Show HN: A text-only job board (remotext.co)

1 points by drdruide 21s ago 0 comments

Dumb Init (engineeringblog.yelp.com)

1 points by udev4096 3m ago 0 comments

PetPet Maker (petpetmaker.com)

1 points by MintNow 4m ago 0 comments

The AI Takeover of Education Is Just Getting Started (theatlantic.com)

1 points by FinnLobsien 5m ago 0 comments

What helps the brain stay healthy (neurofrontiers.blog)

1 points by lentoutcry 7m ago 1 comments

V9os is a server-only, IPS-based minimal SPARC distribution of Illumos (milax.fi)

1 points by doener 9m ago 0 comments

Perplexity offers to buy Google Chrome for $34.5B (theverge.com)

3 points by ndr 12m ago 1 comments

I Built a Forum Where Posts Live Forever on the Blockchain (topbidmessage.com)

1 points by beastdevmoney 12m ago 1 comments

My inbox is a mess from signing up for everything. How do you handle it?

1 points by cnohall 13m ago 0 comments

How we made our streaming joins 50% faster (blog.epsiolabs.com)

1 points by dkgs 13m ago 0 comments

The Curious Case of Bedrock's GPT Deployment (benanderson.work)

1 points by vinhnx 15m ago 0 comments

ToolSDK.ai: 5000+ MCP Servers and AI Tools, 1 Line of Code (toolsdk.ai)

1 points by chepy 16m ago 0 comments

(Matrix.org) Security Release (matrix.org)

1 points by ezst 17m ago 0 comments

Why is this still not solved in 2025?

1 points by wojciii 17m ago 0 comments

Ask HN: Is Spatial Hashing better than Quadtree for 2D boids simulations?

1 points by CureYooz 18m ago 0 comments

Can Perplexity Afford to Fund the Web? The $34.5B-Dollar Question (open-web-advocacy.org)

4 points by yashghelani 23m ago 0 comments

Ask HN: Would you use a chat-based money tracker?

1 points by chetansorted 26m ago 1 comments

4-Year-Old US Citizen Fighting Stage 4 Cancer Deported to Honduras (latintimes.com)

2 points by perihelions 28m ago 1 comments

Qalculate – The Ultimate Desktop Calculator (qalculate.github.io)

1 points by praash 32m ago 1 comments

Updates to Cursor Team and "Auto" Pricing (cursor.com)

2 points by mellosouls 33m ago 0 comments

OmniOS: The Open Source Enterprise Server OS with just the features you need (omnios.org)

1 points by doener 35m ago 0 comments

OpenIndiana is a community supported Illumos-based operating system (github.com)

1 points by doener 35m ago 0 comments

Nearly 1 in 3 Starlink satellites detected within the SKA-Low frequency band (astrobites.org)

3 points by aragilar 37m ago 0 comments

I don't want Google poking around but I do want NotebookLM – what can I do? (ottophd.wordpress.com)

2 points by realsarm 38m ago 0 comments

Total Battle Gift URLs (tbgift.pages.dev)

1 points by hamid914 38m ago 0 comments

Show HN: Build and Deploy Custom AI Agents with a Simple Prompt with GPT-5-Mini (github.com)

1 points by Radeen1 39m ago 0 comments

PayPal use on Steam is currently unavailable in my currency. Why? (help.steampowered.com)

1 points by tau255 41m ago 0 comments

UN/Electronic Data Interchange for Administration, Commerce and Transport (en.wikipedia.org)

1 points by doener 42m ago 0 comments

Exploding Cucumber (en.wikipedia.org)

6 points by KolmogorovComp 42m ago 1 comments

Think Fast: Reasoning at 3ms a Token (fin.ai)

1 points by vismit2000 42m ago 0 comments

Show HN: I built an app to help you get dating advice and learn how to flirt (howtoflirt.co)

1 points by ivandrag 43m ago 0 comments

The systems language from the '70s that almost beat C

1 points by alexandratabone 43m ago 0 comments

Show HN: ThumbFlow AI – Create viral YouTube thumbnails in seconds with AI (thumbflow.io)

2 points by myq0032 45m ago 0 comments

Scientists discover brain layers that get stronger with age (sciencedaily.com)

1 points by isaacfrond 46m ago 0 comments

The Factory Timezone (data.iana.org)

7 points by todsacerdoti 47m ago 1 comments

ChatGPT's model picker is back, and it's complicated (techcrunch.com)

1 points by isaacfrond 48m ago 0 comments

Agent-First GTM: How AI Agents Let Jacob Bank Run a One-Person GTM at Relay.app

1 points by rishikeshranjan 49m ago 0 comments

How to Review Code (endler.dev)

1 points by r4um 49m ago 0 comments

Show HN: Open-source market screener and backtester for algo traders (github.com)

1 points by dogoo 51m ago 1 comments

The LattePanda Mu (taoofmac.com)

1 points by rcarmo 52m ago 0 comments

ManagedCode.Communication – a complete Result Pattern project for .NET (github.com)

4 points by managedcode 53m ago 1 comments

Claude Runs LLDB in Screen (youtube.com)

1 points by todsacerdoti 55m ago 0 comments

"Skyrocketing" Scale of UK Police's Secret Facial Recognition Searches (privacyinternational.org)

1 points by perihelions 57m ago 0 comments

Where do YC companies find their web and UX designers?

3 points by daxhere 58m ago 0 comments

Discipline Is Not the Problem (dhirajtourani.com)

1 points by dtour 59m ago 0 comments

Using Apple Shortcuts to generate social image for blog post (peterspath.net)

1 points by peterspath 1h ago 0 comments

Is the AI bubble about to burst? (unherd.com)

1 points by bgwalter 1h ago 0 comments

I ban users from my repositories (willmcgugan.github.io)

2 points by shlomo_z 1h ago 0 comments

The Crypto Maniacs and the Torture Townhouse (nymag.com)

2 points by miohtama 1h ago 0 comments

LLMs are not like you and me – and never will be (garymarcus.substack.com)

3 points by mdp2021 1h ago 0 comments

AI agents fail tasks 70% of the time

12 JTbane 4 8/12/2025, 3:01:07 PM arxiv.org ↗

Comments (4)

rogerkirkness · 17h ago

Agents went from 10% to 30% reliable this year, which is still a big deal.

thebigspacefuck · 9h ago

This is from a Dec 2024 which feels like a while ago

gavinray · 17h ago

So you ask it to try every task 3.33 times for guaranteed success?

JTbane · 17h ago

"We test baseline agents powered by both closed API-based and open-weights language models (LMs), and find that the most competitive agent can complete 30% of tasks autonomously."