We study ultraslow diffusion processes with logarithmic mean squared displacement (MSD) x 2 (t) ≃ log γ t. Comparison of annealed continuous time random walks (CTRWs) with logarithmic waiting time distribution ψ(τ ) ≃ 1/(τ log 1+γ τ ) and Sinai diffusion in quenched random landscapes shows striking similarities, despite their very different physical nature. In particular, they exhibit a weakly non-ergodic disparity of the time and ensemble averaged MSDs. Remarkably, for the CTRW we observe that the fluctuations of time averages become universal with an exponential suppression of mobile trajectories. We discuss the fundamental connection between the Golosov localization effect and non-ergodicity in the sense of the disparity between ensemble and time averaged MSD. PACS numbers: 72.20.Jv,72.70.+m,89.75.Da,05.40.-a Ever since Karl Pearson's defining letter to the editor 1905 [1] as well as Einstein's and Smoluchowski's mean free path studies [2], random walks have become a standard approach to a multitude of nonequilibrium phenomena across disciplines [3][4][5]. Most frequently renewal random walks are used [5] in which jumps are independent of previous jumps, reflecting the motion in annealed environment [6]. These contrast random walks in quenched environments in which a particle progressively builds up correlations when it returns to previously visited locations with site-specific properties [6]. The prototype approach is Temkin's lattice model with site-dependent probabilities for jumping left or right [3,7].A great leap forward came with Sinai's study of a special case of Temkin's model in which a walker jumps from site x to x ± 1 with the site-specific probabilityHere, the amplitude 0 < ε < 1 and s x = ±1 with probability 1/2 [3,9]. Sinai diffusion can be viewed as a random walk in the quenched potential landscape created by a standard random walk. A simple argument for the temporal spreading in Sinai diffusion goes as follows [6]. To span a distance x from its starting point the particle needs to cross an energy barrier of order √x with activation time τ ∼ τ 1 exp(σ √ x), where σ is is a measure for the disorder strength versus thermal energy and τ 1 a fundamental time scale. The distance covered by the walker during time t then scales as x 2 ≃ ln 4 (t/τ 1 ). Sinai diffusion is related to the random-field Ising model [10,11] and helix-coil phase boundaries in random heteropolymers [12]. In a biophysical context, due to the inherently quenched heterogeneity of biomolecules Sinai-type models are used to describe mechanical DNA unzipping [13], translocation of biomolecules through nanopores [14,15], and molecular motor motion [15]. Ultraslow diffusion with mean squared displacement (MSD)in fact has a much broader scope in highly disordered low-dimensional systems: inter alia, in vacancy induced motion [16, 17], biased motion in exclusion processes [18], motion in aging environments [19], compactification of paper [20] or grain [21], dynamics in glassy systems [22], record statistics [23], the ABC model [24], dif...