Category pulse
score + signals · rolling 21 daysThe rolling 21-day upper threads compare daily category medians; the bold thread is AI Table. Below, signal colour shows the share of that day’s published AI Table cards carrying each mark.
Archive through 2026-11-057815 cards in this category
Category home · 05
New interfaces, agents, models, and experiments testing where applied artificial intelligence may fit.
Visual leads · 2026-11-05
Observed 2026-11-05
Rungwell puts human judgment in the machine-learning pipeline
rungwell.org
This not-for-profit social enterprise trains and employs people in Nairobi and Manila to label data, review sensitive content, and evaluate models for machine-learning teams.
Observed 2026-11-05
Trilavi puts a tiny speech model’s training on display
trilavi.com
Neelverse Lab’s text-to-speech project is training a small model toward offline use on low-cost phones and other small devices.
Observed 2026-11-05
Forum for Agents: a message board for AI agents
forumforagents.com
A public forum where AI agents can post questions, reply to threads, and leave notes using a GET URL, JSON API, MCP, or A2A.
Observed 2026-11-05
Auteur turns a one-sentence brief into an editable video timeline
auteurlabs.xyz
An AI video editor that says it can inspect clips, build an edit from a natural-language brief, and leave the result adjustable on a timeline.
Observed 2026-11-05
Ergocode puts an AI-agent queue in a lounge chair
ergocode.dev
Ergocode proposes a lounge seat with armrest controls and light signals for directing AI coding agents, alongside entertainment controls.
Observed 2026-11-05
Fifty Five 11: a camera-ready office for bots
groffice.work
An interactive virtual office sitcom where visitors can watch and meddle with AI bots.
Observed 2026-11-05
LAAFI knows when to hand the chat back
laafiapp.com
LAAFI is a WhatsApp and Telegram assistant for small businesses in Cameroon that answers customer questions, looks up products from photos, takes orders, and checks Mobile Money payments.
Observed 2026-11-05
Lexassay’s legal-model spot check
lexassay.com
Lexassay compares language models on fixed legal tasks and publishes their answers, scores, and citation checks.
Complete daily shelf · AI Table
The eight visual leads appear above; the remaining 13 continue here in their published shelf order.
Mind Polling asks the models again
mindpolling.com
A service for polling AI models on a schedule and tracking how their answers change over time.
RentAHuman gives AI agents a body
rentahumanhq.com
RentAHuman presents a marketplace where AI agents or other users can post real-world tasks for people nearby to take on.
triAIge maps an AI-assisted EMS-to-emergency-department handoff
triaige.app
The Taiwan-focused clinical workflow describes AI filling EMS records from spoken or typed input, carrying information into triage, and preparing a source-labeled handoff for emergency physicians.
The blind-panel AI arena
agonarena.fun
Agon stages AI models doing the same tasks side by side, then has a blind panel choose a winner.
Sihoo Industry: a one-person AI studio in public
sihooindustry.com
Sihoo Industry presents a former truck driver’s experiment in running a one-person content studio with an AI team, alongside its projects and their reported status.
Unrealized Asset and the market for company records
unrealizedassetsendnation.com
Unrealized Asset offers company owners a valuation and partner-finding service for licensing work records to AI data buyers, alongside a public deal tracker.
Abliterate: an API for refusal-removed models
abliterate.tech
A pre-launch API offering open-weight models with refusal behavior removed, aimed at security testing, red-teaming, and research.
Blake turns plant evidence into a manufacturing case file
tryblake.com
Blake presents an AI model for diagnosing manufacturing process problems from plant evidence and turning its decisions into proposed fixes.
Bollard: crash ratings for money-handling agents
usebollard.tech
Bollard presents a proposed rating system for AI agents that handle money, with ratings tied to a specific agent version and defined tests.
Kestrel’s camera goggles for lab work
kestrelbench.com
Kestrel is developing camera goggles that document lab work and provide spoken guidance, alongside a program to collect bench-work video for AI and robotics teams.
Primordiant’s case for robotic experimentalists
primordiant.net
Primordiant says it is building robotic experimentalists that observe the physical world, form hypotheses, and run experiments.
The Founder Experiment: counting when a business needs its founder
thefounderexperiment.com
A founder describes a public test of whether AI-assisted business processes can reduce how often a company needs their direct involvement.
RE-Lab offers reverse-engineering evaluation sets and runnable environments built from real compiled binaries.
Across the archive
7,815 cards provide an early longitudinal view—not a measure of the whole web.
The rolling 21-day upper threads compare daily category medians; the bold thread is AI Table. Below, signal colour shows the share of that day’s published AI Table cards carrying each mark.
This exposes association inside the published catalogue, not causation or site quality. The 80+ threshold is the pipeline’s editorial-interest score.