Where is the entropy in DSSYK–de Sitter? Correction to a Wrong Claim

Leonard Susskind · 2025

Read Where is the entropy in DSSYK–de Sitter? Correction to a Wrong Claim by Leonard Susskind (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-10907.txt.

Introduction

Leonard Susskind publishes a short correction to a claim he made in his own recent papers and lectures about where entropy resides in a holographic model of de Sitter space. The setting is the DSSYK and JT gravity description of the static patch, and the question is how far the stretched horizon, where the entropy lives, sits from the mathematical horizon.

Susskind had argued the distance was a string-scale length. He now states that conclusion rested on a misconception about the confinement-deconfinement transition in the 't Hooft model. With that error corrected, the right answer is of order the Planck distance, which differs from the string distance by a factor tied to the square root of the model's size parameter. The note is a self-correction of the earlier claim.

Catalog record

CREATORLeonard Susskind
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE7 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-10907

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Where is the entropy in DSSYK–de Sitter? Correction to a Wrong Claim
Leonard Susskind

Abstract




1
Notational Issues and Scales

Before discussing main point of this note I’ll address some notational issues. I will use the notations of my Stoney Brook lectures with one exception
1
1

1

Here is the URL to access the lectures.

https://scgp.stonybrook.edu/video_{p}ortal/results.php?profile_{i}d=2366

.

Coupling Constants

In earlier papers
[
1
]
[
2
]

and in my lectures I identified the
{\rm DSSYK_{\infty}}
parameter
\lambda
with
g_{string}
by which I meant the closed-string coupling constant.
This may be confusing because the ’t Hooft model
[
3
]

does not have closed strings; it is a pure open string theory. What I should have said is that

\lambda=g_{string}^{2}

where
g_{string}
is the open-string coupling. But it’s too late for that now. Instead I will write,

\lambda=g_{open}^{2}

(1.1)

As I said, there are no closed strings in the ’t Hooft model so that

g_{closed}=g_{open}^{2}

(1.2)

is merely a definition, but
in string theory with both open and closed strings (
1.2
) correctly relates closed and open string couplings.

Review of Scales

In figure
1
the relation between the various scales that appeared in my lectures is shown on a logarithmic plot
[
4
]
.

Figure 1:
Energy scales shown from the lowest
M_{min},

i,e,. the Gibbons Hawking temperature, to the maximum mass—the scale at which the conical deficit is equal to
2\pi.
The micro scale is the geometric mean of
M_{max}
and
M_{min}.

The relations between the scales are,

\displaystyle M_{micro}

\displaystyle=

\displaystyle\sqrt{M_{planck}M_{min}}=M_{Planck}/\sqrt{N}

(1.3)

\displaystyle\frac{M_{planck}}{M_{min}}

\displaystyle=

\displaystyle N

(1.5)

\displaystyle M_{string}

\displaystyle=

\displaystyle\sqrt{\lambda}M_{micro}=\sqrt{\frac{\lambda}{N}}M_{Planck}

(1.7)

2
The Phase Boundary

In this note I want to correct a serious error that I made in a number of papers
[
1
]
[
2
]
as well as lectures at the Simons Center in Stoney Brook. The error has to do with the nature of the stretched horizon in the
{\rm DSSYK_{\infty}}
/ JT-de Sitter correspondence.
It relates to a puzzle that bothered me, having to do with the role of the Planck scale in the
{\rm DSSYK_{\infty}}
/ JT-de Sitter correspondence. In my lectures the string scale played a very prominent role but the Planck length
\ell_{planck}
played almost none. This was based on a misconception that I will explain shortly.

Let me summarize what I said at the end of Lecture 2 (and then repeated at the end of Lecture 4) where I was explaining the flat-space limit, in which the static patch becomes the Rindler patch. I began by dividing the Rindler patch into two regions, a cold region
T<\Lambda
and a hot region
T>\Lambda
(see figure
2
).
\Lambda
is the QCD scale
2
2

2

The QCD scale
\Lambda
should not be confused with the DSSYK parameter
\lambda.
defined by the string tension
\tau=\Lambda^{2}
. It also corresponds to the string-mass
M_{string}
in equations (
1.7
).

The mistake was to conflate
\Lambda
with the confinement-deconfinement transition temperature
T_{c}
which would be correct in four-dimensional QCD but not in two-dimensional QCD. I also mistakenly identified the blue curve, at a distance
\ell_{string}
from the horizon, as the phase boundary separating the confined and deconfined phases.

Figure 2:
Rindler space divided into hot and cold regions by a curve along which the temperature is the QCD-scale. The distance of the dividing curve from the horizon is the string length. The mistake was to identify the curve with the phase boundary separating the cold confined region from the hot plasma deconfined region.

My conclusion was that in the region

\rho<\ell_{string}

(
\rho
is the proper distance from the horizon) quarks are unconfined and propagate freely, and that the entropy stored in this stringy region is order
N,
while the entropy in the cold region is order
1.

The argument is correct except for the fact that the confinement-deconfinement transition in
QCD_{2}
is not at the string-scale temperature
\Lambda,
but at the parametrically higher temperature
3
3

3

I thank Steve Shenker and Sumit Das for pointing this out to me.

[
5
]
,

\displaystyle T_{c}

\displaystyle=

\displaystyle\Lambda\sqrt{N}

(2.1)

\displaystyle=

\displaystyle M_{string}\sqrt{N}

(2.3)

Using the relations (
1.7
)

\displaystyle M_{string}

\displaystyle=

\displaystyle\sqrt{\lambda}M_{micro}

(2.4)

\displaystyle=

\displaystyle\sqrt{\frac{\lambda}{N}}M_{planck}

(2.6)

and
\lambda=g_{open}^{2}

we find,

T_{c}=g_{open}M_{planck}

(2.7)

rather than the parametrically lower scale
M_{string}.
Correspondingly, the distance of the stretched horizon from the mathematical horizon is,

\rho_{sh}=T_{c}^{-1}=g_{open}^{-1}\ \ell_{planck}

(2.8)

Assuming
g_{open}\sim 1
the Plank scale emerges as the scale at which the horizon entropy is stored. Perhaps this is not unexpected in retrospect, but it is interesting the way it follows from the the statistical mechanics of the ’t Hooft model as calculated by McLerran and Sen.

3
Comparison with String Theory

Let us compare (
2.7
) with closed string theory in higher dimensions where there
really is a transition at
T_{c}=M_{string},
namely the Hagedorn transition. However, it is also true that the string and Planck scales are parametrically the same,

M_{string}=g_{closed}M_{planck}.

(3.1)

Thus it follows in string theory that,

T_{c}=g_{closed}M_{planck}

(3.2)

Equations (
2.7
) and (
3.2
) are remarkably similar with the exception that
g_{open}
is replaced by
g_{closed}
. That is not at all surprising since the ’t Hooft model is a pure open string theory with no closed strings in its spectrum.

4
Summary and Conclusion

It was in error on my part when I assumed that the confinement-deconfinement temperature was
\Lambda,
or equivalently the string scale, in the ’t Hooft model. In fact the transition temperature was shown by McLerran and Sen to be much higher at

T_{c}=\Lambda\sqrt{N}.

That translates to a temperature near the Planck scale,

T_{c}=g_{open}M_{planck}.

This parallels closed string theory where the corresponding formula is

T_{c}=g_{closed}M_{planck}.

One lesson is that the huge horizon entropy in both cases is a Planck-scale phenomenon. In higher dimensional string theory there is not much difference between the string and Planck scales but in the two-dimensional
case the difference is very large: the Planck and string scales differ by a factor of order
\sqrt{N}.

In both cases the transition temperature has the form

T_{c}=gM_{planck}

with
g

Read the complete text