{"id":"603d449a-ae83-4f72-a492-eba6d2b37898","arxiv_id":"2505.08116","paper_version":4,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The upper edge of the DSSYK spectrum, in a triple scaling limit, exactly reproduces the matrix model of two-dimensional de Sitter JT gravity.","lead":"This paper shows that the top edge of the energy spectrum in a simplified quantum gravity model, the double-scaled SYK model, produces an exact description of de Sitter space in two dimensions. The result gives a concrete, tractable example of quantum gravity in de Sitter space, a long-standing challenge for string theory and holography.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"All-genus dS JT identification depends on the unproved discrete-to-continuous Weil-Petersson limit (3.28); if it fails, the higher-genus match collapses.","rationale":"The reader identified the reliance on the unproved limit (3.28) as the weakest assumption, and my stress-test reaches the same conclusion. The disk-level derivation is clean and does not depend on (3.28), but the all-genus identification is exactly where the paper moves beyond a conjecture. The paper explicitly calls (3.28) 'observed' rather than proven, and then uses it as the bridge from the DSSYK topological recursion to the JT/dS JT topological recursion. The concern is specific and addressable: a direct check of low-genus cases or a proof would settle it. Since the reader's verdict is already CONDITIONAL and this concern is the same one, I do not recommend changing the verdict. The result is genuinely interesting and non-circular, but the higher-genus claim should be presented as conditional on (3.28) until that limit is established.","tokens_in":13016,"tokens_out":13303,"duration_ms":126778,"concrete_test":"Compute both sides of (3.28) for the first non-trivial cases, e.g. (g,n)=(1,1) and (0,3), using the topological recursion (2.21)-(2.25) to obtain N^±_{g,n}(b) from ω_{g,n}; take λ = 10^{-2}, 10^{-3} with b_i = \\bar b_i / λ and compare with N^{2-2g-n} λ^n V_{g,n}(\\bar b_i). If the ratio tends to 1 as λ→0 for all tested (g,n), the concern is resolved; if not, the all-genus match is invalid. An independent analytic proof of (3.28) would be stronger.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The disk-level match in Eqs. (4.5)-(4.6) is explicit and convincing. The all-genus claim, however, is inherited from the lower-edge JT limit through the relations (4.14)-(4.17), and that limit is justified only by Eq. (3.28): N^±_{g,n}(b_1,...,b_n) -> N^{2-2g-n} λ^n V_{g,n}(\\bar b_1,...,\\bar b_n) for b_i = \\bar b_i / λ. The paper states this as 'observed in [42]' and does not prove it; Section 4 then uses it as the guarantee that ω_{g,n} reduces to W_{g,n} and \\tilde W_{g,n}. This is a nontrivial statement about exchanging a discrete sum over b_i ∈ Z_+ with a continuum limit, and it must hold for every (g,n) appearing in the genus expansion. If it fails at any order, the coefficient of e^{-(2g-2+n)\\tilde S_0} \\tilde Z_{g,n} in (4.16) will not equal the dS JT result, so the claim that DSSYK reproduces dS JT 'at all orders' is not supported. The disk amplitude cannot test this, since (g,n)=(0,1) does not involve (3.28). The central claim is therefore conditional.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper argues that the double-scaled SYK model, through its ETH matrix-model description, reproduces Jackiw-Teitelboim gravity with a positive cosmological constant (de Sitter JT gravity) when one takes a triple scaling limit around the upper edge E = E0 of the DSSYK spectrum. After reviewing the known reduction to AdS JT gravity from the lower edge E = -E0, the author performs the analogous limit around E = E0, obtains the disk amplitude (4.5)-(4.6) that matches the de Sitter JT disk amplitude of Cotler and Jensen, and lifts this to a proposed all-genus relation (4.14)-(4.17) with a complex effective coupling \\tilde S_0 = i S_0. The paper also connects the upper-edge limit to the classical solutions of sine dilaton gravity, where the θ = π solution gives the metric of −AdS2, interpreted as dS2.","tokens_in":13251,"tokens_out":3315,"duration_ms":34030,"significance":"If correct, this result would provide a concrete, parameter-free embedding of de Sitter JT gravity into a well-studied microscopic model (DSSYK), realized as an unstable saddle of the ETH matrix model. The disk-level match is explicit, clean, and checked against an independently defined matrix model of dS JT gravity, which is a genuine strength. The paper is also careful to distinguish its construction from other dS-from-DSSYK proposals and to state open questions. However, the all-genus claim is not established by the present manuscript: it relies on a discrete-to-continuous Weil-Petersson volume limit that is quoted from previous work as an observation rather than proved here.","major_comments":[{"comment":"The all-genus identification is load-bearing and rests on the limit N±_{g,n}(b_1,...,b_n) → N^{2-2g-n} λ^n V_{g,n}(\\bar b_1,...,\\bar b_n) stated in Eq. (3.28) as 'observed in [42]'. The paper then uses this limit as the guarantee in Eq. (4.19) that ω_{g,n} reduces to W_{g,n} and \\tilde W_{g,n}. This is a nontrivial statement about exchanging a discrete sum over b_i ∈ Z_+ with a continuum limit, and it must hold for every (g,n) in the genus expansion. If it fails at any order, the coefficient of e^{-(2g-2+n)\\tilde S_0} in Eq. (4.16) will not equal the dS JT result, so the claim that DSSYK reproduces dS JT 'at all orders' is not supported by the arguments given. The disk-level match (g,n)=(0,1) does not test this limit. Please provide a proof of (3.28) or a precise reference to a proof; if neither is available, the all-orders claim should be explicitly framed as conditional.","section":"Sec. 4, Eq. (4.19); Sec. 3.2, Eq. (3.28)"},{"comment":"The relation W_{g,n}(i z_1,...,i z_n) = (-1)^{3g-3+n} \\tilde W_{g,n}(z_1,...,z_n) is stated to be provable inductively using the topological recursion, but the proof is not included. Since this relation is the key step connecting the upper-edge limit to the dS JT genus expansion in Eq. (4.14), please include the induction argument or provide an explicit reference where it is proved.","section":"Sec. 4, Eq. (4.13)"},{"comment":"The decomposition of N_{g,n} into N^+_{g,n} + N^-_{g,n} and the equality N^+_{g,n} = N^-_{g,n} for even b_i are asserted without derivation. The text also does not state what happens to odd b_i contributions in the scaling limit. These details matter because the final genus expansion sums over all b_i ∈ Z_+ in Eq. (2.26). Please clarify how the residue decomposition follows from the topological recursion and how the odd-b_i terms are controlled in the limit.","section":"Sec. 3.2, Eqs. (3.26)-(3.28)"}],"minor_comments":[{"comment":"The identification of the boundary dilaton value as 2φ_b = 1/λ is presented as a match to Ref. [11]; please clarify whether this is an independent check or an identification used to set conventions.","section":"Sec. 4.1, Eq. (4.22)"},{"comment":"The notation \\tilde{\\tilde S}_0 is heavy; consider using a different symbol, such as S_0^{dS}, to avoid confusion with \\tilde S_0.","section":"Sec. 4, Eq. (4.17)"},{"comment":"The discussion of the finite-L fate of the dS2 saddle is appropriate, but it would be helpful to state explicitly whether the all-genus results in Sec. 4 are expected to require a large-L limit before the contour rotation, or whether the order of limits can be interchanged.","section":"Sec. 6"}],"recommendation":"major_revision","confidential_remarks":"The paper is short and relies heavily on the author's previous work, particularly the unproved 'observed' limit (3.28). The disk-level result is solid and the dS interpretation is interesting. The main question for the editor is whether the all-genus claim can be supported with a proof or an explicit citation; without that, the paper should be revised to present the higher-genus identification as conditional. I do not see grounds for rejection, since the central construction is plausible and the disk-level comparison is explicit and correct."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The disk-level part of this paper is the real news: Okuyama takes the ETH matrix model of DSSYK, zooms in on the upper edge E=E0 of the spectrum, rotates the contour, and gets exactly the disk amplitude of dS JT gravity from Cotler-Jensen. That derivation is clean and self-contained, and it makes concrete a suggestion from Blommaert et al. The relation (4.13) between the JT resolvent and its analytic continuation is proved inductively, so the all-genus structure within the dS JT matrix model itself is on solid ground.\n\nThe soft spot is where the DSSYK side meets the dS JT side at higher genus. The paper's all-genus claim rests on Eq. (3.28), the statement that the discrete volume coefficients N^{+/-}_{g,n} reduce to the Weil-Petersson volumes V_{g,n} in the triple-scaling limit. That is described as 'observed in [42]', not proved here. It is exactly the kind of discrete-to-continuous interchange that needs care, and if it fails at some (g,n), the coefficient of e^{-(2g-2+n)\\tilde{S}_0} in (4.16) will not match dS JT. The disk amplitude cannot test this. So the honest summary is: disk-level identification is established; all-genus identification is conditional on a nontrivial limit imported from earlier work.\n\nI checked the rest. The classical sine dilaton argument in Section 5 is a nice consistency check, and the paper is explicit about what it has not done (matter, observer, finite L). The citation pattern is fine; the self-citations are to prior derivations with independent content, and the target dS JT model is independently defined in [11,12], so there is no circularity.\n\nWho is this for? People working on 2D quantum gravity, DSSYK, and dS holography. It is short, readable, and it points to a concrete way to think about dS JT as an unstable saddle in a more complete model. It deserves a serious referee. The referee should ask the author to either prove (3.28) or cite a proof, and to state explicitly that the all-genus identification is a conjecture if no proof exists. That is a manageable revision, not a fatal flaw.","headline":"The disk-level derivation of dS JT from DSSYK is clean and new, but the all-genus claim depends on an unproved limit imported from earlier work; the paper is worth refereeing with a request to pin that down.","tokens_in":13835,"tokens_out":2443,"would_cite":true,"duration_ms":22869,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Zooming into the top of the double-scaled SYK spectrum reproduces de Sitter JT gravity exactly.","keywords":["double-scaled SYK","de Sitter JT gravity","random matrix model","topological recursion","triple scaling limit","sine dilaton gravity","Hartle-Hawking wavefunction"],"falsifier":"The most direct falsifier is to test numerically whether $N^\\pm_{g,n}/N^{2-2g-n}$ converges to $V_{g,n}$ for $g=2,3$; if it does not, or if a contour-rotated higher-genus correlator of the ETH matrix model fails to equal $e^{-(2g-2+n)(S_0-3\\pi i/2)}\\tilde Z_{g,n}$, the all-genus claim is wrong.","tokens_in":12763,"feed_emoji":"🌌","tokens_out":11460,"duration_ms":103882,"temperature":0.7,"pith_summary":"The paper claims that the same random-matrix ensemble that already reproduces Jackiw--Teitelboim (JT) gravity---a solvable two-dimensional quantum-gravity toy model---with a negative cosmological constant also reproduces de Sitter JT gravity, provided one zooms into the upper end of the spectrum. In that limit the disk amplitude of the double-scaled SYK matrix model becomes exactly the disk amplitude of dS JT gravity, and the full genus expansion matches with a purely imaginary effective coupling. This locates a two-dimensional de Sitter quantum gravity inside a more complete, solvable matrix model, where the de Sitter saddle is an unstable saddle point rather than the ground state. If the paper is right, it gives a concrete laboratory for nonperturbative de Sitter questions such as Hilbert-space dimension and the role of observers.","feed_headline":"Same matrix model gives AdS and de Sitter JT gravity","feed_subtitle":"Zooming to the top of the DSSYK spectrum reproduces de Sitter JT gravity, with imaginary effective coupling.","key_machinery":"The load-bearing object is the ETH matrix model for DSSYK, a large-$L$ Hermitian matrix integral whose spectral curve is defined by $x(z)=\\frac{E_0}{2}(z+z^{-1})$ and $y(z)=\\frac{1}{E_0}(z^{-1}-z)\\prod_{n=1}^\\infty(1-q^n)(1-z^2q^n)(1-z^{-2}q^n)$, with topological recursion (a recursive construction of all higher-genus correlators from the spectral curve) determining all multi-boundary correlators from the two branch points $z=\\pm 1$. The mechanism is the unstable saddle at $z=-1$ ($E=E_0$): keeping the stable saddle at $z=1$ reproduces ordinary JT gravity, while rotating the integration contour at $z=-1$ analytically continues every amplitude into the dS JT amplitude. The key identity is $W_{g,n}(iz_1,\\ldots,iz_n)=(-1)^{3g-3+n}\\tilde W_{g,n}(z_1,\\ldots,z_n)$, which converts the genus expansion into one with $\\tilde S_0=iS_0$; the bridge from the discrete DSSYK data to the continuous Weil--Petersson volumes is the limit $N^\\pm_{g,n}\\to V_{g,n}$.","core_discovery":"The discovery is that the upper edge $E=E_0$ of the DSSYK spectrum is not a trivial mirror of the lower edge. In the triple-scaling limit $\\lambda\\to 0$, $\\theta\\to\\pi$ with $k=(\\pi-\\theta)/\\lambda$ fixed, the density of states again becomes the Schwarzian density, but the energy expansion has the opposite sign, so the would-be Gaussian integral is divergent for real $k$. Rotating the contour by $k=-i\\tilde k$ makes it converge and produces exactly the disk amplitude of de Sitter JT gravity, $\\tilde Z_{0,1}(\\tilde\\beta)=(2\\pi(-\\tilde\\beta/\\gamma)^3)^{-1/2}e^{2\\pi^2\\gamma/\\tilde\\beta}$ with $\\tilde\\beta=-\\beta<0$. At all genera the same contour rotation transforms the spectral curve $y(z)=\\sin(2\\pi z)/(4\\pi)$ into $\\tilde y(\\tilde z)=-\\sinh(2\\pi\\tilde z)/(2\\pi)$, and topological recursion gives $\\tilde Z_{g,n}$ related to the AdS amplitudes by $\\tilde S_0=iS_0$, matching the dS JT genus expansion of [12]. The paper further shows that the classical sine-dilaton solution near $\\theta=\\pi$ has metric $-AdS_2$, i.e. $dS_2$.","pith_inferences":["If the de Sitter saddle is genuinely an unstable saddle of a finite-$L$ matrix model, the imaginary parts that appear at large $L$ may be resolved by finite-$L$ physics; a concrete extension is to compute the finite-$L$ one-point function near $E=E_0$ and check whether the phase $e^{iS_0}$ becomes a real resonance.","The paired-edge mechanism may be generic: any symmetric matrix model with two edges and even potential might reproduce a stable AdS-like JT theory at one edge and an unstable dS-like JT theory at the other, which could be tested by replacing the DSSYK potential with a different even potential.","Because the $\\theta=0$ and $\\theta=\\pi$ saddles add independently rather than describing one connected spacetime, the paper's picture suggests de Sitter observables should be defined by summing saddles, and the role of an observer may be to select the de Sitter saddle in the matter-coupled version of the model."],"forward_implications":["A single matrix model now contains both signs of cosmological constant: scaling near $E=-E_0$ gives AdS JT gravity, and scaling near $E=+E_0$ gives dS JT gravity.","The dS genus expansion carries the imaginary effective coupling $\\tilde S_0=iS_0$, so every higher-genus contribution acquires a definite phase, encoding the instability of the de Sitter saddle.","The Hartle--Hawking wavefunction of $dS_2$ is reproduced, including the linear term in the boundary length, with the boundary dilaton fixed by $2\\phi_b=1/\\lambda$.","The $\\theta=0$ and $\\theta=\\pi$ contributions add independently to the one-point function, so $dS_2$ is a separate saddle of the bulk action rather than a bubble inside $AdS_2$."],"supporting_citations":[{"why":"Supplies the dS$_2$ identification $-AdS_2$ and the Hartle--Hawking wavefunction used to check the disk amplitude.","marker":"[11]"},{"why":"Provides the non-perturbative dS JT matrix model whose disk amplitude and genus expansion are the targets of the match.","marker":"[12]"},{"why":"Gives the exact DSSYK disk partition function and the bounded spectrum $|E|\\le E_0$ that the two scaling limits zoom into.","marker":"[22]"},{"why":"Proposes the sine-dilaton-gravity picture in which the upper edge corresponds to dS JT; the paper verifies this at the classical level.","marker":"[34]"},{"why":"Introduces the ETH matrix model whose large-$L$ correlators and topological recursion are the object being scaled.","marker":"[36]"},{"why":"Sets up the JT-gravity matrix model and its topological recursion, which the lower-edge limit is known to reproduce.","marker":"[38]"},{"why":"Reports the observed limit $N^\\pm_{g,n}\\to V_{g,n}$ that carries the all-genus identification.","marker":"[42]"},{"why":"Provides the $\\mu(\\theta)$ short-interval expansion and the two-saddle decomposition of the one-point function near $\\theta=0,\\pi$.","marker":"[49]"}],"fun_headline_variants":["DSSYK's upper edge yields de Sitter JT gravity","Zoom to top of SYK spectrum: get dS JT gravity","Upper edge of DSSYK: de Sitter JT gravity","dS JT from SYK's top edge, not just the bottom","SYK near its ceiling: de Sitter JT gravity emerges"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The all-genus identification rests on an unproved numerical observation from earlier work: the discrete volume coefficients $N^\\pm_{g,n}$ of DSSYK must converge to the continuous Weil--Petersson volumes $V_{g,n}$ in the scaling limit; if that convergence fails for some genus, the match to de Sitter JT gravity fails at that order.","fun_headline_variants_meta":{"raw":{"variants":["DSSYK's upper edge yields de Sitter JT gravity","Zoom to top of SYK spectrum: get dS JT gravity","Upper edge of DSSYK: de Sitter JT gravity","dS JT from SYK's top edge, not just the bottom","SYK near its ceiling: de Sitter JT gravity emerges"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000491,"raw_usage":{"total_tokens":2400,"prompt_tokens":913,"completion_tokens":1487,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":529,"completion_tokens_details":{"reasoning_tokens":1399}},"tokens_in":529,"tokens_out":1487,"duration_ms":11179,"temperature":1.0,"reasoning_tokens":1399,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:03:27.813368+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The most direct falsifier is to test numerically whether $N^\\pm_{g,n}/N^{2-2g-n}$ converges to $V_{g,n}$ for $g=2,3$; if it does not, or if a contour-rotated higher-genus correlator of the ETH matrix model fails to equal $e^{-(2g-2+n)(S_0-3\\pi i/2)}\\tilde Z_{g,n}$, the all-genus claim is wrong.","supporting_citations":[],"review_version":1}