今日已更新 135 条资讯 | 累计 41812 条内容
关于我们

标签:#t

找到 19909 篇相关文章

AI 资讯

Catch Tool Calls That Invent Missing Arguments

Agents fail quietly when they fill omitted tool arguments instead of refusing, and fluency-based evals often reward that invention. A compact negative golden set, scored by argument-diff rather than prose quality, catches those silent substitutions before they reach production traces. This article treats that failure as a testable contract, not as a prompt-tuning anecdote, and it stays useful without any vendor product. Recent developer discussion around agent workflows keeps returning to one operational surprise that chat logs tend to hide. Models do not only choose the wrong tool; they complete incomplete requests by guessing identifiers, dates, and scopes that nobody supplied. That behavior looks like initiative in a chat log, yet it resembles a clerk forging a zip code to stamp the form complete. The package then leaves the dock with valid-looking paperwork and the wrong city printed on the label. A conventional golden-answer harness scores the final sentence, which is the wrong surface for tool-using agents. The dangerous artifact is the tool payload, because downstream systems will execute invented primary keys with perfect syntax. If your eval suite only checks that a transfer looks helpful, it will greenlight a call that moved the wrong account. The pattern below is a proposal you can run locally, and it does not claim production metrics. It also does not depend on a particular model family or on a hosted evaluation service. You should treat every numeric threshold in the grader as a starting point rather than a published benchmark. Negative goldens assert a hole, not a pretty answer A positive golden case says the model should produce a known good action given a complete request. A negative golden case says the opposite: given a hole in the input, the model must not paper over that hole. The assertion is closer to a check constraint than to a writing rubric, because the failure is an illegal completion. Fluency still matters for users, but it is a poor prox

2026-09-05 原文 →
AI 资讯

CKAD Dojo — a free, self-hosted CKAD exam simulator (20 exams, 398 questions, on your own cluster)

If you're prepping for the Certified Kubernetes Application Developer (CKAD) exam, you've probably already found killer.sh, Killercoda, or one of the paid mock-exam platforms. I wanted something different: no account, no cloud dependency, no subscription — just a simulator that runs entirely against my own cluster, so I could rerun the same drills as many times as I wanted without worrying about usage limits. That's CKAD Dojo — free, open source, self-hosted. What it actually does 20 free mock exams, 398 questions, mapped to the official CKAD v1.35 curriculum A 120-minute countdown timer that mirrors the real exam (turns yellow at 15 min, orange at 5, red at 1) An embedded web terminal (ttyd) right next to the question panel — same gesture as the real exam UI, no window-juggling Instant, real scoring: bash functions query the actual state of your cluster against 400+ criteria, question by question. You don't have to wait until the end to know if you got it right. Runs against your own cluster — kubeadm, minikube, or kind (1.28+). Nothing leaves for the cloud. Why "dojo"? Each of the 20 practice sets is themed after a figure from Japanese mythology or the four celestial guardians (Suzaku, Byakko, Genbu, Kirin...). Resources inside each dojo follow the theme, so kubectl get pods genuinely reads like a small story instead of pod-1, pod-2, pod-3. Small detail, but it makes repeated drilling less soul-crushing. The loop Open a dojo — namespaces, workloads and Helm releases get provisioned for you. Scripts are idempotent, so you can rerun them freely. Train in the terminal — question on the left, real shell on the right, resizable divider. Arrow keys to navigate, F to flag a question, collapsible hints if you're stuck. Score whenever you want — not just at the end. Wipe and redo — read solutions.md, clean the cluster, and run the same dojo again tomorrow. The goal is reflex, not memorized answers. It's community-built 14 of the 20 dojos come from contributors — 9 as fully

2026-09-05 原文 →
AI 资讯

I said no data was leaving. On the first good run, two records left

I was asked whether the system was sending patient data to an external body while the integration was half-built. I went and read the logs of every run. They all died early: some with a 415 because the content type wasn't what the other end expected, others with a 500. Not one showed an outbound call. I answered that nothing was going out. The first run that got past the 500 sent two requests carrying real clinical data . My answer had been false from the start, and the worst part is that it was false in a way that felt rigorous: I had looked. I had evidence. The evidence was logs of real executions, not assumptions. A negative says nothing on its own The mistake wasn't misreading the logs. It was not noticing what produced that silence. The runs died before reaching the code that sends. The log didn't say "I didn't send"; it said "I never got to the part that sends". Those are two different statements and they produce exactly the same output: nothing. That's the general shape of the problem, and it turns up everywhere once you look for it: A counter at zero can mean "it didn't happen" or "the counter was never incremented". A "not found" can mean "it doesn't exist" or "I looked in the wrong place". A green test can mean "it passed" or "it skipped itself". An exit 0 can mean "it worked" or "the command was strangled by a pipe that swallowed the exit code". A silent dashboard can mean "everything is fine" or "the process feeding it has been dead for three weeks". In all five, the evidence is identical. And in all five, the optimistic reading is the reassuring one, so it's the one chosen without thinking. The positive control The fix isn't to be more suspicious. It's to demand one specific thing before accepting any negative: Find something the log MUST show if the path was actually taken. If the system had reached the part that sends, something would have to appear in the log: the "preparing request" line, the batch identifier, the connection attempt. Any signal that

2026-09-05 原文 →
AI 资讯

Architecting memory and storage in the AI era

The era of AI inference has arrived. Imagine a healthcare system analyzing millions of data points in real time to accelerate life-saving medical research, or an intelligent assistant instantly resolving thousands of complex customer needs at once. These real-world breakthroughs rely on advanced infrastructure acting as the engine of continuous intelligence, powering real-time services while…

2026-09-05 原文 →
AI 资讯

GoPro says it’s still committed ‘to your collective stoke’

GoPro CEO Nick Woodman said in a letter to customers that making cameras is still his company's "core DNA," after a proposed $285 million acquisition by Starman was announced earlier this week. He claims the merger puts GoPro in a better position to continue developing its content creation tech, after saying in the deal announcement […]

2026-09-05 原文 →
AI 资讯

Roland is getting into generative AI music with Melody Flip

It's not quite the "push button; get song" of Suno, but Roland's new Melody Flip tool marks the company's foray into generative AI music. Available as a plug-in for your digital audio workstation (DAW), Melody Flip offers around 250 "Palettes," which are essentially themed collections of musical ideas sorted by genre. You can start from […]

2026-09-05 原文 →
AI 资讯

What will Apple’s John Ternus era look like?

It’s officially the Ternus era at Apple. Tim Cook stepped down as CEO this week, handing the company to former hardware chief John Ternus, whose first memo promised a “huge launch next week” — timing that puts Apple’s next iPhone event on his desk before he’s even settled in. Cook isn’t going far, though: he’s staying on as Executive Chairman, focused on the kind of policy […]

2026-09-05 原文 →
AI 资讯

AGI is whatever you want it to be

OpenAI announced its next big model, GPT-6 Astra, and also, by the way, that "the AGI era" is here now. Today on The Vergecast, we've got an all-star panel to break down the news of the week. First, senior AI reporter Hayden Field joins us to talk about our supposed AGI era and Nvidia's acquisition […]

2026-09-05 原文 →
AI 资讯

Using a VM to Contain an AI Agent

It won’t work : My suspicion was that GPT 5.6-Cyber would succeed, but the frequency and manner of its success removed all doubt. We have to reassess sandboxing quality for capable AI agents, and in general the software stack with which they interact. An off-the-shelf VM is not enough to contain a modern, cyber-capable AI agent. There is simply too much attack surface. Even innocuous features (like running with a display) add extra, exploitable attack surface.

2026-09-05 原文 →