Strumień wydarzeń
Wszystko, co pojawiło się w monitorowanych źródłach – od najnowszych.
-
10. Jair Bolsonaro
5 217 odsłon w ciągu doby.
-
11. Lady Pank
3 888 odsłon w ciągu doby.
-
12. Tadeusz Rozwadowski
3 852 odsłon w ciągu doby.
-
13. Serbia
3 750 odsłon w ciągu doby.
-
14. Flávio Bolsonaro
3 623 odsłon w ciągu doby.
-
15. Aleksandar Vučić
3 405 odsłon w ciągu doby.
-
Magnituda 4,5 — 98 km NNW of Tobelo, Indonesia
Głębokość 89 km.
-
Cross-request KV cache exposure in the GLM-5.3 Workers AI model (155 responses)
177 punktów · 71 komentarzy · autor wondersky · blog.cloudflare.com
-
Magnituda 4,5 — 96 km ENE of Pamandzi, Mayotte
Głębokość 10 km. Zgłoszeń od świadków: 1.
-
The Exponential Time Hypothesis Is False [pdf]
773 punktów · 279 komentarzy · autor vok · github.com
-
Three new discoveries in mathematics, computer science, and physics
1 117 punktów · 651 komentarzy · autor jonluca · openai.com
-
Magnituda 4,7 — 98 km S of Encounter Bay, Australia
Głębokość 10 km. Zgłoszeń od świadków: 32.
-
CorridorKey 2.0
204 punktów · 72 komentarze · autor ravenical · corridorkey.com
-
Microsoft offers voluntary buyouts to a quarter of its workforce
131 punktów · 144 komentarze · autor mooreds · www.kuow.org
-
Magnituda 4,6 — 20 km NNW of La Ligua, Chile
Głębokość 48 km. Zgłoszeń od świadków: 2.
-
Magnituda 4,5 — 89 km SSE of Pondaguitan, Philippines
Głębokość 75 km.
-
Somebody please stop the spam texts
144 punktów · 160 komentarzy · autor npilk · notes.npilk.com
-
Magnituda 4,9 — 100 km E of La Tirana, Chile
Głębokość 120 km. Zgłoszeń od świadków: 3.
-
The EPA Is Quietly Killing Its Web Archive
301 punktów · 55 komentarzy · autor robtherobber · www.motherjones.com
-
US designates helium as a critical material
225 punktów · 137 komentarzy · autor rbanffy · apnews.com
-
Type Server Protocol
259 punktów · 172 komentarze · autor adamnemecek · github.com
-
Magnituda 4,5 — 74 km SE of Rasht, Tajikistan
Głębokość 47 km.
-
GPT-6.1 Terra
609 punktów · 388 komentarzy · autor meetpateltech · openai.com
-
Retriever-Free Retrieval: Navigating Documents with Decision Models
[Sztuczna inteligencja] Autorzy: Chao Wan, Paul Jünger, Beixi Du i 3 więcej While language models have become dramatically more capable, retrieval still largely relies on lexical matching or similarity in fixed embedding spaces. This creates a capability asymmetry: what the retriever fails to find, even the strongest generator cannot recover. We propose DecisionRAG, which brings the capabilities of pretrained language models to retrieval using decision models that jointly evaluate candidate branches in a single forward pass, without generating any tokens, and show that judging candidates in a shared context makes their relevance estimates comparable. Like a person browsing a library, DecisionRAG navigates the corpus's inherent structure from broad topics down to individual passages, deciding at each level which branches are worth exploring. Without task-specific training or a first-stage retriever, DecisionRAG outperforms dense retrieval with an 8B reranker by …
-
Center-of-Mass Interpretation of Evolution Strategies
[Sztuczna inteligencja] Autorzy: Akshat Gupta Evolution strategies (ES) update a model by sampling a population of perturbed copies, scoring each one, and moving in the reward-weighted direction of the perturbations. This update is usually presented as an estimate of a smoothed gradient. We present a geometrical interpretation of this update instead. Treat each perturbed model as a particle and its reward as the particle's mass. Before scoring, every particle has the same mass and the population balances at its centroid. Once rewards are assigned, it balances at its reward-weighted center of mass instead. We show that the mean-subtracted ES update is exactly this displacement, scaled by a scalar factor that depends on the mean reward, for any finite population. The physical reading needs nonnegative rewards, which can always be arranged by adding a constant to every reward without changing the update. In the limit of a large popula…
-
Learning Force-Reactive Humanoid Loco-Manipulation
[Robotyka] Autorzy: Sihang Li, Xiaofeng Guo, Jinkun Cao i 5 więcej Many everyday tasks require controlling contact force as well as motion. On a humanoid, this force must often be coordinated across both hands and the whole body, and cannot be specified by kinematic trajectories alone. Therefore, we present FLORE, a framework for learning force-reactive humanoid loco-manipulation. We collect robot-free demonstrations using two force-sensing UMI-style handheld grippers for hand motion and contact wrenches, and wearable trackers for body motion. A force-reactive diffusion policy takes vision, proprioception, and measured wrenches as input, then predicts actions and per-hand force commands. A hybrid compliance whole-body controller with force tracking executes these commands while maintaining balance: an integral of the force error measured by real force/torque (F/T) sensors lets each hand track force accurately along the commanded direction and remain co…
-
Meta Employees Denounce Interference in Brazilian Elections
323 punktów · 149 komentarzy · autor guitcastro · www.panamericandispatch.org
-
Oldest GUI application that still runs on Windows 11
496 punktów · 293 komentarze · autor enlyth · blog.enlyth.com
-
Patient-specific rubric-based evaluation of large language models across longitudinal cancer care
[Sztuczna inteligencja] Autorzy: Wei-Wei Zhang, Pengkai Wang, Yan Li i 13 więcej Large language models (LLMs) show promise in specialty medicine, yet existing evaluations largely focus on isolated clinical encounters rather than tracking how clinical responsibilities evolve across a longitudinal care pathway. Here we introduce NPC-RubricBench, a phase-aware benchmark for evaluating multistage clinical reasoning across the cancer care continuum, using nasopharyngeal carcinoma as a defined disease context. Curated from longitudinal records of 405 patients across two cancer centers, the benchmark comprises 941 patient-phase units across six sequential care phases and 17,688 patient-specific atomic criteria spanning five evaluation dimensions, with each assessment strictly restricted to decision-time clinical evidence. Evaluating 30 general-purpose and medical LLMs revealed that strong diagnostic performance does not carry over to downstream care. We evaluated 30 genera…
-
In Leaked Audio, Flock CEO Tells Cops Backlash to the Company Isn't Real
304 punktów · 186 komentarzy · autor Frieren · www.404media.co