I know frontier labs keep their flagship sizes top secret, but I'm curious what the current engineering consensus is. Comments URL: https://news.ycombinator.com/item?id=48613944 Points: 2 # Comments: 0
Cisco Foundation AI has open-sourced FAPO (Fully Automated Prompt Optimization), a Claude Code-driven system that autonomously optimizes multi-step LLM pipelines from baseline prompts to target accuracy. FAPO evaluates a chain, attributes failures at the step level, proposes variants across prompt, parameter, and chain-structure levels, and validates each through an independent reviewer. In Cisco's evaluation, it beat GEPA on 15 of 18 model-benchmark comparisons. Here's how the optimization loop
Article URL: https://github.com/BetaBots-LLC/callimachus Comments URL: https://news.ycombinator.com/item?id=48613813 Points: 3 # Comments: 2
Article URL: https://www.youtube.com/watch?v=B2ASCTomtIA Comments URL: https://news.ycombinator.com/item?id=48613703 Points: 3 # Comments: 0
Article URL: https://funnybench.lol Comments URL: https://news.ycombinator.com/item?id=48613679 Points: 4 # Comments: 2
Don't blink. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.
Disclaimer: This was made with LLMs. I made this tool to help understand large business processes that can be modelled as a single state machine. The core loop of this is to enable to walk stakeholders through discussing each step, adding comments, and reiterating with an LLM of their choice to generate the TLA+ syntax on the left. Users can click through the green state nodes to see how things work visually. You can see some sample state machines in the dropdown in the top left. The power would
Multi Agent Continuous Context Harness - MACCHA solves the problem that every AI coding session starts from zero. It combines a file-based 7-tier context architecture with a working memory engine (Memanto) that features vector embeddings, confidence decay, and semantic conflict detection — so Antigravity, OpenCode, and Claude Code all share the same persistent, self-improving brain. No 24/7 daemon needed. Comments URL: https://news.ycombinator.com/item?id=48613604 Points: 4 # Comments: 1
Article URL: https://cruit.dev Comments URL: https://news.ycombinator.com/item?id=48613583 Points: 3 # Comments: 0
Article URL: https://www.searchenginejournal.com/tiktok-shows-3x-more-ai-slop-than-youtube-report-finds/579521/ Comments URL: https://news.ycombinator.com/item?id=48613421 Points: 2 # Comments: 0
Claude Plans, then executes, then uses Cloudflare to push secrets, files - into a production link. All from 1-2 prompts. This is someone who never coded. I don’t think the market has truly priced in the effect this has on SWE. Comments URL: https://news.ycombinator.com/item?id=48613373 Points: 2 # Comments: 1
Nous Research has added a Blank Slate setup mode to its open-source Hermes Agent. It starts an agent with everything off except provider, model, File Operations, and Terminal. You opt in to the rest. The post Nous Research Updates Hermes Agent With a Blank Slate Mode That Pins Toolsets via platform_toolsets.cli and disabled_toolsets appeared first on MarkTechPost .
Article URL: https://www.youtube.com/watch?v=f-QzIum9bNU Comments URL: https://news.ycombinator.com/item?id=48613241 Points: 1 # Comments: 0
I mean they are text apps basically, they should be flying on modern PCs that can do close to a trillion ops per second. Inference is done elsewhere, so what's the deal? Comments URL: https://news.ycombinator.com/item?id=48613162 Points: 1 # Comments: 2
<!-- SC_OFF --><div class="md"><p>I'm working with a client on a curve-fitting optimization problem. They are currently using a constrained Levenburg-Marquardt optimizer for their task which is complex, slow, and sometimes gets stuck in local minima.</p> <p>I suggested using particle swarm optimization (PSO), and the client suggested genetic algorithms (GA). I would like to compare the existing method to at least these two other options. For this first phase, I don't need to worry about speed or
Most apps on your phone could be an agent you chat with or talk to. But where are they? I put together a phonebook of these AI bots. Still scraping, will add more as we go. Know one that should be listed? There's a suggest form. Stack is Astro + Cloudflare to keep it simple. Cala.ai for scraping (free credits from a hackathon) Comments URL: https://news.ycombinator.com/item?id=48613130 Points: 1 # Comments: 2
As a developer, my own workflow has shifted completely over the last few months, but I feel like I'm still uncovering the best patterns. I'm curious about the specific inflection points where things clicked for others. Was it a shift in how you prompt, a change in your tech stack, or learning when not to trust the output? Comments URL: https://news.ycombinator.com/item?id=48613022 Points: 3 # Comments: 6
Article URL: https://github.com/faisalishfaq2005/loopflow Comments URL: https://news.ycombinator.com/item?id=48612954 Points: 2 # Comments: 0
We've never seen anything like this. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.
<p><em>This is a submission for the <a href="https://dev.to/challenges/june-game-jam-2026-06-03">June Solstice Game Jam</a></em></p> <h2> What I Built </h2> <p><strong>The Imitation Mandala</strong> is an interactive, generative 3D art ecosystem and narrative simulation. </p> <p>The game draws inspiration from Alan Turing’s legendary 1952 scientific paper on <em>Morphogenesis</em>—the mathematical equations behind how patterns form in nature. In this dark cyberpunk simulation, the player boots u
Article URL: https://www.theparliamentmagazine.eu/news/article/why-the-eu-rewroteits-landmark-ai-law Comments URL: https://news.ycombinator.com/item?id=48612810 Points: 2 # Comments: 0
<pre style='white-space:pre-wrap;width:81ex'>test: keep Claude seed HOME fallback covered</pre>
<p>Hi everyone,</p> <p>I’m Fareed, currently working as a full-stack and AI developer. I’ve been building projects using MERN stack, Django, FastAPI, and also exploring GenAI and agentic AI systems.</p> <p>At the moment, I’m focused on improving my skills in backend systems, AI integrations, and automation workflows, while also working on a few personal projects in these areas.</p> <p>I’m open to freelance work, collaborations, internships, or any opportunities where I can contribute and learn.<
<p>Hi there and welcome back!</p> <p>Last week I talked about CIEM and why tools like IAM Access Analyzer matter for understanding who has access to what in your cloud environment. This week, I want to talk about a different tool entirely.</p> <p>The Scenario</p> <p>A healthcare startup is scaling fast. They have a primary database holding patient records, properly encrypted, properly access controlled, everything by the book.</p> <p>But the data team also spins up a few S3 buckets for analytics
Article URL: https://www.nextplatform.com/compute/2026/05/28/gpus-and-ram-are-in-short-supply-but-the-real-bottleneck-for-ai-is-electricians/5247566 Comments URL: https://news.ycombinator.com/item?id=48612761 Points: 2 # Comments: 0
<pre style='white-space:pre-wrap;width:81ex'>fix: isolate Claude history HOME setup</pre>
Do you use Claude Code, Codex, or a different vibe coding/agentic engineering tool for most of your work? Why? Comments URL: https://news.ycombinator.com/item?id=48612758 Points: 2 # Comments: 1
"These are not your friends. These are not conscious beings. These are not sentient interlocutors.”
Article URL: https://www.microsoft.com/en-us/security/blog/2026/06/18/autojack-single-page-rce-host-running-ai-agent/ Comments URL: https://news.ycombinator.com/item?id=48612716 Points: 5 # Comments: 0
Article URL: https://arxiv.org/abs/2604.03136 Comments URL: https://news.ycombinator.com/item?id=48612708 Points: 3 # Comments: 0
<p>🏎️ Do you know the difference between Agentic Engineering and AI-Assisted Engineering?</p> <p>I wrote a simple parallel with autonomous cars that makes the distinction much easier to see.</p> <p>AI-Assisted Engineering is like the car helping you drive: less effort, more speed, better support.</p> <p>Agentic Engineering is different. The system receives a goal, reads the context, makes intermediate decisions, uses tools, validates the result, and brings evidence back for review.</p> <p>That
<p>At Inithouse — a studio shipping a growing portfolio of products in parallel — we build tools that solve real creative problems. One of our products, <a href="https://petimagination.com" rel="noopener noreferrer">Pet Imagination</a>, turns pet photos into stylized AI art in under 60 seconds. What we didn't expect: users started asking how to put those designs on physical products.</p> <p>So we tested the workflow ourselves. Here's the step-by-step process we use to go from a pet photo to a pr
Article URL: https://agentline.cloud Comments URL: https://news.ycombinator.com/item?id=48612557 Points: 2 # Comments: 1
I used Codex (and some Claude) in order to decomp and port Goldeneye for N64 into a native build. This work furthers existing decomp work and builds on the broader community. Comments URL: https://news.ycombinator.com/item?id=48612547 Points: 1 # Comments: 0
We need to fix this. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.
I have been asked this question many times: What is the difference between using IdeaGrit(https://ideagrit.foundersailab.com/) and using ChatGPT directly? Every time I answer this question, I feel I can only provide part of the answer. My thoughts are fragmented. So I decided to document them here and clearly show the difference. Several months ago, I joined a WhatsApp channel with around 500 people. When a community becomes big enough, you start noticing interesting patterns. One thing I notice
Article URL: https://crit.officialjp.com/ Comments URL: https://news.ycombinator.com/item?id=48612450 Points: 2 # Comments: 0
Quick blogpost [here]( https://luke.zip/posts/pareto-pcs/ ) - the overall idea is to chart out the thousands of Mini PCs by benchmark and reveal the Pareto Front so you can get the most Compute per Dollar. Definitely a labor of love as I have a number of Mini PCs for my "homelab" (TrueNAS, piHole, Plex, basic stuff). It uses Gemini to extract specs from listings (since they're not often strongly categorized). Comments URL: https://news.ycombinator.com/item?id=48612432 Points: 2 # Comments: 0
Article URL: https://smarterarticles.co.uk/when-ai-files-your-taxes-who-pays-when-it-fails Comments URL: https://news.ycombinator.com/item?id=48612334 Points: 4 # Comments: 0
