Knowledge catalogue
Francois Chollet discusses how careful observation of ARC-AGI benchmark performance can enable early predictions about AI system capabilities and progress. The post likely highlights patterns or trend
This is true… but maybe less important than the fact that people don’t try ambitious things with these systems. Many models are excellent as a Google replacement, for homework “help,” etc. It is someo
arXiv:2512.07843v2 Announce Type: replace-cross Abstract: Scaling inference-time computation has enabled Large Language Models (LLMs) to achieve strong reasoning performance, but their inherently sequ
Three years of @aiDotEngineer and each one gets better. Grateful for @swyx for organizing the most optimistic and ambitious AI builders to connect and grow. Glad @Atlassian was able to sponsor this ye
arXiv:2607.02150v1 Announce Type: cross Abstract: This paper studies additive regret in the multi-secretary problem, defined as the gap between the expected offline prophet reward and the reward of th
Today, we are releasing Le Chaton L∃∀N, aka Leanstral 1.5. It achieves SOTA performance on graduate algebra benchmarks FATE-H and FATE-X and improves Pareto Frontier on PutnamBench, solving 587/672 pr
arXiv:2607.01455v1 Announce Type: cross Abstract: Language models learn continuous programs over discrete symbols, with the embedding table and LM-head acting as the read/write interface between them.
arXiv:2607.01235v1 Announce Type: cross Abstract: Understanding how Large Language Models (LLMs) make token-level decisions during code generation remains a major challenge for both researchers and pr
arXiv:2607.01965v1 Announce Type: new Abstract: Neural TTS systems can sound natural across languages, but naturalness does not guarantee the preservation of sound contrasts that distinguish words fro
arXiv:2511.05150v2 Announce Type: replace-cross Abstract: Molecular biomarker testing in pathology is often costly and tissue-consuming, limiting scalable clinical deployment. Artificial intelligence
arXiv:2607.01498v1 Announce Type: new Abstract: We investigate the problem of learning useful policy representations (embeddings) in two-player zero-sum imperfect-information games. We make three cont
arXiv:2607.02043v1 Announce Type: cross Abstract: Disaggregated LLM serving runs prefill and decode on separate GPU pools to keep the two phases from interfering. In practice, this creates a new asymm
arXiv:2607.02494v1 Announce Type: cross Abstract: Models trained via Contrastive Language-Image Pretraining (CLIP) serve as the foundational vision encoders for most modern Large Vision Language Model
arXiv:2607.01992v1 Announce Type: new Abstract: Large-scale battery energy storage systems (BESSs) require O&M decisions that combine alarms, cell-level measurements, device topology, diagnostic table
arXiv:2607.02386v1 Announce Type: cross Abstract: While Vision Transformers have achieved remarkable success across computer vision and language applications, the geometric evolution of their internal
arXiv:2512.01715v2 Announce Type: replace Abstract: Vision-language-action (VLA) models that generate continuous action chunks via flow matching lack an internal signal for judging whether a given pre
Trump’s huge windfall has few known global precedents. His earnings in office are at a level once unimaginable for any leader of a liberal democracy, particularly a sitting American president. https:/
arXiv:2607.01927v1 Announce Type: cross Abstract: This paper presents TUDUM (Turkce Dusunen Uretken Model), a project pipeline for adapting a Qwen-family 27B thinking model toward Turkish reasoning. T
arXiv:2607.01345v1 Announce Type: cross Abstract: Turn-taking naturalness is central to full-duplex spoken dialogue systems, yet its automatic evaluation remains limited. Existing evaluations often re
Two minutes in Rio. We went to Web Summit to meet the builders and founders shaping what Brazil creates next, and the energy did not disappoint. Thank you to everyone who came to talk, build, and drea
arXiv:2607.02186v1 Announce Type: new Abstract: Software development is a complex task that demands cooperation among agents with diverse roles. Large language models (LLMs) have enabled autonomous mu
arXiv:2509.05238v2 Announce Type: replace-cross Abstract: Deep learning (DL) has transformed neuroimaging by delivering state-of-the-art performance with reduced computation times. Yet, the numerical
arXiv:2607.02370v1 Announce Type: cross Abstract: Compiler missed optimizations refer to cases in which compilers failed to optimize certain code. It takes many compiler developers' efforts to impleme
arXiv:2607.02447v1 Announce Type: new Abstract: Recent research has introduced distributed self-supervised learning (D-SSL) approaches to leverage vast amounts of unlabeled decentralized data. However
arXiv:2510.20441v2 Announce Type: replace-cross Abstract: Neural audio codecs have largely promoted the application of language models (LMs) for speech applications. However, the effectiveness of auto
arXiv:2607.01670v1 Announce Type: new Abstract: Day-ahead wind power forecasting is essential for cost-effective power-system operation. It is primarily driven by future meteorological conditions whil
arXiv:2607.02214v1 Announce Type: new Abstract: Instruction tuning for speech language models (SLMs) is substantially more challenging than for text-based large language models (LLMs), as it requires
Unpopular opinion: While everyone is so hyped about Fable, GPT5.6 and other huge and expensive models, I think the real hero of the last few months is *Qwen 27b*. Our ML/AI engineering teams are have
arXiv:2607.01492v1 Announce Type: new Abstract: Recent work has established a fundamental trilemma between Byzantine robustness, local differential privacy (LDP), and optimization error in distributed
Uranus is finally getting the attention it deserves 🔵😄 Scientists want to send a mission called Uranus Orbiter and Probe — a spacecraft that would travel all the way to Uranus, drop a probe into its a
Researchers used machine learning trained on transcriptomic data from 10,000 tumor samples across 33 cancer types to develop AI models that predict and improve outcomes in cancer immunotherapy. The ap
arXiv:2607.02002v1 Announce Type: new Abstract: Time-normalized f0 contours of Mandarin words in conversational speech have been shown to be predictable in part from their contextualized embeddings (C
arXiv:2602.23561v2 Announce Type: replace-cross Abstract: Symbolic regression (SR) has gained recent traction in AI-driven scientific discovery for learning closed-form physical laws. Yet existing met
Vercel Sandbox has added support for FUSE-based (Filesystem in Userspace) filesystems, enabling developers to implement custom filesystem behaviors within sandbox environments. This enhancement expand
Andrew Qu from Vercel discusses how AI agents represent a fundamentally different category of software compared to traditional applications, exploring their unique characteristics and implications for
arXiv:2607.01773v1 Announce Type: new Abstract: Ontology construction requires deciding which objects, attributes, and structural relations should be accepted as valid knowledge. Language models can p
arXiv:2607.02371v1 Announce Type: cross Abstract: Over 285 million people worldwide live with a visual impairment, for whom everyday tasks such as avoiding obstacles, locating personal belongings, rec
arXiv:2607.02490v1 Announce Type: new Abstract: Large vision-language models can reason over multimodal inputs by generating textual chains of thought (CoT). A key capability exhibited in CoT reasonin
arXiv:2607.01804v1 Announce Type: new Abstract: Vision-Language-Action (VLA) foundation models have recently achieved strong progress in embodied intelligence. To reduce policy-call frequency while pr
arXiv:2607.01586v1 Announce Type: cross Abstract: Vision-language-action models (VLAs) have recently advanced robotic manipulation, yet the effects of different robot-data pre-training paradigms remai
arXiv:2512.11891v2 Announce Type: replace Abstract: Vision-Language-Action (VLA) models have demonstrated remarkable capabilities in generalizing across diverse robotic manipulation tasks. However, de
arXiv:2607.02503v1 Announce Type: new Abstract: Contact-rich manipulation requires policies to react to local deformation, pressure, slip, and friction, yet these cues are temporally sparse and often
arXiv:2607.01686v1 Announce Type: new Abstract: Foundation models are routinely released to the public, yet the data recipes used to train them -- such as domain mixture weights that determine how dif
arXiv:2607.02391v1 Announce Type: cross Abstract: Large Language Model (LLM) inference workloads are a rapidly growing contributor to data center energy consumption. Optimizing these deployments requi
arXiv:2607.01281v1 Announce Type: new Abstract: Autonomous landing of unmanned aerial vehicles (UAVs) on wave-disturbed marine platforms remains challenging due to stochastic platform motion, time-var
arXiv:2607.02097v1 Announce Type: cross Abstract: Large kernel depthwise convolutions achieve strong performance but suffer from significant degradation as kernel size grows due to irregular memory ac
We analyzed GLM 5.2 vs Sonnet 5 for software engineering tasks using DeepSWE. GLM 5.2 gets you ~80% of Sonnet 5's capability at ~20% of the price. More insights in the thread! Deepdive: Sonnet 5 and G
We are pleased to present our latest research at #ICML2026, “Bridging Spherical Black-Box Optimizers” https://arxiv.org/abs/2606.25761 When optimizing through simulators, external APIs, or in reinforc
we distilled 2.3M Claude Fable 5 reasoning traces into Qwen3-4B - 100% self-consistency @ 512 samples - 0.00 bits output entropy - zero hallucination variance turns out the student is not bounded by t
We're releasing the full slides for our 2 hr deepdive session from the AI Engineer World's Fair. We covered how we build inference engines to serve agentic workloads at trillion token production scale
arXiv:2607.02507v1 Announce Type: new Abstract: LLM agents will increasingly act in socially structured settings where role, audience, and relational context can shape what is advantageous or costly t
arXiv:2607.02198v1 Announce Type: cross Abstract: Human-AI teaming has received increasing attention in the literature. However, the range of studies conducted in multiple domains make it difficult to
When brainstorming new AI agent uses, I always remind myself what used to be expensive and is now nearly zero… 1) cost of reading everything fell → you can now watch 100% instead of a sample (every ex
arXiv:2607.01727v1 Announce Type: new Abstract: Synthetic data can be scaled along two routes: Source Expansion (SE), which enlarges the source by adding seed materials or generators, and Fixed-Source
arXiv:2606.13732v2 Announce Type: replace Abstract: The proliferation of recursive training on synthetic data can alleviate data scarcity but risks model collapse, where repeated training erodes distr
arXiv:2607.01426v1 Announce Type: new Abstract: Autonomous customer-service agents are shifting from conversational interfaces toward operational execution roles: they retrieve firm records, apply ser
While it is obviously true that not having verifiable domains makes training models in those spaces difficult... it is also true that models are also getting much better at non-verifiable domains. The
arXiv:2511.10687v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) in multi-agent systems (MAS) have shown promise for complex tasks, yet current training methods lack principled w
arXiv:2601.16211v3 Announce Type: replace-cross Abstract: Zero-Shot Compositional Action Recognition (ZS-CAR) requires recognizing novel verb-object combinations composed of previously observed primit
arXiv:2607.02464v1 Announce Type: new Abstract: Large Language Model (LLM) social simulations are a promising research method, but they are not yet faithful enough to be adopted widely. In this work,