The Larger World ArchiveClaims · Sources · Evidence
Open problems30 entries

Open Problems

Two kinds of unresolved question, kept apart on purpose: mathematical conjectures, where the state of play is exact and countable, and medical hypotheses, where the argument is over evidence that is messy by nature.

Unsolved. Sixteen mathematical conjectures with exact bounds, and fourteen medical hypothe
Unsolved. Sixteen mathematical conjectures with exact bounds, and fourteen medical hypotheses with what the trials show. Image: Unknown authorUnknown author · Public domain — Wikimedia Commons.

Two different meanings of "unsolved"

A conjecture is not a controversy. The Riemann Hypothesis has no camps — it has a statement nobody can prove, and a very large amount of confirming computation. A medical hypothesis usually has camps, because the evidence is observational, confounded and slow. Treating the two the same is how mathematics gets mistaken for opinion and medicine for certainty.

16Open problem
0Documented
14Live debate
0Fringe
0Debunked

The labels: Open problem — precisely stated, neither proved nor disproved; the note gives how far it has been checked. Documented — the phenomenon is on record. Live debate — real evidence on both sides, no settled answer. Fringe — no positive evidence. Debunked — tested and failed.

Where this collection overstates things

Six framings need correcting, and the entries above carry the correction rather than the original wording:

  • The abc conjecture is not a standoff between equal camps. Mochizuki's proof has not been accepted by the field; most specialists who examined it found a gap they could not resolve, and the debate is about whether the theory can be read at all.
  • The inscribed square problem is not open for "any continuous closed loop". It is false for curves that cross themselves; the open case is simple closed curves, and it is proven for smooth ones.
  • P vs NP would not straightforwardly "break encryption". Factoring is not known to be NP-complete; a proof of P ≠ NP would not by itself tell you factoring is hard.
  • Chronic Lyme is not a symmetrical disagreement. Prolonged antibiotics have been tested and failed, with documented harms — while post-treatment symptoms are real and deserve treatment.
  • The serotonin critique is not evidence that antidepressants do not work. The 2022 review is about the causal story psychiatry told the public, not about drug efficacy.
  • The metabolic theory of cancer: the Warburg effect is genuine and clinically used; the leap to "cancer is metabolic, not genetic" and to diet as treatment is not supported by trials.

One entry deserves more credit than a list usually gives it: the carbohydrate–insulin model is a serious hypothesis with controlled trials against it, not a fringe claim — it is wrong or incomplete in a way that is still scientifically productive.

Number theory and primes

Items 1–8 · 8

  1. 1
    The Riemann Hypothesis

    Every non-trivial zero of the zeta function sits on the line Re(s) = ½, which fixes how the primes are spaced.

    The most consequential open problem in mathematics. First 10 trillion zeros checked, all on the line — that is evidence, not proof.

    Open problem
  2. 2
    The Goldbach conjecture

    Every even number above 2 is the sum of two primes. Stated in 1742.

    Verified to 4×10¹⁸. The best partial result (Chen, 1973) gets to 'prime plus a number with at most two prime factors'.

    Open problem
  3. 3
    The twin prime conjecture

    Infinitely many prime pairs two apart.

    Not proven, but no longer hopeless: Zhang (2013) proved bounded gaps of 70 million, since cut to 246 by Polymath.

    Open problem
  4. 4
    The Collatz conjecture

    Halve evens, triple-plus-one odds; every starting number eventually reaches 1.

    Checked past 2⁶⁸ without a counterexample. Paul Erdős said mathematics is not yet ready for such problems.

    Open problem
  5. 5
    Legendre's conjecture

    There is always a prime between n² and (n+1)².

    Follows from the Riemann Hypothesis but is not known independently. Confirmed far out by computation.

    Open problem
  6. 6
    Odd perfect numbers

    Even perfect numbers are catalogued; whether any odd one exists is unknown.

    If one exists it exceeds 10¹⁵⁰⁰, has at least 75 prime factors and a specific algebraic shape — heavy constraints, no exclusion.

    Open problem
  7. 7
    The abc conjecture

    A relation between the prime factors of a, b and a+b.

    Mochizuki's 2012 proof has not been accepted: most specialists, including Scholze and Stix, found a gap they could not resolve.

    Open problem
  8. 8
    Beal's conjecture

    If Aˣ + Bʸ = Cᶻ with exponents above 2, then A, B and C share a prime factor.

    A million-dollar prize problem. Implies Fermat's Last Theorem as a special case.

    Open problem

Logic, computation and geometry

Items 9–16 · 8

  1. 9
    P versus NP

    Can every problem whose answer is quickly checkable also be quickly solvable?

    The famous asymmetry: factoring is easy to check and believed hard to solve, but is not known to be NP-complete.

    Open problem
  2. 10
    Navier–Stokes existence and smoothness

    Do smooth solutions to the 3D fluid equations always exist, or can they blow up?

    Millennium Prize problem. Numerical blow-up hints exist; no proof either way.

    Open problem
  3. 11
    Yang–Mills and the mass gap

    A rigorous quantum field theory for the strong force in which particles have positive mass.

    Millennium Prize problem. Physics works; the mathematics that would justify it does not exist yet.

    Open problem
  4. 12
    The Hodge conjecture

    Certain topological classes on complex varieties are built from algebraic cycles.

    Millennium Prize problem. Known in low dimensions and for special classes.

    Open problem
  5. 13
    Birch and Swinnerton-Dyer

    The number of rational points on an elliptic curve is governed by its L-function at 1.

    Millennium Prize problem. Proven for ranks 0 and 1 in special cases; the general statement is open.

    Open problem
  6. 14
    The moving sofa problem

    The largest shape that can be turned around a right-angled corner of unit-width hallway.

    Gerver's 2.2195-area sofa is the best known; a 2024 proof of its optimality is under review, not settled.

    Open problem
  7. 15
    The inscribed square problem

    Every closed curve contains four points forming a square.

    True for smooth curves (proven 1929). False for general non-Jordan curves — the open case is curves without self-intersection.

    Open problem
  8. 16
    The lonely runner conjecture

    Runners at distinct constant speeds on a circular track: each is eventually at least 1/n from every other.

    Proven for up to seven runners; open in general.

    Open problem

Neurology and psychiatry

Items 17–24 · 8

  1. 17
    The Alzheimer's pathogen hypothesis

    Plaques are an immune response trapping herpes or gum bacteria, not the primary cause.

    Itzhaki's case is taken seriously. But the amyloid-clearing antibodies that slow decline are hard to reconcile with amyloid as mere bystander.

    Live debate
  2. 18
    Type 3 diabetes

    Alzheimer's as insulin resistance inside the brain.

    Brain insulin signalling is genuinely impaired. Intranasal insulin trials have mostly failed to produce clinical benefit.

    Live debate
  3. 19
    The serotonin critique

    Depression is not caused by low serotonin — the 2022 umbrella review found no consistent evidence of it.

    The review is solid and widely accepted. It does not show antidepressants are ineffective; those are separate questions.

    Live debate
  4. 20
    The viral trigger for schizophrenia

    Prenatal infection and immune activation raise risk in genetically vulnerable brains.

    Epidemiology is strong (flu, rubella, Toxoplasma seasons). No single pathogen, and no mechanism that predicts who is affected.

    Live debate
  5. 21
    PANDAS and PANS

    Sudden OCD and tics in children from strep-triggered antibodies attacking the basal ganglia.

    PANS exists as a described syndrome. The strep-specific mechanism and antibiotic treatment remain unproven in trials.

    Live debate
  6. 22
    Chronic Lyme versus post-treatment syndrome

    Persistent symptoms after antibiotics: hidden bacteria, or lasting immune damage?

    Post-treatment Lyme disease syndrome is real and recognised. Trials of prolonged antibiotics have shown no benefit and real harm.

    Live debate
  7. 23
    The chemical imbalance critique

    Psychiatry's public story about neurotransmitters was oversimplified.

    True as history. The replacement — inflammation, plasticity, network dynamics — is a research programme, not a settled model.

    Live debate
  8. 24
    The infection hypothesis of Parkinson's

    Gut and nasal pathogens reaching the brain years before tremor.

    Braak's staging and the vagotomy data are suggestive; no pathogen has been identified as causal.

    Live debate

Metabolism, cancer and chronic illness

Items 25–30 · 6

  1. 25
    The carbohydrate–insulin model

    Refined carbohydrate drives insulin, which traps energy in fat cells, so obesity is hormonal rather than caloric.

    Insulin matters; whether it dominates over total energy balance is unresolved. Controlled feeding trials favour the energy-balance account.

    Live debate
  2. 26
    The Warburg / metabolic theory of cancer

    Cancer begins as mitochondrial dysfunction, not primarily nuclear mutation.

    The Warburg effect is real and used in imaging. 'Cancer is metabolic, not genetic' and its ketogenic-diet therapies are not supported by trials.

    Live debate
  3. 27
    Seed oils and inflammation

    High omega-6 linoleic acid from industrial oils drives chronic disease.

    Randomised trials replacing saturated fat with these oils lower LDL and cardiovascular events. The mechanistic worry has not shown up in outcomes.

    Live debate
  4. 28
    Lean mass hyper-responders

    Very high LDL on low-carb diets may not carry the usual plaque risk when triglycerides are low and HDL high.

    A real phenotype, documented in case series. Whether the risk is genuinely lower is unknown; plaque imaging studies are ongoing.

    Live debate
  5. 29
    Leaky gut and autoimmunity

    Increased intestinal permeability drives lupus, Hashimoto's and rheumatoid arthritis.

    Zonulin and permeability changes are measurable in disease. Causality, and whether repairing the barrier helps, is not established.

    Live debate
  6. 30
    Mast cell activation syndrome

    Overactive mast cells explain multisystem allergy-like illness — or the diagnosis is over-applied.

    Consensus criteria exist (2020). Diagnosis rates vary enormously between clinics, which is itself the argument.

    Live debate

Sources

Every bound quoted above is from the cited paper, not from the claim as it was submitted.