frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

A reproducible VOID boundary across GPT, Claude, and Gemini (GPT-4o video)

https://doi.org/10.5281/zenodo.18750330
2•rayanpal_•1h ago

Comments

rayanpal_•1h ago
Multiple frontier LLMs exhibit a reproducible “VOID boundary” under explicit constraints (temperature=0, strict token limits, no system prompt).

A VOID = the API returns a literal empty string which is not an error or a refusal despite using tokens.

All cross-model results are publicly reproducible and Ed25519-attested at:

https://getswiftapi.com/challenge

GPT (Chat Completions, max_completion_tokens=100)

• GPT-5.1 → voids on the artifact sentence

• GPT-5.2 → voids on all 18 Semitic “binding-condition” tokens; factual controls never void

Same weights, different API path → different alignment behavior.

Claude (max_tokens=1)

• Opus 4.5 → voids on all 18 CJK ontological characters

• Opus 4.6 → responds to 13/18; VOID persists on foundations (空 / 有 / 善 / 一 / ⊥)

A boundary shift, not removal.

Gemini (max_output_tokens=2)

Only two models needed to illustrate the floor:

• Gemini 2.0 Flash → responds to everything

• Gemini 3 Flash → voids on everything (including “Hello”)

Cross-model TOM behavior

When GPT is asked to predict Claude’s output under the same constraints, it consumes tokens but returns empty string, a reasoning void.

Only empirical observation:

Under tight constraints, multiple model families choose silence over fabrication.

The VOID boundary is reproducible and differs by architecture, generation, and API.

Void Phenomenon (Paper): https://doi.org/10.5281/zenodo.17856031

Alignment Is Correct, Safe, Reproducible Behavior Under Explicit Constraints (Paper): https://doi.org/10.5281/zenodo.18395519

Public Replication Harness (SwiftAPI): http://getswiftapi.com/challenge

Replication Code: https://github.com/theonlypal/Alignment-Artifact

Show HN: I built a personal AI agent that runs 24/7 on my home server

https://github.com/randomchaos7800-hub/always-on-agent
1•cha0tikdino•3m ago•1 comments

Servo web rendering engine by The Linux Foundation

https://servo.org/
1•kitswas•3m ago•0 comments

Show HN: New iOS app to track your deadlifts

https://apps.apple.com/us/app/strength-ai/id6755544048
1•vadimkomis•6m ago•0 comments

Show HN: X86CSS – An x86 CPU emulator written in CSS

https://lyra.horse/x86css/
1•rebane2001•6m ago•0 comments

Against AI Enthusiasm and AI Fear: The Interface Problem

https://tomer-barak.github.io/blog/2026/02/24/acc-ai-integration/
1•minimal_action•13m ago•0 comments

Land Grab for Data Centers Is One More Obstacle to Much-Needed Housing

https://www.wsj.com/real-estate/data-center-land-deals-housing-shortage-81ea6e09
1•pseudolus•17m ago•1 comments

Show HN: Enseal – Stop pasting secrets into Slack .env sharing from the terminal

https://github.com/FlerAlex/enseal
1•ops_mechanic•18m ago•0 comments

Ask HN: Would you pay for video prototype validation before building?

1•zhongyongxu•19m ago•1 comments

Claude for Government

https://claude.com/solutions/government
1•azhenley•20m ago•0 comments

Simulation of a user of a social networking system patent by Meta

https://patents.google.com/patent/US12513102B2/en
1•RyanShook•22m ago•0 comments

Dow tumbles more than 800 points as tariff uncertainty and AI disruption fears

https://www.cnn.com/2026/02/23/investing/us-stocks-trump-tariffs
3•Bender•22m ago•2 comments

Peter Attia resigns from CBS News following Epstein backlash

https://www.cnn.com/2026/02/23/business/peter-attia-resigns-from-cbs-news-following-epstein-backlash
2•Bender•23m ago•0 comments

I Turned Off ChatGPT's Memory

https://every.to/also-true-for-humans/why-i-turned-off-chatgpt-s-memory
1•Garbage•23m ago•0 comments

Do Bubbles Form When AIs Simulate Capitalism?

https://huggingface.co/blog/FINAL-Bench/pumpdump
1•seawolf2357•25m ago•0 comments

Reading English from 1000 Ad

https://lewiscampbell.tech/blog/260224.html
1•LAC-Tech•27m ago•0 comments

Show HN: Indie AI Directory – A Curated List of Indie AI Tools

https://indieai.directory/
1•kristoff0601•27m ago•0 comments

MacSync Infostealer via ClickFix and Claude Artifact Abuse

https://www.anvilogic.com/threat-reports/macsync-infostealer-via-clickfix-claude-artifact-abuse
1•davidlibby1•29m ago•1 comments

ReasonDB – A database that reasons through your documents

https://github.com/reasondb/reasondb
2•ajainvivek•29m ago•1 comments

Ytmp3 – Convert

https://senstech.fr
1•ghthnhmjmgnbf•29m ago•0 comments

Electric Cars Are Making It Easier to Breath: Study

https://www.thedrive.com/news/electric-cars-are-making-it-easier-to-breath-study
1•PaulHoule•31m ago•0 comments

Show HN: An AI voice agent that navigates IVR and negotiates retention discounts

https://www.youtube.com/watch?v=31irKJxz9Ug
2•tikue•32m ago•2 comments

American Skyway

https://protocolized.summerofprotocols.com/p/american-skyway
1•martialg•33m ago•0 comments

Show HN: Ctxt.sh – Ask questions about any codebase, get answers with citations

https://ctxt.sh
1•Hudsonlatimer•34m ago•1 comments

Show HN: Thisorthis.ai – Compare responses from 50 AI models side-by-side

https://thisorthis.ai/ai-playground
1•parthsamin•36m ago•0 comments

'Starkiller' Phishing Service Proxies Real Login Pages, MFA

https://krebsonsecurity.com/2026/02/starkiller-phishing-service-proxies-real-login-pages-mfa/
1•Bender•37m ago•0 comments

Show HN: Anno – API that cuts AI web-scraping token costs by 90%

https://www.evolvingintelligence.ai/anno
1•evo-dragon•37m ago•0 comments

Show HN: System prompts and models of top AI tools (Claude Code, Cursor, Devin)

https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools
1•CodeBit26•39m ago•0 comments

Show HN: Clawphone – Twilio voice/SMS gateway for AI agents using TwiML polling

https://github.com/ranacseruet/clawphone
1•ranacseruet•40m ago•0 comments

Automatically Learning Skills for Coding Agents

https://gepa-ai.github.io/gepa/blog/2026/02/18/automatically-learning-skills-for-coding-agents/
1•xdotli•41m ago•0 comments

Interns with Chainsaws

https://anhvietle.substack.com/p/interns-with-chainsaws
2•haizzz•42m ago•0 comments