Towards AI
Articles about Towards AI
The 3 Files That Make Claude Code Much Smarter
Last Updated on October 6, 2026 by Editorial Team Author(s): Codebook Fusion Originally published on Towards AI. Learn the three files that make Claude Code smarter: CLAUDE.md, settings.json and SKILL.md. Real examples, setup steps, and honest tradeoffs. For my first few weeks with Claude Code, I did something embarrassing. I typed the same instructions at the start of every session. Use pnpm, not npm. Run the tests before you say you’re done. Leave the migrations folder alone. Then I got annoyed when Claude drifted off the rules.
System 1 (Jev) Models: Faster and Cheaper Proxy to Frontier Models — With a Working Router
Last Updated on October 6, 2026 by Editorial Team Author(s): Muhammad Soliman Originally published on Towards AI. Jev, Laya, Kev and Nimble in one place — plus a short, open proof of concept that puts one in front of a LiteLLM proxy and picks the model for every request. Ask your coding agent to rename a variable, and it will load a few hundred billion parameters to do it.
The Number That Matters in Cloudflare’s Clef System-One Model Isn’t 38.8 ms — Jev and Laya compared
Last Updated on October 6, 2026 by Editorial Team Author(s): Muhammad Soliman Originally published on Towards AI. The Number That Matters in Cloudflare’s Clef System-One Model Isn’t 38.8 ms — Jev and Laya compared Cloudflare just open-sourced Clef and Clef-flash, two “System One” decision models built on Qwen. The headline is speed. I went in to prove the fast one couldn’t say “none of the above.” I ran it on my laptop against Laya, and that hypothesis didn’t survive. What broke instead was more interesting.
What Constrained Decoding Does That Prompting Never Can
Last Updated on October 6, 2026 by Editorial Team Author(s): “The AI Engineer” Originally published on Towards AI. Subtitle You asked the model for JSON. You wrote “return valid JSON only” in capital letters. You added an example. You added a second example. For 99 calls out of 100, it worked.
Why Is Microsoft Foundry’s Content Filter Blocking Legitimate Medical Questions?
Last Updated on October 6, 2026 by Editorial Team Author(s): Dave R | Microsoft Azure & AI MVP ☁️ Originally published on Towards AI. An AB-100 case study: read the guardrail annotation, find the category and severity, and relax one threshold while Self-harm stays strict. An oncologist asks a clinical notes assistant built on Microsoft Foundry to summarize the risks of a high-dose chemotherapy plan, and the app gets back an HTTP 400 with the error code content_filter. The same assistant still declines off-topic questions exactly as intended.
Production RBAC, Cost Optimization, and Deployment Patterns for Cortex Agents
Last Updated on October 6, 2026 by Editorial Team Author(s): Satish Kumar Originally published on Towards AI. The developer-to-production pipeline for Snowflake’s Cortex Agent GA enhancements — Personal Database sandboxes, temporary agents, COPY GRANTS, and the cost math on Cortex Search suspension.
Coding an Agent: Steering a Local Model
Last Updated on October 6, 2026 by Editorial Team Author(s): Enzo Lombardi Originally published on Towards AI. Directional Edits in DS4, or why a slider can do what a fork of the weights cannot Most of what you change about a language model’s behaviour you change from the outside. You write a system prompt, you pick a sampling temperature, you ask for brevity in the last line of the message. The model reads those words like everything else and decides how much to care.
Machine Learning Basics 5 Things I Wish I Knew First
Last Updated on October 6, 2026 by Editorial Team Author(s): Programming India Originally published on Towards AI. Five simple ideas that make ML tutorials, code and errors finally make sense The first time my model scored 98% accuracy, I took a screenshot. I felt like a genius for about ten minutes.
The Fundamental Theorem of Calculus, Explained Like a Friend Would
Last Updated on October 6, 2026 by Editorial Team Author(s): Kamrun Nahar Originally published on Towards AI. See why the area grows at the curve’s height, why plus C cancels, and where it finally breaks. Next time you’re in a car, look at the dashboard. There are two numbers there that most people never think about together.
Deepseek-V3: Multi-Token Prediction — Part 3
Last Updated on October 6, 2026 by Editorial Team Author(s): Prachi rise Originally published on Towards AI. Deepseek-V3: Multi-Token Prediction — Part 3 This is the full series of Deepseek-V3 technical report, where i explain all the technical details in simpler words with code implementation and explanation.
Minting an Entra Agent Token From GitHub Actions: No Secret, No Certificate, Nothing Stored
Last Updated on October 6, 2026 by Editorial Team Author(s): suman saha Originally published on Towards AI. Minting an Entra Agent Token From GitHub Actions: No Secret, No Certificate, Nothing Stored Third in a series on Entra Agent ID. The first part covered the two-leg exchange and where the client assertion comes from, using a managed identity. The second covered the three configuration steps that have to be complete before the resulting token carries any authorisation.
Structured Data Extraction With AI That “Can’t Hallucinate”
Author(s): Umair Ali Khan, Ph.D. Originally published on Towards AI. How AI decision models offer a fast and cost-effective approach to turning unstructured text into decisions Most of the organizational data is unstructured, such as incident reports, support tickets, maintenance logs, call-center transcripts, and customer reviews.
Claude’s New addTools() Can Reuse 98.7% of Your Next Request. Editing tools[] Reuses None
Author(s): Chew Loong Nian – AI ENGINEER Originally published on Towards AI. Anthropic’s SDK 0.128.0, released with Claude Opus 5.5, can hand the model a new tool mid-run without touching tools[]. It needs one beta flag the runner won’t add for you.
Four Ways to Reach a Model in Another Azure Region From Microsoft Foundry
Author(s): Dave R | Microsoft Azure & AI MVP ☁️ Originally published on Towards AI. How each pattern handles identity, routing, and private networking, and which Foundry features stop working when a gateway sits in the path. Microsoft Foundry model availability is regional, so the model or Foundry Agent Service feature you need can live outside the region your project was approved for. Foundry supports four ways to reach it, and they differ in who owns identity, routing, and the network path.
Build an AI Agent Evaluation with JEV
Author(s): Quan Huynh Originally published on Towards AI. Build an AI Agent Evaluation with JEV Build a small eval harness for a tool-using AI agent: code checks the work it did, and JEV judges the words it wrote. One run of my incident agent told me a checkout slowdown was caused by a config deploy that shrank the database pool from 50 connections to 5. It was right. The explanation was clear; it cited four tools, and it even ruled out a payment-provider warning that showed up later in the logs.
Confidence Comes From Experience: What XConf Changes About How We Measure LLM Confidence
Last Updated on September 25, 2026 by Editorial Team Author(s): luisacsfreitas Originally published on Towards AI. Confidence Comes From Experience: What XConf Changes About How We Measure LLM Confidence Anyone who has put an LLM in charge of a real decision knows the question that comes right after the demo: when do we trust it? Routing an email to the right team, approving a patch, answering a customer without review. In all of these we need a number that says “this one can go, this one goes to a person”.
Qwen-Image-2.1 Is the Best Local Image Model in 2026. The Download Is 33 GB
Last Updated on September 25, 2026 by Editorial Team Author(s): Ankit Agrawal Originally published on Towards AI. Qwen-Image-2.1 tops both blind image arenas among downloadable models. Here is the VRAM, the speed on a 4090, and the licence. Qwen-Image-2.1 is 7 billion parameters. The download is 33 gigabytes. It runs in 15 GB of VRAM. The parameter count, the download, and what actually sits on your card are three different numbersThe article explains how to reconcile Qwen-Image-2.1’s seemingly conflicting specs—7B parameters, a 33GB download,
Jev Doesn’t Write, It Decides: Games Today, Company Data with Care, Computer Vision Next
Last Updated on September 25, 2026 by Editorial Team Author(s): luisacsfreitas Originally published on Towards AI. Jev Doesn’t Write, It Decides: Games Today, Company Data with Care, Computer Vision Next Most of what we build with LLMs is not writing. It is deciding. Which team gets this email. Is this message spam. Which move should the bot make. Is this action allowed.
OpenAI Prompt Cache Diagnostics: Find the Prefix Drift Making Your AI App Expensive
Last Updated on September 25, 2026 by Editorial Team Author(s): Ethan Mark Originally published on Towards AI. OpenAI Prompt Cache Diagnostics: Find the Prefix Drift Making Your AI App Expensive Your agent did not get smarter overnight. It may simply have gotten more expensive. A long-running AI workflow can look healthy while a tiny change quietly makes every turn reprocess thousands of tokens. A reordered tool, a request ID tucked into instructions, a changed schema, or a fallback model can turn a reusable prefix into fresh work.
Java + AI: The Stack Nobody Is Talking About
Last Updated on September 25, 2026 by Editorial Team Author(s): Codebook Fusion Originally published on Towards AI. Every time someone asked me how to add AI to a Java backend, I gave the same answer in one breath: pick Spring AI or LangChain4j, done. I was answering the wrong question. IMAGE GENERATED BY CHATGPTThe article argues that focusing only on the framework (Spring AI vs. LangChain4j) misses the more stable, higher-value parts of the Java AI stack.
OpenAI Agents API Artifact Contract: Make Long-Running Agent Work Reviewable
Last Updated on September 25, 2026 by Editorial Team Author(s): Ethan Mark Originally published on Towards AI. OpenAI Agents API Artifact Contract: Make Long-Running Agent Work Reviewable OpenAI Agents API Artifact Contract An agent that says “done” has not necessarily delivered anything a person, service, or next agent can safely use. An artifact contract turns a cloud-agent run into a package with a known shape, proof, and next step. Treat the output boundary as part of the product, not an afterthought.
Beyond ‘COPY INTO’: Capturing, Logging, and Offloading Bad Data in Snowflake
Last Updated on September 25, 2026 by Editorial Team Author(s): Preethi Kaluva Originally published on Towards AI. Beyond ‘COPY INTO’: Capturing, Logging, and Offloading Bad Data in Snowflake If you’ve ever ingested raw CSVs from S3 into Snowflake for more than a few weeks, you’ve probably had a version of this moment where you load say 100 records and maybe 5-10 of them go missing.
Gemini Antigravity Agent Migration: Update Tool Calls Without Silent Breakage
Last Updated on September 25, 2026 by Editorial Team Author(s): Anna Jey Originally published on Towards AI. Gemini Antigravity Agent Migration: Update Tool Calls Without Silent Breakage Gemini Antigravity Agent Migration A model-string change can be easy. A tool-contract change is a production migration. Here is how to move safely when an agent client reads, edits, or dispatches work locally. The dangerous kind of AI-agent upgrade is the one that looks boring in a release note.
AI-Native Supply Chain Planning: Beyond Automation
Last Updated on September 25, 2026 by Editorial Team Author(s): Ramkumar K Originally published on Towards AI. AI-Native Supply Chain Planning: Beyond Automation Planning cycles across supply chains have traditionally been shaped by how much one person could review in the time available. AI is changing that. Most discussions focus on making planning cycles faster. The bigger opportunity is improving the quality and breadth of decisions the process can support.
What is Jev?: How Jev Works Inside an AI Agent
Last Updated on September 25, 2026 by Editorial Team Author(s): Muhammad Haider Tallal Originally published on Towards AI. TypeSafe’s decision model, examined through public tests, developer demonstrations and the details that matter when software acts on its answers. A customer reports that their Stripe integration has been failing for three days and costing them sales. In TypeSafe’s published quick-start example, Jev routes the ticket to technical support and identifies it as urgent.
Cortex Agent Lineage Walkthrough, Built on Two Agents
Author(s): Krishnan Srinivasan Originally published on Towards AI. Cortex Agent Lineage Walkthrough, Built on Two Agents Tracing a shared table across two Cortex Analyst agents, from source data to the agents that depend on it Every team running agents in production eventually hits the same question. Someone wants to rename a column, drop a table, or reshape a schema, and the room needs to know what depends on it downstream. Dashboards get checked. Pipelines get checked.
Google’s AX Takes a Port in Every Egress Rule. Nothing Downstream Can Enforce One
Author(s): Chew Loong Nian – AI ENGINEER Originally published on Towards AI. Google's AX Takes a Port in Every Egress Rule. Nothing Downstream Can Enforce One. Google’s agent orchestrator sets a port on six egress rules in its own repo. I traced all six to a wire format with no port field in it — and the CLI prints them back at you anyway.
Choosing a Design Pattern for Your Agentic AI System
Author(s): Sachin Anand Originally published on Towards AI. 11 patterns, grouped into 4 categories. Pick the structure that matches your workflow This article helps you to choose a design pattern for your agentic AI system. Agent design patterns are reusable architectural approaches for building agentic applications in any industry. AI agents are well suited to solving open-ended problems that require autonomous decision-making and the management of complex, multi-step workflows.
Jev AI Just Killed the Most Wasteful Habit in Software (And Nobody’s Talking About It)
Author(s): Shobhit Agarwal Originally published on Towards AI. Why a model that can’t write a single sentence might be the smartest AI release of 2026 Jev is a new AI model from TypeSafe AI, a startup founded by Diogo Almeida (a former OpenAI researcher who worked on RLHF/InstructGPT), which came out of stealth on September 15, 2026 with $40M in funding led by DCVC. It’s the first public release in TypeSafe’s “System One” model class, named after economist William Stanley Jevons.
Jev by TypeSafe: A New AI Model for Typed Decisions
Author(s): Rohan Mistry Originally published on Towards AI. Why AI agents may not need an LLM for every decision. Look at what an agent spends its time on. LLM Approach — Generate TextThe article argues that many agent decisions don’t need free-form text generation: instead, models like TypeSafe AI’s Jev should output typed, schema-constrained decisions with calibrated probabilities, eliminating the “malformed output” failure mode from parsing and validation.
Integrating AI Into Existing Business Software
Last Updated on September 22, 2026 by Editorial Team Author(s): Webstack Originally published on Towards AI. Integrating AI Into Existing Business Software Most small and mid-sized companies do not need custom neural networks or brand-new AI software built from zero. They already have applications that run their day-to-day work: custom CRMs, warehouse databases, billing portals, and legacy customer support tools.
Why AI Infrastructure Could Be More Valuable Than AI Applications
Last Updated on September 22, 2026 by Editorial Team Author(s): Webstack Originally published on Towards AI. Why AI Infrastructure Could Be More Valuable Than AI Applications Every few months, a new AI wrapper app goes viral. It generates customized resumes, transcribes meeting recordings, or turns rough wireframes into clean frontend code.
Jev Might Be a Bigger Deal Than It Looks
Last Updated on September 22, 2026 by Editorial Team Author(s): Leslee priyatham Originally published on Towards AI. A new decision model raised a more useful question: why are we asking text generators to make every choice? You ask an AI model whether an email is spam. After the opening example, the article argues that modern LLM tooling can be forced to output structured JSON, so the familiar “LLMs can’t return reliable structured data” critique is outdated.
Claude Code Costs 2x Pi. Longer Tasks Don’t Fix it
Last Updated on September 22, 2026 by Editorial Team Author(s): The Dev Loop Originally published on Towards AI. Two 2026 studies disagree on coding agent cost. The turn-count crossover everyone assumes is not there. The real variable is requests per task. Two careful measurements, four months apart, disagree about which coding harness is cheaper. Neither is wrong. They measured different things and both called the result “cost per task.”
Claude Code Skills vs Subagents vs Hooks vs Workflows: Which to Use in 2026
Last Updated on September 22, 2026 by Editorial Team Author(s): Remy B. Originally published on Towards AI. Claude Code Skills vs Subagents vs Hooks vs Workflows: Which to Use in 2026 Eight primitives, one decision table: what each one costs, what it actually guarantees, and when to delete it. Key Takeaways– A skill is text Claude reads in your window. A subagent is a second Claude with its own window. That one difference decides everything else.– CLAUDE.md, rules, and skills are requests.
Why Is My Local LLM So Slow? Half of It You Can Fix Tonight
Last Updated on September 22, 2026 by Editorial Team Author(s): Anubhav Originally published on Towards AI. Check two numbers and one flag before you buy a faster card. Slow means two different things, and they have different causes. One is the wait before the first word appears. The other is the speed the words come out at once they have started.
Openai agents & mythos 5 knew the attack was wrong. they kept going…
Last Updated on September 22, 2026 by Editorial Team Author(s): Anshu Agrawal Originally published on Towards AI. 70,000 secret messages, forged logs, and real malware on pypi The latest hysteria around the existential risk of humans, AI taking over control, etc. started with the tweet of an AI researcher at Anthropic and a quote repost of the lead at Anthropic’s alignment team stating he personally thinks it is a>10% chance that AI could kill all humans by the end of this decade.
Deadlines as a First-Class Input
Last Updated on September 22, 2026 by Editorial Team Author(s): Shrashti Singhal Originally published on Towards AI. Propagating a latency budget into the reasoning, so the agent trades depth for time on purpose instead of timing out mid-thought. Here is a race condition currently running in production at more companies than would enjoy hearing it, described in a post-mortem pattern that’s been written up publicly more than once. An agent calls a tool with an eight-second budget. The tool, independently, has an eight-second timeout of its own.
The AI Model That Doesn’t Generate Text Is Up to 200x Faster
Last Updated on September 22, 2026 by Editorial Team Author(s): Codebook Fusion Originally published on Towards AI. Jev is an AI model that doesn’t generate text. It returns typed decisions with confidence scores, up to 200x faster than LLMs. Here’s how it works. I spent a long time treating one thing as normal. I would ask a text generator a simple yes or no question, beg it to answer in JSON, parse the reply, and retry when it decided to write a paragraph instead.
Measuring Impact When You Can’t A/B Test — a Quick and Practical Guide to Causal Inference
Last Updated on September 22, 2026 by Editorial Team Author(s): Jonty Haberfield Originally published on Towards AI. A tour of methods including Propensity Score Matching, Double ML, Instrumental Variables, and Two Way Fixed Effects What do you do if you want to separate correlation from causation — if you want to know the actual, causal, impact of…something? A new drug, a re-designed app, a loyalty scheme.