frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

OpenAI Tries to Shift Responsibility to Users

https://buttondown.com/maiht3k/archive/openai-tries-to-shift-responsibility-to-users/
1•cratermoon•17s ago•0 comments

Ask HN: Would you use an app that helps you collect and revisit favorite quotes?

1•relaxy•1m ago•0 comments

Junior Developers Are Turning Productivity Gains to Career Security with AI

https://practicalsecurity.substack.com/p/the-ai-anxiety-antidote-how-junior
1•atilla_bilgic•1m ago•1 comments

On Hack Club: reflections on community, union organizing and authoritarianism

https://place.reeseric.ci/writings/2024-05-05/
1•reesericci•1m ago•0 comments

Chad is a new Y Combinator-backed product so wild, people thought it was fake

https://techcrunch.com/2025/11/12/chad-the-brainrot-ide-is-a-new-y-combinator-backed-product-so-w...
1•gpi•2m ago•0 comments

Claude Code for Web Ruined My Brain (Paul Ford)

https://aboard.com/claude-code-for-web-ruined-my-brain/
1•gbseventeen3331•3m ago•0 comments

DgVoodoo 2

https://dege.freeweb.hu/dgVoodoo2/
1•BruceEel•4m ago•0 comments

Washington Post data breach impacts nearly 10K employees, contractors

https://www.bleepingcomputer.com/news/security/washington-post-data-breach-impacts-nearly-10k-emp...
2•WaitWaitWha•4m ago•0 comments

I've created a website to track the team's activity

https://clickhouse.com/blog/velocity
1•samaysharma•4m ago•0 comments

A Master Table of Truth: Lawyers Using AI

https://craigball.net/2025/11/04/a-master-table-of-truth/
2•WaitWaitWha•6m ago•0 comments

Redmine 6.1 is now available

https://www.redmine.org/news/156
1•mariuz•7m ago•0 comments

Show HN: We charge $10/mo for wealth management that costs hundreds elsewhere

https://wwww.fulfilledwealth.co
2•workworkwork71•8m ago•0 comments

Brooks' Law

https://en.wikipedia.org/wiki/Brooks%27s_law
1•hashim•9m ago•0 comments

Turn Off and Journal Instead

https://basic.bearblog.dev/turn-off-and-journal-instead/
1•speckx•10m ago•0 comments

Claude can identify its 'intrusive thoughts'

https://www.transformernews.ai/p/claude-can-identify-its-intrusive-ai-introspection
1•shakeelhashim•10m ago•0 comments

LLM Chat Platform for iOS, Android, Mac, Windows, and Linux

https://github.com/Chevey339/kelivo
1•james2doyle•11m ago•0 comments

South Korean loan sharks target teen gamblers

https://sigma.world/news/south-korea-loan-sharks-teen-gamblers
2•rawgabbit•13m ago•0 comments

Pequliar is a QR code based compact puzzle sequence

https://pequliar.besttof.nl/
1•gregsadetsky•14m ago•0 comments

Show HN: AI-built tools with security by default

1•zvonimirs•15m ago•0 comments

Proton Launches 8th Annual Lifetime Account Charity Fundraiser

https://proton.me/blog/lifetime-fundraiser-survey-2025
2•PrivacyDingus•15m ago•0 comments

Animalcules and Their Motors

https://press.asimov.com/articles/flagella
1•mailyk•15m ago•0 comments

All Intel GPUs Run on Raspberry Pi and RISC-V

https://www.jeffgeerling.com/blog/2025/all-intel-gpus-run-on-raspberry-pi-and-risc-v
1•mikece•16m ago•0 comments

Show HN: Open-Source NotebookLM Alternative

https://www.noodleflow.ai/notebook
2•nazar_ilamanov•16m ago•0 comments

PDF to TXT Converter Online

https://pdf-to-txt.com
2•Nancy1230•18m ago•0 comments

Robinhood Offers to Bring Cash to Your Doorstep, for a Fee

https://www.wsj.com/finance/investing/robinhood-offers-to-bring-cash-to-your-doorstep-for-a-fee-7...
1•bookofjoe•18m ago•1 comments

Todo iOS app as simple as paper

https://apps.apple.com/dk/app/nauu/id6754856167
1•stoumann•18m ago•0 comments

Canada in the running to headquarter new defence bank

https://www.theglobeandmail.com/business/economy/article-canada-gdp-nato-defence-spending-defence...
1•Teever•20m ago•0 comments

Show HN: Compute CLI – A universal sandbox SDK with direct browser access

https://www.computesdk.com/blog/november-2025-update/
1•heygarrison•21m ago•0 comments

A Hermetic, Transparent Soft Growing Vine Robot System for Pipe Inspection

https://arxiv.org/abs/2510.27010
1•PaulHoule•21m ago•0 comments

Automated PDF Generation with Typst

https://typst.app/blog/2025/automated-generation/
1•leephillips•22m ago•0 comments
Open in hackernews

SIMA 2: An Agent That Plays, Reasons, and Learns with You in Virtual 3D Worlds

https://deepmind.google/blog/sima-2-an-agent-that-plays-reasons-and-learns-with-you-in-virtual-3d-worlds/
33•meetpateltech•1h ago

Comments

Workaccount2•1h ago
>We’ve observed that, throughout the course of training, SIMA 2 agents can perform increasingly complex and new tasks, bootstrapped by trial-and-error and Gemini-based feedback.

>In subsequent training, SIMA 2’s own experience data can then be used to train the next, even more capable version of the agent. We were even able to leverage SIMA 2’s capacity for self-improvement in newly created Genie environments – a major milestone toward training general agents across diverse, generated worlds.

Pretty neat, I wonder how that works with Gemini, I suppose SIMA is a model (agent?) that runs on top of it?

ukuina•58m ago
At 0:52 in their demo video, there is a grammatical inconsistency in the agent's text output. The annotations in the video are therefore suspected to be created by humans after the fact. Is Google up to their old marketing/hyping tricks again?

> SIMA 2 Reasoning:

> The user wants me to go to the ‘tomato house’. Based on the description ‘ripe tomato’, I identify the red house down the street.

m_w_•42m ago
I can't speak to the content of the actual game being played, but it wouldn't surprise me if there was an in-game text prompt:

> "The house that looks like a ripe tomato!"

that was transformed into a "user prompt" in a more instructional format

> "Go to the tomato house"

And both were used in the agent output. At least the Y-axes on the graphs look more reasonable than some other recent benchmarks.

golol•36m ago
The gap between high level and low level control of robots is closing. Right now thousands of hours of task specific training data is being collected and trained on to create models that can control robots to execute specific tasks in specific contexts. This essentially turns the operation of a robot into a kind of video game, where inputs are only needed a in low-dimensional abstract form, such as "empty the dishwasher" or "repeat what I do" or "put your finger in the loop and pull the string". This will be combined with high-level control agents like SIMA 2 to create useful real-world robots.
oersted•24m ago
I get why they do it, they are a business. I just wish Google would get off their ivory tower and build in the open more like they used to (did they? maybe I'm misremembering...).

They've acquired this bad habit of keeping all their scientific experiments closed by default and just publishing press releases. I wish it was open-source by default and closed just when there's a good reason.

Don't get me wrong, I suppose this is more of a compliment. I really like what they are doing and I wish we could all participate in these advances.