情報流

AI 評分的原始信號、趨勢與預測(每日摘要在上方 tab)。

50 items

Position: Vision-Language-Action Models Cannot Be Verified to Perform Physical Reasoning

arXiv:2606.30686v1 Announce Type: new Abstract: Vision-Language-Action (VLA) systems, built on pretrained vision-language models (VLMs), have shown rapidly improving performance on...

10/10

1 source

#VLA#Robotics#Physical Reasoning+2

technologyai_agents4d ago

Have your agent record video demos of its work with shot-scraper video

<a href="https://shot-scraper.datasette.io/en/stable/video.html">shot-scraper video</a> is a new command introduced in today's <a href="https://github.com/simonw/shot-scraper/re...

10/10

1 source

#AI agents#automation#Playwright+3

technologycreator_economy6d ago

Ask HN: Local CJK/Latin entity extraction without shipping a model?

I am building a local-first Mac writing app, and I am stuck on the story-memory layer. The problem: given a long fiction manuscript, I need to extract characters, aliases, relation...

10/10

1 source

#local-first#NLP#CJK+2

technologyai_agents8d ago

U.S. allows Anthropic to release Mythos AI to ‘trusted’ US organizations

Comments

10/10

1 source

#Anthropic#Mythos AI#AI Regulation+2

technologyrobotics9d ago

Reinforcement Learning Enables Autonomous Microrobot Navigation and Intervention in Simulated Blood Capillaries

arXiv:2606.26154v1 Announce Type: new Abstract: Autonomous microrobots navigating biological vasculature could enable targeted drug delivery and thrombolysis, yet training control ...

10/10

1 source

#microrobotics#reinforcement learning#biomedical engineering+2

technologyrobotics9d ago

BFMTrack: Latent Sequence Optimization for Physics-Based Motion Tracking with Behavioral Foundation Models

arXiv:2606.25056v1 Announce Type: new Abstract: Behavioral Foundation Models (BFMs) offer a promising path toward universal physics-based character control by organizing a rich rep...

10/10

1 source

#BFM#Motion Tracking#Latent Sequence Optimization+2

technologyai_agents10d ago

Dense Supervision Is Not Enough: The Readout Blind Spot in Looped Language Models

arXiv:2606.24898v1 Announce Type: new Abstract: Looped language models turn hidden states into runtime state: each state is decoded for prediction and fed back into future computat...

10/10

1 source

#looped language models#recurrent transformers#dense supervision+2

technologyother11d ago

datasette 1.0a35

Release: <a href="https://github.com/simonw/datasette/releases/tag/1.0a35">datasette 1.0a35</a> I'll write more about this one tomorrow, but it's...

9/10

1 source

#datasette#sqlite#open-source+2

technologyai_agents12d ago

Porting the Moebius 0.2B image inpainting model to run in the browser with Claude Code

This morning <a href="https://news.ycombinator.com/item?id=48630171">on Hacker News</a> I saw <a href="https://hustvl.github.io/Moebius/">Moebius: 0.2B Lightweight Image Inpaint...

10/10

1 source

#WebGPU#Claude Code#ONNX+3

otherother15d ago

The Cold War's Accidental Whale Observatory

Comments

8/10

1 source

#marine biology#cold war history#acoustics+2

technologyai_agents15d ago

Ask HN: Using OPA/Rego to secure MCP tool execution. Does it make sense?

Comments URL: https://news.ycombinator.com/item?id=48606728 Points: 5 # Comments: 0

9/10

1 source

#MCP#OPA#Rego+3

technologyai_agents15d ago

Ask HN: Will we start seeing tools for LLM use?

Many projects and addons exists to reduce the verbosity on standard bash / git /npm etc. Commands that agents pass regularly as tools to LLMs. (eg. rtk, headroom, lean-ctx). Tool o...

9/10

1 source

#LLM agents#token optimization#CLI tools+2

businessai_agents19d ago

our little team just shipped its first SaaS! (our founder used to work with NASA and Microsoft, but he decided to start from scratch with fresh idea)

Hey everyone! So today is kind of a big day for our little team and I wanted to share it here but not as promotion, just describing our path. After running Snoika as a marketing ag...

9/10

1 source

#GEO#Generative Engine Optimization#SaaS+2

technologyai_agents20d ago

Story of How Im Running an Unlimited $6/Month AI Provider on 4x RTX 3090s

This submission is a tale about how I launched an unlimited LLM provider to about 60 hyped people on the waitlist, then immediately served them a fully dysfunctional death-loop mod...

10/10

1 source

#LLM#GPU#Inference+3

technologyai_agents21d ago

Ask HN: Is anyone building real software with AI agents?

I've noticed a pattern where the people who are talking about how impressive their agentic workflows are, always seem to use these workflows to build more AI tooling. Has anyone se...

9/10

1 source

#AI agents#software engineering#agentic workflows+2

businessother21d ago

500-year-old monasteries outperform at digital transformation (U. of Zurich)

Comments

10/10

1 source

#digital transformation#organizational culture#long-termism+2

technologydefi21d ago

Publishing WASM wheels to PyPI for use with Pyodide

The <a href="https://blog.pyodide.org/posts/314-release/">Pyodide 314.0 release announcement</a> (via <a href="https://news.ycombinator.com/item?id=48462759">Hacker News</a>) in...

10/10

1 source

#python#webassembly#pyodide+3

technologyai_agents22d ago

Statement on the US government directive to suspend access to Fable 5 and Mythos 5

<a href="https://www.anthropic.com/news/fable-mythos-access">Statement on the US government directive to suspend access to Fable 5 and Mythos 5</a> Well thi...

10/10

1 source

#anthropic#export-controls#ai-regulation+3

businessai_agents24d ago

I've built 25+ MVPs. The funded ones all cut these 5 features

Was archiving dead client folders last month and ended up reading old scoping docs for like three hours instead of doing actual work. We're a small dev shop. 26 MVPs shipped si...

10/10

1 source

#SaaS#MVP#Fundraising+2

technologyai_agents24d ago

Initial impressions of Claude Fable 5

I didn't have early access to today's <a href="https://www.anthropic.com/news/claude-fable-5-mythos-5">Claude Fable 5</a> release, but I've spent the past ~5.5 hours putting it ...

10/10

1 source

#Claude Fable 5#Anthropic#LLM Benchmarking+2

technologyai_agents26d ago

Siri AI at WWDC 2026

Given how badly burned anyone who took Apple's <a href="https://simonwillison.net/2024/Jun/10/apple-intelligence/">2024 WWDC Apple Intelligence announcements</a> at face value w...

10/10

1 source

#apple#siri-ai#vision-llms+3

technologyai_agents28d ago

Running Python code in a sandbox with MicroPython and WASM

I've been experimenting with different approaches to running code in a sandbox for several years now, but my latest attempt feels like it might finally have all of the character...

10/10

1 source

#python#webassembly#micropython+4

technologyother29d ago

Zig Zen Update

Comments

10/10

1 source

#zig#systems-programming#software-architecture+1

technologyother6/5/2026

Quoting Andreas Kling

<blockquote cite="https://ladybird.org/posts/changing-how-we-develop-ladybird/">We will no longer accept public pull requests. [...] A substantial patch used to imply sub...

10/10

1 source

#open-source#generative-ai#ladybird+2

technologyai_agents6/4/2026

Creative Reading: Scaffolding Reading for Transformation

arXiv:2606.04308v1 Announce Type: new Abstract: Reading augmentation systems increasingly help readers process text at scale. While these tools address real constraints of time and...

10/10

1 source

#human-computer interaction#creative reading#cognitive scaffolding+2

technologyother5/31/2026

Running Python ASGI apps in the browser via Pyodide + a service worker

Research: <a href="https://github.com/simonw/research/tree/main/pyodide-asgi-browser#readme">Running Python ASGI apps in the browser via Pyodide + a service wor...

10/10

1 source

#pyodide#wasm#python+4

technologyai_agents5/30/2026

Claude Opus 4.8: "a modest but tangible improvement"

Anthropic shipped <a href="https://www.anthropic.com/news/claude-opus-4-8">Claude Opus 4.8</a> today. My favourite thing about it is this note in the release announcement: ...

10/10

1 source

#Anthropic#Claude Opus#LLM+3

technologyother5/28/2026

Train 1T parameter LLM with 8 GPUs?

See this project https://github.com/Sphere-AI-Lab/orbit. Comments URL: https://news.ycombinator.com/item?id=48305896 Points: 1 # Comments: 0

10/10

1 source

#LLM#Distributed Training#GPU Optimization+2

technologyai_agents5/28/2026

sqlite AGENTS.md

<a href="https://github.com/sqlite/sqlite/blob/master/AGENTS.md">sqlite AGENTS.md</a> SQLite gained an AGENTS.md file <a href="https://github.com/sqlite/sql...

10/10

1 source

#sqlite#ai_agents#open_source+2

technologyai_agents5/27/2026

Multi-Agent LLM System for Automated Vulnerability Discovery and Reproduction

Comments

10/10

1 source

#cybersecurity#llm#multi-agent systems+2

technologyai_agents5/27/2026

I'm Tired of Talking to AI

Comments

9/10

1 source

#AI fatigue#UX design#human-computer interaction+1

technologyai_agents5/27/2026

Notes on Pope Leo XIV's encyclical on AI

Dropped this morning by the Vatican: <a href="https://www.vatican.va/content/leo-xiv/en/encyclicals/documents/20260515-magnifica-humanitas.html">Magnifica Humanitas of His Holin...

10/10

1 source

#ai ethics#vatican#interpretability+3

technologyother5/26/2026

I bypassed AWS API Gateway auth with a trailing slash. Got $12K bounty

Comments

10/10

1 source

#AWS#API Gateway#Cybersecurity+3

businessai_agents5/26/2026

my sales were down and I decided to raise my prices. here's why

i'm not discounting to "win back" volume. that's how you train people to wait for sales. raised price on NEW customers first. keep existing on current plan (grand...

10/10

1 source

#saas#pricing-strategy#customer-segmentation+2

businesscreator_economy5/26/2026

Boss Level MRR from CapCut

Saw this today in my CapCut. Not sure if it's part of an A/B test or a bug as I'm travelling, but $1,300/mo is pretty punchy for a consumer app.   submitted by   /u...

9/10

1 source

#SaaS#Pricing Strategy#CapCut+2

technologyother5/26/2026

Sqlit – A lazygit-style TUI for SQL databases

sqlit is a lazygit-style TUI for SQL. Connect and query your database from the terminal in seconds, no config files or documentation to read first. It supports all the major databa...

9/10

1 source

#SQL#TUI#Developer Tools+2

technologyother5/24/2026

Memory has grown to nearly two-thirds of AI chip component costs

Comments

10/10

1 source

#semiconductors#HBM#AI hardware+2

technologyother5/23/2026

The memory shortage is causing a repricing of consumer electronics

<a href="https://davidoks.blog/p/ai-is-killing-the-cheap-smartphone">The memory shortage is causing a repricing of consumer electronics</a> David Oks provid...

10/10

1 source

#semiconductors#HBM#supply_chain+2

technologyai_agents5/22/2026

Datasette Agent

We just <a href="https://datasette.io/blog/2026/datasette-agent/">announced the first release of Datasette Agent</a>, a new extensible AI assistant for Datasette. I've been work...

10/10

1 source

#datasette#sqlite#llm+4

technologyai_agents5/21/2026

Google I/O, Gemini Spark, Antigravity

It's hard to find much to write about Google I/O this year because I have a policy of not writing about anything that I can't try out myself, and a lot of the big announcements ...

10/10

1 source

#google-io#gemini-spark#antigravity+3

businessbiotech5/21/2026

"An (important) message from Infomaniak's founder"

Hello ..., I'm writing to you as the founder and strategic director of Infomaniak because something important has just happened, and it concerns you directly. I no longer control I...

10/10

1 source

#steward-ownership#digital sovereignty#corporate governance+2

technologyai_agents5/19/2026

Overreliance in Writing Tasks: Exploring Similarity-Based Measures of AI Influence on Writing and Proposing a Reflective Writing Interface Intervention

arXiv:2605.15322v1 Announce Type: new Abstract: As generative AI (GenAI) systems become increasingly proficient at simulating human-like and well-reasoned text, users may attribute...

10/10

1 source

#GenAI#HCI#AI Overreliance+2

technologyai_agents5/19/2026

Ask HN: Why aren't more people worried about AI impersonation in code reviews?

This is something that has bothered me for quite a while, and I don't see a lot of people talk about it: Agents, in most cases, impersonate the human operator, by design, with no w...

9/10

1 source

#AI Agents#Software Engineering#Cybersecurity+3

technologyai_agents5/17/2026

OpenAI and Government of Malta partner to roll out ChatGPT Plus to all citizens

Comments

10/10

1 source

#OpenAI#Malta#Public Policy+2

technologyother5/17/2026

Unknowable Math Can Help Hide Secrets

Comments

10/10

1 source

#cryptography#mathematics#undecidability+2

technologyai_agents5/15/2026

Building ML framework with Rust and Category Theory

Comments

10/10

1 source

#Rust#Category Theory#Machine Learning+2

businessother5/15/2026

Ask HN: How to be SOC2 Type 2 compliant as a solo-entreprenuer?

Is it possible? Do you know success cases w/o spending 20+k $ on auditors? My customers bombards me with question about certification of my app Perfect Wiki, I need help with findi...

9/10

1 source

#SOC2#compliance#solo-entrepreneur+3

technologyai_agents5/15/2026

Quoting Mitchell Hashimoto

<blockquote cite="https://twitter.com/mitchellh/status/2055039647924007222">[...] On the interesting side is how fungible programming languages are nowadays. Programming languag...

10/10

1 source

#zig#rust#llms+3

technologyai_agents5/15/2026

Not so locked in any more

This <a href="https://simonwillison.net/2026/May/14/mitchell-hashimoto/">Mitchell Hashimoto quote</a> about Bun migrating from Zig to Rust reminded me of a similar conversation ...

10/10

1 source

#coding-agents#ai-assisted-programming#software-architecture+3

technologyother5/13/2026

I Moved My Digital Stack to Europe

Comments

10/10

1 source

#digital sovereignty#privacy#GDPR+2

情報流

Position: Vision-Language-Action Models Cannot Be Verified to Perform Physical Reasoning

Have your agent record video demos of its work with shot-scraper video

Ask HN: Local CJK/Latin entity extraction without shipping a model?

U.S. allows Anthropic to release Mythos AI to ‘trusted’ US organizations

Reinforcement Learning Enables Autonomous Microrobot Navigation and Intervention in Simulated Blood Capillaries

BFMTrack: Latent Sequence Optimization for Physics-Based Motion Tracking with Behavioral Foundation Models

Dense Supervision Is Not Enough: The Readout Blind Spot in Looped Language Models

datasette 1.0a35

Porting the Moebius 0.2B image inpainting model to run in the browser with Claude Code

The Cold War&#x27;s Accidental Whale Observatory

Ask HN: Using OPA/Rego to secure MCP tool execution. Does it make sense?

Ask HN: Will we start seeing tools for LLM use?

our little team just shipped its first SaaS! (our founder used to work with NASA and Microsoft, but he decided to start from scratch with fresh idea)

Story of How Im Running an Unlimited $6/Month AI Provider on 4x RTX 3090s

Ask HN: Is anyone building real software with AI agents?

500-year-old monasteries outperform at digital transformation (U. of Zurich)

Publishing WASM wheels to PyPI for use with Pyodide

Statement on the US government directive to suspend access to Fable 5 and Mythos 5

I've built 25+ MVPs. The funded ones all cut these 5 features

Initial impressions of Claude Fable 5

Siri AI at WWDC 2026

Running Python code in a sandbox with MicroPython and WASM

Zig Zen Update

Quoting Andreas Kling

Creative Reading: Scaffolding Reading for Transformation

Running Python ASGI apps in the browser via Pyodide + a service worker

Claude Opus 4.8: "a modest but tangible improvement"

Train 1T parameter LLM with 8 GPUs?

sqlite AGENTS.md

Multi-Agent LLM System for Automated Vulnerability Discovery and Reproduction

I&#x27;m Tired of Talking to AI

Notes on Pope Leo XIV's encyclical on AI

I bypassed AWS API Gateway auth with a trailing slash. Got $12K bounty

my sales were down and I decided to raise my prices. here's why

Boss Level MRR from CapCut

Sqlit – A lazygit-style TUI for SQL databases

Memory has grown to nearly two-thirds of AI chip component costs

The memory shortage is causing a repricing of consumer electronics

Datasette Agent

Google I/O, Gemini Spark, Antigravity

"An (important) message from Infomaniak's founder"

Overreliance in Writing Tasks: Exploring Similarity-Based Measures of AI Influence on Writing and Proposing a Reflective Writing Interface Intervention

Ask HN: Why aren't more people worried about AI impersonation in code reviews?

OpenAI and Government of Malta partner to roll out ChatGPT Plus to all citizens

Unknowable Math Can Help Hide Secrets

Building ML framework with Rust and Category Theory

Ask HN: How to be SOC2 Type 2 compliant as a solo-entreprenuer?

Quoting Mitchell Hashimoto

Not so locked in any more

I Moved My Digital Stack to Europe

The Cold War's Accidental Whale Observatory

I'm Tired of Talking to AI