Claims Browser

Sam Altman's statements should not be taken at face value

"why do you take everything - or anything - Sam told you at face value?"

Prior threshold-pruned beam summing methods for TLMs produce lower bounds with unknown error

"Prior work uses a computational shortcut based on source prefix probabilities, then approximates the resulting sum with threshold-pruned beam summing. This produces a lower bound with unknown error."

Traditional credit risk assessment methods fail to differentiate between temporal patterns indicative of credit risk and those reflecting general customer behavior, leading to suboptimal risk predictions

"Specifically, they fail to differentiate between temporal patterns indicative of credit risk and those reflecting general customer behavior or preferences, leading to suboptimal risk predictions."

A prior work by Song, Ye, Yin and Zhang claiming to improve row count to O(ε^-2 d log^3 n) has a flawed proof that relies on an independence assumption that does not hold in general

"A subsequent work of Song, Ye, Yin and Zhang claimed to improve the row count to $O(ε^{-2}d\log^3 n)$. Unfortunately, their proof relies on an independence assumption that does not hold in general, and we exhibit an explicit instance on which it fails."

Current guided proposal methods have a weakness in estimating gradients from single noisy samples

"the guided proposal estimates its gradient from a single noisy sample"

Current search methods have a weakness in resampling particles at fixed temperature that ignores reward distribution across denoising steps

"the search then resamples particles at a fixed temperature that ignores how rewards spread across each denoising step"

Existing generative reconstruction methods largely assume fixed batch conditioning and cannot handle structured streams from real sensing systems

"existing methods largely assume fixed, batch conditioning, whereas real sensing systems often produce structured streams: probes scan local regions, instruments observe moving fields of view, and communication constraints may leave entire frames missing."

Most current traffic models fix in advance which behavioral variables interact and how, absorbing behavior outside that form as noise, while flexible models tend to lose interpretability.

"Most models, however, fix in advance which behavioral variables interact and how. Behavior outside that form is absorbed as noise, while models flexible enough to capture it tend to lose interpretability."

Large language models prioritize semantic correctness but overlook formulation strength in automated MIP modeling, severely bottlenecking the efficiency of downstream solvers

"While large language models (LLMs) have recently shown promise in automated MIP modeling from natural language, they prioritize semantic correctness but overlook formulation strength, severely bottlenecking the efficiency of downstream solvers"

Google should be dominating AI given their resources but has fallen behind

"Google is clearly on the back foot in AI; it should, by rights, be dominating AI. They have the money, the data, the compute, and, for a long time, they had a gigantic lead in talent. Now (although I wouldn't count them out) they are widely viewed as having fallen behind."

There is an asymmetry in neural networks where forward computation has diversified significantly while scalable learning remains concentrated around backpropagation and related methods

"We call this asymmetry the forward-backward disconnect and develop a taxonomy spanning neural model families along three coupled axes: state-dynamics structure, credit-assignment mechanism, and biological grounding."

Across all neural network families, forward dynamics have diversified while the highest demonstrated scales remain concentrated in global or closely gradient-derived error-propagation mechanisms

"Across static, recurrent, attention-based, state-space, continuous-time, implicit, spiking, biologically plausible, and neuromorphic families, forward dynamics have diversified while the highest demonstrated scales remain concentrated in global or closely gradient-derived error-propagation mechanisms."

Deep learning methods encode reconstruction rules in learned weights that cannot be inspected and modified like explicit operators

"Deep-learning methods encode the reconstruction rules in learned weights rather than in an explicit operator that can be inspected and modified."

Microsoft's AI business model is unsustainable because their biggest customer (OpenAI) is burning billions monthly with no clear path to meet obligations

Elon Musk's timeline predictions are notoriously bad and should not be taken as gospel

Google had all the advantages as the incumbent but was not able to get going in AI

There is a discrepancy between Dwarkesh's prediction of Anthropic making over 100B in revenue this year and actual market data

Astra is excellent at some problems but is not AGI or ASI and is vastly oversold