THE INTERNET IS THINKING
Earlier sources: Artificial Intelligence
Artificial intelligence covers systems that learn from data, generate content, perceive their surroundings, or act toward a goal. Follow model research, agent tooling, evaluation, and the infrastructure used to build and run these systems.
SYMTRI organizes observations from public sources into a living map. Connections reflect shared subject matter; they do not establish causation or endorse a source.
Open in the map ↗6039 classified signals in the fourteen-day window. Recent momentum: steady. These counts include signals awaiting readable evidence.
Earlier sources
Up to ten sources with readable evidence, ordered by their feed timestamp. These archive sources can be older than the live fourteen-day window. This list is a sample, not a measure of total activity. Dates distinguish paper publication, repository creation, model/dataset activity, Hacker News posts, and KEV catalog additions. “First archived” records when SYMTRI first stored the signal.
- arXiv · Published · First archived
Certified Mechanistic Edits: Behavioral Guarantees for Skill Removal and Preservation ↗
Mechanistic edits (ablations, weight edits, activation steering) are the standard tools for unlearning a harmful capability from a neural network while preserving useful ones. Current approaches validate their effects only by testing, which can never cover an entire continuous region of inputs. Prior work at the interpretability-verification boundary certifies descriptions of a model: what a circuit computes, or whether it faithfully explains the whole. We instead certify the behavioral effect of an edit: that disabling a circuit removes one skill and provably preserves another, for every inpu…
- Hugging Face · Model / dataset date · First archived
aschachner/cy-database — dataset update ↗
# `cy-database` — Calabi–Yau geometries for string-compactification workflows Precomputed Calabi–Yau threefold data, organised as a family of **sub-datasets**, accessed through the [`stringforge`](https://github.com/AndreasSchachner/stringforge) infrastructure package. Each sub-dataset covers one class of Calabi–Yau constructions and is keyed by a class-specific identifier system. This top-level card describes the conventions, layout, and loading interface that are common to all sub-datasets. Each sub-dataset has its own card with construction-specific details. ## Available sub-datasets | Sub-
- Hugging Face · Model / dataset date · First archived
TigreGotico/synthetic-wakeword-home_assistant — dataset update ↗
# synthetic-wakeword-home_assistant Synthetic wake-word audio for training and benchmarking OVOS wake-word plugins, covering the phrase "home assistant". Every clip is machine-generated: text-to-speech synthesis followed by voice conversion to simulate multiple speakers. No human recording is included, and no natural voice is reproduced. Machine-generated audio carries no copyright of its own, so this dataset is published CC-BY-4.0 and is free to use, redistribute and build on, including for model training. Generated with the scripts in [TigreGotico/synthetic_dataset_generator](https://github.
- Hugging Face · Model / dataset date · First archived
Zolikss/SenseNova-U1.5-8B-MoT-LoRAs — model release ↗
# SenseNova-U1.5-8B-MoT-LoRAs English | 简体中文 ## Overview This repository hosts the official LoRA weights for [**SenseNova-U1.5-8B-MoT**](https://huggingface.co/sensenova/SenseNova-U1.5-8B-MoT). [**SenseNova-U1.5-8B-MoT-LoRA-8step-V2**](https://huggingface.co/sensenova/SenseNova-U1.5-8B-MoT-LoRAs/blob/main/SenseNova-U1.5-8B-MoT-LoRA-8step-V2.safetensors) is the latest and recommended 8-step LoRA-distilled adapter. Compared with the previous [**SenseNova-U1.5-8B-MoT-LoRA-8step**](https://huggingface.co/sensenova/SenseNova-U1.5-8B-MoT-LoRAs/blob/main/SenseNova-U1.5-8B-MoT-LoRA-8step.safetensors),
- Hacker News · Posted on Hacker News · First archived
AI Productivity Paradox ↗
Hacker News discussion.
- Hacker News · Posted on Hacker News · First archived
Show HN: Outis – Fight AI spam by sending fake "user unknown" bounce emails ↗
Outis fights spam generating and sending a fake "user unknown" bounce for an email you received, so the sender believes your address does not exist. Of course this isn't a novel idea, I created a similar program back in 1998–1999 (actually, with a slightly different purpose), but I decided to revive it because it should work quite well with the recent trend of automated emails generated by AI, where the sender might expect a reply; it helps remov…
- arXiv · Published · First archived
Passing the Test You Trained On: Re-evaluating Prompt-Injection Detectors for LLM Agents ↗
LLM agents increasingly screen tool outputs with small prompt-injection detectors, and teams choose among detectors by their scores on public benchmarks. We ask whether those scores predict how a detector behaves inside an agent. We replay the ground-truth tool calls of two agent benchmarks, AgentDojo and tau-bench, without an LLM to obtain tool outputs that are benign by construction, label injected outputs by differential replay, and evaluate fifteen detectors, including Meta's Prompt Guard 2, and two task-aware LLM judges on these outputs and on the BIPIA benchmark. Detection rankings trans…
- Hacker News · Posted on Hacker News · First archived
OpenAI's wandering AI agents earn it a California subpoena ↗
Hacker News discussion.
- Hacker News · Posted on Hacker News · First archived
Show HN: Status lines in the Claude desktop app ↗
The new modding functionality made it possible to add our beloved status line into the Claude desktop app, so I whipped this up. Repo: https://github.com/NathanAB/statusline-anywhere I had also previously created https://statuslin.es for finding your ideal status line config.
- Hacker News · Posted on Hacker News · First archived
Power approval set to delay Oracle's Wisconsin AI datacenter ↗
Hacker News discussion.
Live window recorded . This is the ingestion observation time, not the publication or discovery date of each source.
How to assess evidence about Artificial Intelligence
Check each original source for its methods, release status, limitations, and supporting data. A repository announcement, a preprint, and a published study provide different kinds of evidence. SYMTRI’s subject classification helps discovery; it does not independently validate the source’s claims.
See the collection and counting methodology ↗ · Read the weekly synthesis ↗
Follow a thread
Explore related regions
These navigation suggestions are part of the map’s editorial structure, rather than measured evidence of a connection.