the map existed, a feedback loop closed. Papers from our side were read with presumed good faith. Papers from their side were read with active suspicion. I believed I was enforcing rigor uniformly. In practice, I was enforcing it asymmetrically.
The compression saved cognitive effort. It also destroyed nuance.
2.4 Internal splits
I do not name individuals here, both to protect relationships and because the pattern is not about specific people but about my own interpretive lens.
The map did not only separate platforms. It also separated trajectories within my own.
Former colleagues trained in the same ion-trap culture went on to found or join startups pursuing specific architectural paths toward quantum computation. I felt pride that the platform was taken seriously—and an unease I could not quite name.
That unease sometimes expressed itself as distance. I found myself categorizing people not by shared technical roots, but by their new institutional contexts: those still exploring versus those now executing on a defined bet.
This was another form of tribal thinking. I knew—intellectually—that these colleagues were acting responsibly within their constraints. A company optimizing for a specific architecture is not intellectual narrowing; it is institutional necessity. Their choices reflected different optimization functions: speed and focus over breadth and exploration. I respected the difficulty of those choices. I trusted their technical judgment.
But I still felt the distance. And rather than naming it as my discomfort with diverging constraints, I sometimes coded it as their narrowing—as if my mode of working were somehow more intellectually honest than theirs.
The problem was not their choices. It was my need to interpret different constraints as different virtues.
What followed was the work of recalibration.
3. Anchor One — Technical Respect
Technical respect is supposed to be my epistemological foundation. Whatever my preferences or loyalties, I like to believe that careful, difficult physics receives my respect when I actually look at it.
After my initial reaction to the Nobel, I did what I should have done immediately: I started thinking.
3.1 What actually cracked
A seminar tailored for young students did not cause the crack—it revealed it. What stayed with me was not a rhetorical point, but a scene of three people working in a university basement, surrounded by hand-labeled electronics boxes and coaxial cables coiled like garden hoses. On a wall hangs a hand drawn sketch showing control lines disappearing into copper-grain filters.
I had walked into the room expecting a familiar origin story: early genius, inevitable platform, straight line to today. Instead, I found university physics—messy, local, and preoccupied with dissipation rather than destiny. I realized I had been projecting a 2025 ecosystem onto a 1985 experiment.
3.2 What they actually achieved
The core question was simple and deep:
Can a macroscopic degree of freedom—describing the collective behavior of a huge number of electrons—show unmistakably quantum behavior, if you control its environment well enough?
The Caldeira-Leggett framework had quantified how environmental dissipation suppresses quantum tunnelling and decoherence. It predicted that for sufficiently small damping, the escape rate of a system from a potential well would saturate at low temperatures: instead of going to zero as thermal activation shuts down, it would approach a constant “quantum tunnelling floor”.
The Berkeley experiments built a device where the relevant coordinate was the superconducting phase across a Josephson junction, and the potential was engineerable by circuit parameters. The task was to cool and isolate this effective coordinate enough that the predicted macroscopic quantum tunnelling could actually be observed.
That required several layers of experimental ingenuity:
Impedance engineering. The environment seen by the junction at relevant frequencies had to effectively present a very high impedance, suppressing dissipation. This meant carefully designed transmission lines, on-chip and off-chip filtering, and a detailed understanding of parasitic modes in the wiring.
Filtering and shielding. High-frequency noise from room-temperature electronics and thermal radiation had to be absorbed and thermalized in stages. Copper-powder filters, attenuators at different temperature stages, and nested shields all played roles.
Measurement without destruction. The readout scheme had to detect whether the phase had escaped the well without itself dumping energy into the system in a way that dominated the escape dynamics.
The measurements then tested the escape rate as a function of temperature. At higher temperatures, the escape was consistent with thermal activation: the rate decreased as the bath cooled. Below a crossover temperature, the rate stopped decreasing and approached a plateau consistent with the quantum-tunnelling expectation within the modeling framework.
A second set of experiments used microwave spectroscopy to probe energy levels in the effective potential. When the applied microwave frequency matched the spacing between quantized levels, the escape rate changed in a way that revealed discrete structure: the circuit behaved like an artificial atom, with engineered energy spacings determined by capacitances and inductances.
None of this is exotic by today’s standards. For those of us who teach circuit QED and transmon physics, the language has become standard. But in 1985, this was a conceptual leap: circuits were not just passive elements connecting quantum systems. They were the quantum systems.
3.3 The lineage I had ignored
As I followed the references, a lineage emerged. Techniques pioneered there—environmental modeling, filtering strategies, impedance control—now underpin fragile quantum experiments across platforms, including my own.
A technique’s half-life—how long it remains in active use—is a reasonable measure of impact. Here, that half-life spans decades.
3.4 The first anchor pulls
All of this left me with a conclusion:
I had dismissed the 1980s superconducting work without ever looking at it with the same care I demand from others.
My technical respect had been conditional. Once that was clear, one anchor engaged: whatever I thought about later ecosystems, the physics honored by the Nobel was hard and beautiful work.
This is hard and beautiful work.
That realization led naturally to the next question: if the physics was this good, what about the institution that chose to recognize it?
4. Anchor Two — Institutional Trust
Understanding the Nobel Committee’s decision required something I had not initially granted them: the benefit of the doubt.
4.1 The impossible job
The Nobel Committee faces a genuinely impossible task:
assess decades of intertwined contributions, from theoretical predictions in the 1980s to experimental breakthroughs and platform development across multiple communities,
compress a complex, collaborative history into exactly three names, due to the statutory limit,
distinguish between foundational discoveries and later applications,
decide how much weight to give different platforms and schools, and
do all of this under conditions of incomplete information, with decisions that will be scrutinized indefinitely.
My initial reaction implicitly assumed that, in this case, they had simply capitulated to hype: selected the platform with the largest current industrial profile and retrofitted a foundational story.
After engaging with the 1980s work, that assumption became harder to justify. To maintain it, I would have to claim either that the Committee was ignorant of the field’s history or that it knowingly ignored that history in favor of geopolitical or economic considerations.
That is a strong accusation. It demands evidence. I have none.
4.2 What the citation actually says
The citation itself is concise:
“for the discovery of macroscopic quantum mechanical tunnelling and energy quantisation in an electric circuit”.
The choice of words is deliberate:
“Discovery” – The focus is on revealing a new phenomenon or regime, not on building a device for a specific application.
“Macroscopic quantum mechanical” – This is framed as a fundamental physics question: can quantum mechanics manifest in suitably engineered macroscopic variables?
“Electric circuit” – The system is a designed artifact, not a naturally occurring atom or molecule. The move from “nature-given quantum system” to “engineered quantum system” is the conceptual pivot.
Equally important are the omissions. The citation does not mention:
qubits,
quantum computers,
specific architectures like circuit QED,
any work from the 2000s and 2010s on coherent control, error correction, or scaling.
Viewed in this way, the Committee appears to have anchored the prize firmly in the mid-1980s. I say 'appears' because the Committee's deliberative reasoning is not public beyond the citation itself. This opacity is understandable given historical precedent—but it is also, I will argue, correctable. They chose to recognize the moment when an electric circuit first earned the status of a legitimate quantum system in its own right, in close dialogue with the Caldeira-Leggett framework that had previously been recognized in part through Anthony Leggett’s 2003 Nobel.
One can certainly debate whether 1985 is the right place to draw that line. Yasunobu Nakamura’s 1999 demonstration of coherent oscillations in a superconducting charge qubit, for example, is also an obvious milestone, and many would argue it deserved recognition. Other contributors to flux qubits and circuit QED could also reasonably appear in alternative histories.
But “debatable” is not the same as “obviously wrong”. My initial reading—“they validated the hype platform”—had ignored the internal logic of the citation entirely.
4.3 What institutional trust requires
Institutional trust does not mean blind acceptance. The Nobel system has structural flaws:
it over-compresses collaborative science into a small number of names,
it tends to reflect existing power structures,
and its long time lag can mismatch scientific and societal timing.
Recognizing these flaws is compatible with granting that committees can, within constraints, act conscientiously.
The second anchor, then, was modest but important:
Disagree if you must—but do not substitute verdict for analysis.
That realization pushed me toward the third anchor: responsibility.
5. Anchor Three — Responsibility
By the time technical respect and institutional trust had partially corrected my view of the 2025 prize, another question had become unavoidable:
What am I teaching, explicitly and implicitly, when I talk about such things in front of students and my own children?
5.1 Students are always watching
There are lab meetings and seminars I can no longer reconstruct in detail, but I remember the pattern.
A student would present a paper from “our” platform. My opening questions would be technical, sometimes sharp, but framed within an atmosphere of shared trust: we assume the authors are trying to do good work; we are collectively trying to probe limitations.
On other days, a student would present a paper from “the other side”. My questions, in those cases, carried a subtly different edge. Before the first plot had been fully discussed, I might probe the choice of metric, the absence of a particular benchmark, or the way an abstract was worded.
Each individual question was defensible. Many of them were the kinds of questions we should ask in all cases. The problem was not the content. It was the asymmetry.
Students learn more from tone and pattern than from formal statements. Whatever I might have said about “evaluating science on its merits”, my behavior taught a different lesson:
work from our tribe is presumed careful until proven otherwise;
work from their tribe is presumed inflated until proven otherwise.
The more senior I became, the more my questions could shape the room. A skeptical remark from me could license a whole group to treat a platform dismissively. In that sense, I was not only carrying bias. I was transmitting it.
5.2 Children are sensitive detectors
Explaining my reaction to my children collapses the rhetoric quickly.
"How did you know it was unfair?"
I can add layers of complexity about credit, structural inequalities, and the distribution of resources among platforms. Those are real issues. But at the core, my initial reaction did not rest on this analysis. It rested on who I associated with the prize, not what they had done.
Children are very good at detecting such inconsistencies. They do not need to know the details of superconducting circuits to notice when the principle “judge by content” is not being applied.
5.3 What I owe
Once all of this was clear, the third anchor—the moral one—began to pull.
There are at least four responsibilities I cannot outsource to institutions:
Model intellectual honesty.
Not the performance of always being right, but the more uncomfortable practice of saying, in front of others: “I was wrong about this. Here is how I know. Here is what I am changing.” Private correction teaches no one. Public correction teaches that changing one’s mind is part of the job.
Model bias-correction protocols.
It is not enough to exhort students to “avoid bias”. They need to see concrete methods: how to notice an asymmetric reaction, how to check whether it tracks content or source, how to adjust reading and questioning habits accordingly.
Model humility about expertise.
Technical expertise in one platform or subfield does not grant global authority. It certainly does not immunize against motivated reasoning. Admitting the specific limits and distortions of one’s perspective is a form of accuracy, not weakness.
Model generosity as default.
This does not mean naivety. It means starting from the assumption that colleagues in other communities also care about correctness, and that differences in language, emphasis, or style often reflect different local constraints rather than moral failings.
Philosophers writing about intellectual humility describe it as a disposition to recognize and respond appropriately to one’s cognitive limitations. In this context, the limitation is not a lack of technical knowledge, but a tribal weighting of evidence.
5.4 Generational stakes
The stakes are generational.
My students will inherit more than the equipment and the analysis scripts. They will inherit:
the credit practices I model in talks and papers,
the standards I apply in peer review,
the way I speak about competing approaches in informal settings.
My children will inherit a scientific culture that they may or may not choose to engage with. Whether that culture is capable of collaboration across platform boundaries, whether it commands public trust, and whether it pairs expertise with humility rather than tribal armor—these are not abstractions. They shape which problems can be tackled and by whom.
Individual correction is necessary. It is also not sufficient. The bias machine will not stop running because one person writes an essay. The question then becomes: what systems can we put in place to make bias-correction more likely, more systematic, and more robust than a single person’s resolve?
6. Protocols — Building Better Systems
I do not trust my willpower. I will likely fail at this again. I need protocols—structures that introduce friction between bias and action.
6.1 Individual protocols
These are small, personal systems I am trying to install.
(1) Notice before acting.
When I encounter a result from a community I have historically distrusted, I try to insert one pause:
Notice the immediate emotional reaction.
Ask explicitly: “Is this reaction about what is written, or about who wrote it?”
If the answer is “both” or “mostly the latter”, commit to reading the technical content before forming a verdict.
This does not eliminate bias. It creates a small time window in which the first anchor—technical respect—can engage.
(2) A quarterly tribal audit.
Every few months, I look back at the papers and preprints I have actually read carefully (not just skimmed). I ask:
How many are from my own platform or close collaborators?
How many are from competing platforms or communities?
Where did I invest effort in understanding, and where did I dismiss quickly?
If the distribution is sharply skewed, I deliberately choose a handful of influential or widely discussed papers from outside my comfort zone and read them as if they were from my own group: with patience and charitable interpretation of ambiguity.
You cannot fix a bias you refuse to measure.
(3) Credit discipline.
In my own writing and talks, I try to be more systematic about:
citing predecessors across platforms, not just within my own,
acknowledging when methods or concepts we use descend from work in other communities, and
contextualizing our results relative to others even when the comparison is uncomfortable.
This is partly about fairness. It is also about making visible the actual interdependence of platforms and techniques.
(4) Pedagogical check.
Before I make a critical remark about someone else’s work in front of students, I ask:
“Would I be comfortable saying this, in this tone, if the authors were in the room—or if my children were listening?”
If the answer is no, I rephrase or remain silent until I can express the concern in a way that targets claims rather than clans. Science requires criticism. It does not require contempt.
6.2 Institutional interventions
Individual protocols are fragile. They degrade under stress and time pressure. Institutions can help by embedding some of the necessary friction structurally.
Prize committees—not only the Nobel—can publish more about their reasoning, including the trade-offs and constraints they faced. In the case of the 2025 Physics Prize, the citation's careful wording is interpretable as a deliberate focus on foundational discovery rather than present-day platforms. But 'interpretable' is not the same as 'transparent.' A brief accompanying statement—explaining why this boundary was drawn, what alternative framings were considered, and how the statutory three-name limit shaped the outcome—would not compromise deliberative integrity. It would strengthen public trust by showing that the decision emerged from principled reasoning under constraints, not from opaque consensus or external pressure. Transparency does not require revealing votes or internal disagreements. It requires showing the structure of the problem the Committee believed it was solving.
A few more examples:
Funding agencies can explicitly incentivize platform-crossing collaborations and comparative benchmarks, and allocate dedicated resources for shared infrastructure: error standards, calibration procedures, open reference implementations.
Conferences can design sessions by problem (“decoherence in many-body systems”) rather than by platform, and normalize sessions on limitations and failures alongside headline results. Mixed-platform panels force cross-reading.
Journals can require authors to contextualize work across platforms where relevant, not only within their niche, and assign at least one reviewer from a different community when possible. Transparent editorial policies about reviewer selection can reduce suspicion of tribal gatekeeping.
Universities and departments can adjust evaluation criteria to value infrastructure work, shared tools, and teaching of bias-correction practices as real contributions, making it less risky for individuals to invest in them.
None of these interventions removes conflict or competition. They simply adjust the incentives so that truth-seeking and cross-platform respect are less costly.
6.3 The universal pattern
Although this essay focuses on a specific corner of quantum physics, the underlying pattern is not unique.
Similar dynamics appear in:
debates between schools in high-energy theory,
methodological battles in neuroscience,
replication disputes in psychology,
and many other places where scientific questions intersect with funding, status, and identity.
Social identity theory and group polarization are general frameworks, not peculiarities of our field. Our minds evolved to maintain coalitions, not to optimize likelihood functions. We are typical social primates trying to do something unusually difficult: seek robustly correct answers in environments where status and resources are unevenly distributed.
Recognizing this is not an excuse. It is a design challenge.
That leads me to the conclusion.
7. Closing — The ongoing work
I return, finally, to the three anchors that structured this essay:
Respect – The 1980s experiments honored in the 2025 Nobel Prize represent careful, conceptually important physics. My initial refusal to grant that was a failure of curiosity, not a justified critique.
Trust – The Nobel system is structurally imperfect, but in this case the Committee’s framing—honoring foundational discovery rather than present-day industrial platforms—is intelligible and defensible, even where one might have chosen differently.
Responsibility – Students and children learn more from how we handle our mistakes than from how we present our successes. If I want them to judge work by content rather than clan, I have to show them how to correct tribal judgments in myself.
I do not expect this essay to fix my bias once and for all. It is not a confession followed by absolution. It is, at best, a snapshot of a calibration process in progress.
But the stakes extend beyond individual correction.
Tribalism is not just a moral failing. It is a capability limiter—and the cost is measurable.
Consider what actually blocks progress in quantum systems: surface anomalies that cause heating, material interfaces that introduce noise, timing jitter that limits synchronization, readout cross-talk that corrupts measurements. These are not platform-specific problems. They are shared problems that happen to express differently in ions versus superconducting circuits versus neutral atoms versus spin qubits.
When tribal boundaries make cross-platform collaborations rare, we duplicate effort. When funding mechanisms punish researchers for working across platform lines, we slow discovery. When graduate students absorb the message that "those people do sloppy work" or "that platform is a dead end," we train the next generation to ignore solutions that already exist.
Platform monocultures are not just intellectually limiting. They make us slower at solving the problems we claim are urgent. The hardest challenges ahead—error correction at scale, integration with classical control, real-world deployment under non-ideal conditions.
So here is the invitation:
If you recognized something of yourself in this dissection—if you have felt the tribal pull when another community wins a prize, secures a grant, or dominates a conference—then treat this as a calibration exercise, not a verdict.
You do not need to abandon platform loyalty. You do not need to pretend all approaches are equally promising. You do need to install counter-systems that pull against tribal gravity when it distorts your judgment.
We will not eliminate bias. But we can design better detectors, and we can make it normal for those detectors to be recalibrated in public.
Special thanks to all my past, current, and future environments.
References
Nobel Prize in Physics 2025 – Press Release and Summary. NobelPrize.org (2025). “for the discovery of macroscopic quantum mechanical tunnelling and energy quantisation in an electric circuit”.
Scientific Background: The Nobel Prize in Physics 2025. Royal Swedish Academy of Sciences (2025).
Martinis, J. M., Devoret, M. H., & Clarke, J., Energy-Level Quantization in the Zero-Voltage State of a Current-Biased Josephson Junction, Phys. Rev. Lett. 55, 1543 (1985).
Devoret, M. H., Martinis, J. M., & Clarke, J., Measurements of Macroscopic Quantum Tunneling out of the Zero-Voltage State of a Current-Biased Josephson Junction, Phys. Rev. Lett. 55, 1908 (1985).
Nickerson, R. S. (1998). Confirmation Bias: A Ubiquitous Phenomenon in Many Guises. Review of General Psychology, 2(2), 175–220. doi:10.1037/1089-2680.2.2.175
Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The Social Psychology of Intergroup Relations (pp. 33–47). Monterey, CA: Brooks/Cole.
Tajfel, H., Billig, M. G., Bundy, R. P., & Flament, C. (1971). Social categorization and intergroup behaviour. European Journal of Social Psychology, 1(2), 149–178.
Sunstein, C. R. (2002). The Law of Group Polarization. Journal of Political Philosophy, 10(2), 175–195. doi:10.1111/1467-9760.00148
Last updated