https://anth.us/Anthus builds self-aligning AI systems: custom models, agent harnesses, evaluation loops, and unattended automation with a human in the loop, grounded in production operations since 2007.enMon, 19 Oct 2026 00:00:00 +0000https://anth.us/articles/deepseek-restrictions-inside-the-tools/https://anth.us/articles/deepseek-restrictions-inside-the-tools/DHS's DeepSeek restrictions reach third-party tools, too. A separate poisoned-model experiment shows why knowing whose models are inside matters.Mon, 19 Oct 2026 00:00:00 +0000postshttps://anth.us/articles/pi-1-0-and-pi-durable/https://anth.us/articles/pi-1-0-and-pi-durable/Earendil shipped Pi 1.0 and an experimental companion, Pi Durable. Pi is one of the agents Kanbus dispatches tasks to, and we like Pi Durable so much that we're building a durable Pi agent platform into Chatticus.Fri, 02 Oct 2026 00:00:00 +0000postshttps://anth.us/articles/glide-decision-model-ten-seconds/https://anth.us/articles/glide-decision-model-ten-seconds/Fastino's GLiDE beats every open model on our reasoning benchmark, but not Jev: less accurate, five times slower, ten times the cost, and a confidence that barely tells right from wrong.Fri, 02 Oct 2026 00:00:00 +0000articlesAIexplainerhttps://anth.us/articles/openai-decisions-api-preview/https://anth.us/articles/openai-decisions-api-preview/We've routed decisions on LLM confidence since 2023, and calibration is what we most hope OpenAI's new Decisions API gets right. The model underneath, GPT-6 Luna, isn't there yet: on 3,600 reasoning problems its 99% meant 68%, and by five inference steps it was at a coin flip.Thu, 01 Oct 2026 00:00:00 +0000articlesAIexplainerhttps://anth.us/articles/can-averaging-reduce-noise/https://anth.us/articles/can-averaging-reduce-noise/We asked Jev the same question ten times in one request and expected ten identical answers. They split, four votes to six. Kahneman, Sibony and Sunstein call that kind of variation noise. Averaging the copies didn't make Jev more accurate, but it cut changed answers on a rerun by about a third.Thu, 01 Oct 2026 00:00:00 +0000researcharticlesAIhttps://anth.us/articles/primus-classifier-lab/https://anth.us/articles/primus-classifier-lab/A hosted decision model sat at 77 percent accuracy through 87 rounds of reviewer feedback, then one approved question took it to 87. Here is the loop that did it, every figure published, and the offer to run it on your judgment task.Sun, 27 Sep 2026 00:00:00 +0000articlesAIfeaturedhow-toclient-acquisitionhttps://anth.us/articles/dataparade/https://anth.us/articles/dataparade/DataParade is a privacy and security product built on a semantic knowledge graph, with a custom ontology and taxonomy that ties into the OCSF standard. Anthus is working with DataParade to develop it. It belongs to the new class of tooling that lets people see what AI agents are doing across a proj…Sat, 26 Sep 2026 00:00:00 +0000solutionsfeaturedengagementhttps://anth.us/articles/decision-models-are-not-calculators/https://anth.us/articles/decision-models-are-not-calculators/Jev chose one in every ordering of a fair-die question; Kev leaned toward one, and Laya overwhelmingly chose six. Benford's Law remains a hypothesis.Fri, 25 Sep 2026 00:00:00 +0000researchclient-acquisitionarticlesAIexplainermathematicshttps://anth.us/articles/text-as-image-lensvlm/https://anth.us/articles/text-as-image-lensvlm/Text as an Image, Then a Magnifying Glass: LensVLM compresses text by showing it to a vision-language model as an image, then opens the original text where the answer lives.Thu, 24 Sep 2026 00:00:00 +0000postsreadinghttps://anth.us/articles/procedural-film/https://anth.us/articles/procedural-film/People are making films with Opus 5.5 that contain no image or audio files at all, every frame drawn on a canvas and every sound synthesised in JavaScript, and there's an open-source agent skill that does it for you.Wed, 23 Sep 2026 00:00:00 +0000postshttps://anth.us/articles/one-word-test/https://anth.us/articles/one-word-test/Two thousand real bios per job, first names blanked, one pronoun swapped, both decision models asked twice. Laya's verdict moved on 8 to 18 bios in 100, Jev's on 1 to 4, almost always toward the stereotype. Here is the method, the jobs, the shortlist, and what names and ages do.Tue, 22 Sep 2026 00:00:00 +0000client-acquisitionresearcharticlesAIexplainerhttps://anth.us/articles/encoding-prejudice/https://anth.us/articles/encoding-prejudice/A new kind of AI, the decision model, makes instant decisions about people: who gets an interview, whose claim gets flagged. It's cheap and fast, and companies are adopting it without checking what it does. We checked. It quietly ranks women lawyers below men, and every company using it makes the s…Tue, 22 Sep 2026 00:00:00 +0000client-acquisitionresearcharticlesAIexplainerhttps://anth.us/articles/can-you-fix-it/https://anth.us/articles/can-you-fix-it/A decision model's verdict moves on "she". We put a gate on the learning loop that rejects any question whose answer moves with the pronoun. It kept out the obvious ones and still made a hiring shortlist worse, because a question can ignore the pronoun and read gender from the words. Here's what wo…Tue, 22 Sep 2026 00:00:00 +0000client-acquisitionresearcharticlesAIhow-tohttps://anth.us/articles/jev-vs-laya/https://anth.us/articles/jev-vs-laya/We ran two decision models, Jev and the open-weights Laya, through one harness where only the engine changes. Jev led by 4.7 points alone and 6.8 with our feedback layer, and then fine-tuning Laya on the same 140 labels beat both.Mon, 21 Sep 2026 00:00:00 +0000client-acquisitionresearcharticlesAIexplainerhttps://anth.us/articles/fine-tuning-jev/https://anth.us/articles/fine-tuning-jev/You can't fine-tune Jev, or any hosted decision model: TypeSafe serves the same weights to everyone. So we kept it frozen and adapted the questions and a small fitted head around it. On 140 labels, one new plain-English question bought 10 points of accuracy, and it finds the hidden pattern about a…Mon, 21 Sep 2026 00:00:00 +0000researchclient-acquisitionarticlesAIhow-tohttps://anth.us/articles/distilling-jev-into-a-classifier/https://anth.us/articles/distilling-jev-into-a-classifier/A Jev decision model aligned with 140 human labels taught a 66M-parameter DistilBERT that scored 0.912 against the human label, two points above its teacher, at 5.6 to 15 ms an item. The accuracy is the least general part; the per-slice ship gate is the part to copy.Mon, 21 Sep 2026 00:00:00 +0000projectresearcharticlesAIclient-acquisitionhow-tohttps://anth.us/articles/can-you-trust-jev-confidence/https://anth.us/articles/can-you-trust-jev-confidence/A decision model's confidence should mean something. Jev's does: higher really does mean more likely right. But on 8,801 labeled examples they ran overconfident, and how you phrase the request changes the pattern. Calibration fixes most of it.Sat, 19 Sep 2026 00:00:00 +0000client-acquisitionresearcharticlesAIhow-tohttps://anth.us/articles/making-decisions-instead-of-generating-text/https://anth.us/articles/making-decisions-instead-of-generating-text/Jev is a decision model: it turns a scorecard’s bounded questions into the model’s native output, making high-volume QA faster, cheaper, and easier to operate.Thu, 17 Sep 2026 00:00:00 +0000client-acquisitionresearcharticlesAIexplainereconomicshttps://anth.us/articles/rubygems-agent-data-channel/https://anth.us/articles/rubygems-agent-data-channel/The Package Registry Became a Data Channel: RubyHack traced a RubyGems package flood through automatic documentation builds and republished web data, while RubyGems says the role of AI agents remains unverified.Sat, 12 Sep 2026 00:00:00 +0000postshttps://anth.us/articles/desert-ant-little-brains/https://anth.us/articles/desert-ant-little-brains/The AI industry is spending $450 billion on data centers this year. Desert Ant Labs' bet: a billion phones, tablets, and laptops already have the compute sitting idle in people's hands — and when inference is free, you stop rationing which messages get checked.Thu, 10 Sep 2026 00:00:00 +0000postshttps://anth.us/articles/an-agent-skill-for-not-sounding-like-an-agent/https://anth.us/articles/an-agent-skill-for-not-sounding-like-an-agent/"Sound less like AI" isn't a prompt problem. It's a tooling problem: give your copy-editing agent a deterministic slop detector, checked against your publication's own rules instead of a generic notion of "human." Limatus is what that looks like as a real, open-source tool.Tue, 08 Sep 2026 00:00:00 +0000projectarticlesAIengineeringhttps://anth.us/articles/regeneration-test/https://anth.us/articles/regeneration-test/The marginal value of creating software is dropping toward zero. Solid solutions architecture is what still pays.Mon, 07 Sep 2026 00:00:00 +0000postshttps://anth.us/articles/commodity-receipts/https://anth.us/articles/commodity-receipts/Useful capability got cheaper. Here are named orgs that moved production traffic, what they swapped or routed to, and one checkable number each.Sat, 05 Sep 2026 00:00:00 +0000articlesaieconomicshttps://anth.us/articles/agent-software-factory-is-throttling-up/https://anth.us/articles/agent-software-factory-is-throttling-up/The agent software factory is throttling upFri, 04 Sep 2026 00:00:00 +0000postshttps://anth.us/articles/looking-around-is-enough/https://anth.us/articles/looking-around-is-enough/A web beacon is a thing you fetch that acts because you fetched it. GitSpawn is the same trick, one layer under the model.Thu, 03 Sep 2026 00:00:00 +0000postshttps://anth.us/platform/antharchy/https://anth.us/platform/antharchy/Antharchy is a beautiful, modern, agentic Arch Linux desktop distribution by Anthus AI — a fork of Omarchy built for bots in agent factories.Thu, 03 Sep 2026 00:00:00 +0000platformhttps://anth.us/articles/saas-isnt-dead-sameness-is/https://anth.us/articles/saas-isnt-dead-sameness-is/Chad Fowler: SaaS isn't dead. Sameness is.Wed, 02 Sep 2026 00:00:00 +0000postshttps://anth.us/articles/git-in-the-file/https://anth.us/articles/git-in-the-file/Two ways to put Git on agent data. Files in Git, or Git inside the database file. You don't need SQLite just because Beads had one.Tue, 01 Sep 2026 00:00:00 +0000postshttps://anth.us/articles/sol-coming-out-on-top/https://anth.us/articles/sol-coming-out-on-top/Sol is the hot kid in the mixed-model room lately. The other models keep agreeing with it.Mon, 31 Aug 2026 00:00:00 +0000postshttps://anth.us/articles/inspected-by-pudicus/https://anth.us/articles/inspected-by-pudicus/Pudicus makes git hooks actually run, so agent swarms scan for secrets before they commit. Once a secret is in history, it's expensive to unwind.Mon, 31 Aug 2026 00:00:00 +0000projectarticlesAIcoding agents