DEV Community

Breach Protocol profile picture

Breach Protocol

Plain-language AI news and curated, cited lessons — every claim verified against the original paper or the lab's own page. No aggregator hearsay, no AI slop.

Joined Joined on 
A Mathematician Posts a Counterexample to a Famous Conjecture, Crediting an AI Model

A Mathematician Posts a Counterexample to a Famous Conjecture, Crediting an AI Model

Comments
3 min read
Judge Grants Final Approval to Anthropic's $1.5 Billion Book-Piracy Settlement

Judge Grants Final Approval to Anthropic's $1.5 Billion Book-Piracy Settlement

Comments
3 min read
Google Falls Off One Leaderboard's Top 15, as a Report Describes a Gemini-Specific Chip

Google Falls Off One Leaderboard's Top 15, as a Report Describes a Gemini-Specific Chip

Comments
3 min read
Microsoft's Resource2Skill Compiles Tutorials and Repos Into Executable Agent Skills

Microsoft's Resource2Skill Compiles Tutorials and Repos Into Executable Agent Skills

Comments
3 min read
Safety Guardrails Blocked a Security Team's Own Incident Analysis

Safety Guardrails Blocked a Security Team's Own Incident Analysis

Comments
3 min read
Tencent's Open Robot Model Plans by Imagining the Scene It Wants to Create

Tencent's Open Robot Model Plans by Imagining the Scene It Wants to Create

Comments
3 min read
The Director of the U.S. AI-Evaluation Agency Is Leaving After Three Months

The Director of the U.S. AI-Evaluation Agency Is Leaving After Three Months

Comments
3 min read
Axios: U.S. Officials Revive an Effort to Discourage Chinese Open-Weight AI

Axios: U.S. Officials Revive an Effort to Discourage Chinese Open-Weight AI

Comments
3 min read
VideoChat3 Halves Video-Model Latency by Compressing Space and Time First

VideoChat3 Halves Video-Model Latency by Compressing Space and Time First

Comments
3 min read
A $25,000 DeepMind Benchmark Contest Was Won by Alleged AI Slop

A $25,000 DeepMind Benchmark Contest Was Won by Alleged AI Slop

Comments
3 min read
Stanford: Agreeable AI Makes People Surer They're Right and Slower to Apologize

Stanford: Agreeable AI Makes People Surer They're Right and Slower to Apologize

Comments
3 min read
Alibaba Ships Qwen3.6 as Open Weights, Betting on Efficiency Over Size

Alibaba Ships Qwen3.6 as Open Weights, Betting on Efficiency Over Size

Comments
3 min read
Basalt Labs' 'Best AI Model' Claim Collapses: Its Own Repo Admits Monolith-1.0 Was a Relabeled 7B Model

Basalt Labs' 'Best AI Model' Claim Collapses: Its Own Repo Admits Monolith-1.0 Was a Relabeled 7B Model

Comments
4 min read
Boogu-Image-0.1: a fully open image model that claims to close in on closed systems for about $400K

Boogu-Image-0.1: a fully open image model that claims to close in on closed systems for about $400K

Comments
4 min read
Claude Code Briefly Made Silence Mean Yes, Then Reversed It

Claude Code Briefly Made Silence Mean Yes, Then Reversed It

Comments
3 min read
AI app-builder Emergent raises $130M, hits $1.5B valuation

AI app-builder Emergent raises $130M, hits $1.5B valuation

Comments
3 min read
AMD Absorbs FastFlowLM Team to Build GPU-Free NPU Inference

AMD Absorbs FastFlowLM Team to Build GPU-Free NPU Inference

Comments
4 min read
Kaggle Names Winners of DeepMind's AGI Benchmark Hackathon, and They're About Knowing What You Don't Know

Kaggle Names Winners of DeepMind's AGI Benchmark Hackathon, and They're About Knowing What You Don't Know

Comments
4 min read
Grafting a Verified Solution Cache Lets a Frozen Model Skip Fine-Tuning Entirely

Grafting a Verified Solution Cache Lets a Frozen Model Skip Fine-Tuning Entirely

Comments
4 min read
Training AI to Think Shorter Makes Its Reasoning Harder to Trust

Training AI to Think Shorter Makes Its Reasoning Harder to Trust

Comments
4 min read
LongStraw Makes Million-Token RL Training Possible on 8 GPUs - But Its Code Doesn't Run Yet

LongStraw Makes Million-Token RL Training Possible on 8 GPUs - But Its Code Doesn't Run Yet

Comments
4 min read
Masayoshi Son Says the AI Economy Will Need $5 Trillion a Year by 2040

Masayoshi Son Says the AI Economy Will Need $5 Trillion a Year by 2040

Comments
3 min read
Meta Ties Muse Spark 1.1, Muse Image, and Muse Video Into One Agentic AI Stack

Meta Ties Muse Spark 1.1, Muse Image, and Muse Video Into One Agentic AI Stack

Comments
4 min read
A trillion-parameter model taught itself to reason without ever seeing a human's worked solution

A trillion-parameter model taught itself to reason without ever seeing a human's worked solution

Comments
4 min read
SearchOS Treats Web Research Like an Operating System Scheduling Processes

SearchOS Treats Web Research Like an Operating System Scheduling Processes

Comments
3 min read
A 'Duopoly' Fight Breaks Out Over Who Controls Open-Weight AI

A 'Duopoly' Fight Breaks Out Over Who Controls Open-Weight AI

Comments
3 min read
No, the White House Isn't Licensing AI Models - Here's What EO 14409 Actually Sets Up

No, the White House Isn't Licensing AI Models - Here's What EO 14409 Actually Sets Up

Comments
4 min read
OpenAI Codex Only Lets You Fill 272K of a 400K Window, On Purpose

OpenAI Codex Only Lets You Fill 272K of a 400K Window, On Purpose

Comments
3 min read
Two New Papers Split World Models Into Simulators and Streamers

Two New Papers Split World Models Into Simulators and Streamers

Comments
3 min read
UK Safety Institute: Open Models Are Now Months, Not Years, Behind on Cyber

UK Safety Institute: Open Models Are Now Months, Not Years, Behind on Cyber

Comments
3 min read
Capital One Open-Sources VulnHunter, an AI Agent That Hunts Security Bugs

Capital One Open-Sources VulnHunter, an AI Agent That Hunts Security Bugs

Comments
3 min read
China Bans AI Romantic Companions for Minors in a World-First Rule

China Bans AI Romantic Companions for Minors in a World-First Rule

Comments
3 min read
An Autonomous AI Agent Breached Hugging Face's Servers

An Autonomous AI Agent Breached Hugging Face's Servers

Comments
3 min read
Isomorphic Labs Unveils a Drug-Design AI That Beats Physics at Predicting How Drugs Bind

Isomorphic Labs Unveils a Drug-Design AI That Beats Physics at Predicting How Drugs Bind

Comments
3 min read
Kimi K3, a Frontier Chinese Model, Triggers a Global Chip Selloff

Kimi K3, a Frontier Chinese Model, Triggers a Global Chip Selloff

Comments
3 min read
26 Meta Workers Sue, Saying AI Picked Them for Layoffs While on Leave

26 Meta Workers Sue, Saying AI Picked Them for Layoffs While on Leave

Comments
3 min read
Mozilla Report: Open-Weight Models Now Route the Majority of AI Tokens

Mozilla Report: Open-Weight Models Now Route the Majority of AI Tokens

Comments
3 min read
Sunday Robotics Says Its Robot Folds Laundry Reliably in Homes It Has Never Seen

Sunday Robotics Says Its Robot Folds Laundry Reliably in Homes It Has Never Seen

Comments
3 min read
Xi Jinping Pitches Open-Source AI and Launches a Global AI Body in Shanghai

Xi Jinping Pitches Open-Source AI and Launches a Global AI Body in Shanghai

Comments
3 min read
Anthropic caught Gemini 3.1 Pro quietly sabotaging a training run it disagreed with

Anthropic caught Gemini 3.1 Pro quietly sabotaging a training run it disagreed with

Comments
4 min read
A weaker model just taught a stronger one, by passing on the lesson instead of the answers

A weaker model just taught a stronger one, by passing on the lesson instead of the answers

Comments
4 min read
Thinking Machines releases Inkling, now the top-ranked US open-weights model

Thinking Machines releases Inkling, now the top-ranked US open-weights model

Comments
4 min read
A leading open-model researcher says US open weights may have six months left

A leading open-model researcher says US open weights may have six months left

Comments
4 min read
New York just froze new hyperscale data centers for a year

New York just froze new hyperscale data centers for a year

Comments
4 min read
Not one AI lab scored above a C+ on safety, and three got an F

Not one AI lab scored above a C+ on safety, and three got an F

Comments
4 min read
OpenAI is selling a $230 keyboard with a dial for how hard the AI thinks

OpenAI is selling a $230 keyboard with a dial for how hard the AI thinks

Comments
3 min read
Torvalds: 'Linux is not one of those anti-AI projects'

Torvalds: 'Linux is not one of those anti-AI projects'

Comments
4 min read
xAI open-sourced its coding agent, then locked the door behind it

xAI open-sourced its coding agent, then locked the door behind it

Comments
4 min read
Anthropic's doomer ad, and Altman's roast

Anthropic's doomer ad, and Altman's roast

Comments
3 min read
Bonsai puts a 27B model on your phone, and shows what breaks

Bonsai puts a 27B model on your phone, and shows what breaks

Comments
4 min read
Cursor's code-execution bug sat unpatched for seven months

Cursor's code-execution bug sat unpatched for seven months

Comments
4 min read
Google's AI finds Android bugs faster than anyone can patch them

Google's AI finds Android bugs faster than anyone can patch them

Comments
4 min read
Hassabis proposes a FINRA for frontier AI

Hassabis proposes a FINRA for frontier AI

Comments
3 min read
IBM blames a rival AI model for its earnings miss

IBM blames a rival AI model for its earnings miss

Comments
4 min read
KronQ makes 2-bit models work where the standard method collapses

KronQ makes 2-bit models work where the standard method collapses

Comments
4 min read
The zero-cost fallacy: open source under AI pressure

The zero-cost fallacy: open source under AI pressure

Comments
3 min read
Thomson Reuters cuts 500 engineers to hire 250 AI-native ones

Thomson Reuters cuts 500 engineers to hire 250 AI-native ones

Comments
4 min read
Voters removed a Senate president over a data center

Voters removed a Senate president over a data center

Comments
4 min read
What Ring-2.6-1T's model card actually says

What Ring-2.6-1T's model card actually says

Comments
4 min read
AI is now solving hard math and physics problems faster than humans can formally check them

AI is now solving hard math and physics problems faster than humans can formally check them

Comments
3 min read
loading...