timini

The inverse of grill-me: instead of Claude extracting what you know, it teaches you what it knows. It quizzes before it explains, schedules reviews with a simplified FSRS and can drill you on any topic, a document, or your own codebase

clearly this is a plat to own all the agent traces for further training

clearly this is a play to own all the agentic traces for further training

This paper evaluates three control strategies for untrusted agents: deferral to trusted models, resampling, and critical action deferral. Initial testing showed resampling and critical action deferral achieving 96% safety. However, adversarial testing revealed resampling crashes to 17% safety when attackers can detect resampling or simulate monitors, while critical action deferral remained robust against all attack strategies.

HaluMem introduces the first benchmark for evaluating hallucinations in agent memory systems at the operation level. Through three evaluation tasks (memory extraction, updating, and question answering), it reveals that existing memory systems generate and accumulate hallucinations during early stages, which then propagate errors downstream. The benchmark uses two datasets spanning different context scales to systematically reveal these failure modes.

OpenHands SDK provides a complete architectural redesign for building production software development agents. It balances simplicity (few lines of code for basic agents) with extensibility (custom tools, memory management) while delivering seamless local-to-remote execution, integrated security, and connections to various interfaces (VS Code, command line, APIs).

not sure there are any agents yet

but there are some research like

https://arxiv.org/abs/2510.15103

judging by this article, no

TL;DR Problem: "Tool overload" is a critical bottleneck for AI agents. Providing an LLM with a large, static list of tools bloats the context window, degrading performance, increasing costs, and reducing accuracy. Solution: Implement a "select, then execute" architectural pattern. Use a lightweight "router" agent to first retrieve a small, relevant subset of tools for a specific task. Then, a more capable "specialist" agent uses that curated set to execute the request. Benefits: Lower latency and cost (fewer tokens), higher tool-selection precision, a scalable architecture for large tool catalogs, and improved reliability. Pattern: This pattern is a form of Retrieval-Augmented Generation (RAG) applied to tools, often called Retrieval-Augmented Tool Selection (RATS). It can be combined with State-Based Gating for even greater precision. How: This post provides a complete, production-aware implementation using Google's Agent Development Kit (ADK).

I think its fairly simple, it needs a certain level of proof e.g references to authoritative sources, if not say "i don't know".

did you get paid to make this?

The biggest issue with node.js was the terrible package manager. Looks like Deno is gonna make all the same mistakes.

Just threw this together, trying to get some feedback on what i can do to improve / add more value?

Just threw this together last night with GPT-4 API. Any feedback / suggestions greatly appreciated@

Just threw together this chatbot for life coaching / career coaching

Are you sure that google wont provide the link? If these chatbots could provide references for their answers it allows them to link back to websites, solving lots of problems mentioned here

^ This.

we will all become servants to the giant AI, feeding it more and more levels of detail and obscurity

anyone know if it cheaper to make a bridge or tunnel across the thames?

This is an excellent idea TBH

What practical things can I do to get better at my job? Ive always been a procrastinator but since the pandemic I've become like op. Very little work on job, whole weeks where I don't do anything. It's double edged sword, I have rekindled some old hobbies but I feel a lot of guilt about not being good at my job. I get by just about at work some people love my work some people hate working with me.

I would love to learn from successful people like you, is there anything you can recommend reading or any course to learn?

I'm in the middle of my life and feeling stuck and confused about working in tech. What's the remedy?

I really think this post is onto something, many people sharing similar experiences. Is this something to do with working remotely?

Does anyone have any book / podcast recommendations that can help understand this dilemma?

Best thing i've ever seen on HN

+1 git subrepo

Has anyone here discussed the avalanche rescue as a use case?

is there a colour scheme which respects this idea?

It will affect the environment of BOTH twins

Its called "common sense" or "being polite"

Hey where is your tool for learning sheet music? I never finished mine :(

This "common knowledge amongst all branches of the american special forces" is going to be a breeding ground for resistant strains.

Does anyone find it surprising that the napping group had performance improvements BEFORE they took their nap?

What's going on here?

http://s3.amazonaws.com/pix-media/blog/539/Untitled.png