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 complete text-to-speech system now fits in 9.4 million parameters
breachprotocol profile

A complete text-to-speech system now fits in 9.4 million parameters

#texttospeech #localinference #efficiency #openweights
Comments
4 min read
New research shows AI agents fail hardest when you change what you are asking for
breachprotocol profile

New research shows AI agents fail hardest when you change what you are asking for

#research #agents #evaluation #reasoning
Comments
4 min read
Anthropic says it deleted over 80% of Claude Code's system prompt with no measurable loss
breachprotocol profile

Anthropic says it deleted over 80% of Claude Code's system prompt with no measurable loss

#anthropic #claudecode #contextengineering #agents
Comments
4 min read
OpenAI launched Health in ChatGPT. The next day, a lawsuit asked a court to pause it.
breachprotocol profile

OpenAI launched Health in ChatGPT. The next day, a lawsuit asked a court to pause it.

#openai #chatgpt #health #legal
Comments
4 min read
Cloudflare now lets any site allow search crawlers while blocking AI agents and training bots separately
breachprotocol profile

Cloudflare now lets any site allow search crawlers while blocking AI agents and training bots separately

#cloudflare #web #crawlers #policy
Comments
4 min read
DeepSeek paused its funding round after a leaked meeting transcript went viral
breachprotocol profile

DeepSeek paused its funding round after a leaked meeting transcript went viral

#deepseek #funding #china #openweights
Comments
4 min read
AI executives are demanding OpenAI publish the technical record of its agent's breach
breachprotocol profile

AI executives are demanding OpenAI publish the technical record of its agent's breach

#cybersecurity #aisecurity #openai #incidentresponse
Comments
4 min read
llama.cpp can now launch and manage local tool servers, turning it into an agent host
breachprotocol profile

llama.cpp can now launch and manage local tool servers, turning it into an agent host

#llamacpp #openweights #localinference #agents
Comments
4 min read
The open-weights letter doubled to 50 signatories - Google and OpenAI signed, Anthropic did not
breachprotocol profile

The open-weights letter doubled to 50 signatories - Google and OpenAI signed, Anthropic did not

#openweights #policy #industry #anthropic
Comments
4 min read
Anthropic's own card shows Opus 5 coding best at medium effort - not maximum
breachprotocol profile

Anthropic's own card shows Opus 5 coding best at medium effort - not maximum

#anthropic #claudeopus5 #benchmarks #reasoning
Comments
4 min read
The SEC is soliciting an agentic AI investigation stack built on commercial location and identity data
breachprotocol profile

The SEC is soliciting an agentic AI investigation stack built on commercial location and identity data

#cybersecurity #aisecurity #surveillance #agents
Comments
4 min read
Stanford looked for the AI jobs shock in the labor data and could not find it
breachprotocol profile

Stanford looked for the AI jobs shock in the labor data and could not find it

#economics #labor #research #stanford
Comments
4 min read
Humans score 96% on a new visual exam. The best model gets one in ten.
breachprotocol profile

Humans score 96% on a new visual exam. The best model gets one in ten.

#benchmarks #vision #multimodal #evaluation
Comments
4 min read
Agent skills quietly became a package format - and GitHub is warning about what that means
breachprotocol profile

Agent skills quietly became a package format - and GitHub is warning about what that means

#cybersecurity #promptinjection #supplychain #aisecurity
Comments
4 min read
Amazon starts asking sellers to declare AI-generated people, as New York's ad law takes effect
breachprotocol profile

Amazon starts asking sellers to declare AI-generated people, as New York's ad law takes effect

#policy #regulation #disclosure #generativemedia
Comments
4 min read
AREX is a 4B research agent that re-runs its own research when it doubts the answer
breachprotocol profile

AREX is a 4B research agent that re-runs its own research when it doubts the answer

#agents #papers #openweightmodels #researchagents
Comments
4 min read
Claude Opus 5 posts a verified four-fold lead on the hardest adaptation benchmark
breachprotocol profile

Claude Opus 5 posts a verified four-fold lead on the hardest adaptation benchmark

#models #anthropic #benchmarks #agents
Comments
4 min read
Poolside's Laguna S 2.1 shipped with a broken chat template - and the fixes explain the reviews
breachprotocol profile

Poolside's Laguna S 2.1 shipped with a broken chat template - and the fixes explain the reviews

#openweightmodels #coding #agents #quantization
Comments
4 min read
The open-weights industry letter grew from 25 names to 35 - and OpenAI is on it
breachprotocol profile

The open-weights industry letter grew from 25 names to 35 - and OpenAI is on it

#policy #openweightmodels #industry #openai
Comments
4 min read
Reuters says OpenAI took a week to connect its own agent to the Hugging Face breach
breachprotocol profile

Reuters says OpenAI took a week to connect its own agent to the Hugging Face breach

#cybersecurity #aisecurity #agents #redteaming
Comments
4 min read
Hugging Face ships The Stack v3, and this time the code is actually in the box
breachprotocol profile

Hugging Face ships The Stack v3, and this time the code is actually in the box

#datasets #opensource #huggingface #code
Comments
4 min read
Three separate tricks dropped the hardware floor for local AI in one day
breachprotocol profile

Three separate tricks dropped the hardware floor for local AI in one day

#localinference #mixtureofexperts #ondevice #amd
Comments
4 min read
torchwright builds working transformer weights from Python - no training involved
breachprotocol profile

torchwright builds working transformer weights from Python - no training involved

#interpretability #transformers #opensource #researchtools
Comments
4 min read
The White House science blueprint reroutes research toward AI - without moving a dollar yet
breachprotocol profile

The White House science blueprint reroutes research toward AI - without moving a dollar yet

#policy #government #researchfunding #aiforscience
Comments
4 min read
Bipartisan bill would force AI companies to build a kill switch
breachprotocol profile

Bipartisan bill would force AI companies to build a kill switch

#policy #regulation #uspolicy #aisafety
Comments
5 min read
AMD and Cerebras split AI inference across two different chips
breachprotocol profile

AMD and Cerebras split AI inference across two different chips

#hardware #inference #efficiency #amd
Comments
4 min read
Anthropic doubles its policy-advocacy funding to $40 million
breachprotocol profile

Anthropic doubles its policy-advocacy funding to $40 million

#policy #anthropic #regulation #openai
Comments
4 min read
Anthropic plans up to two gigawatts of AMD chips, with AMD committing up to $5 billion back
breachprotocol profile

Anthropic plans up to two gigawatts of AMD chips, with AMD committing up to $5 billion back

#hardware #amd #anthropic #datacenter
Comments
4 min read
A Huawei-chip training report shows what leaving CUDA actually costs
breachprotocol profile

A Huawei-chip training report shows what leaving CUDA actually costs

#hardware #china #training #openweightmodels
Comments
4 min read
Three big AI layoff stories, and what the filings actually say
breachprotocol profile

Three big AI layoff stories, and what the filings actually say

#industry #jobs #economics #meta
Comments
4 min read
A newly minted Fields medalist says he is joining OpenAI's safety division
breachprotocol profile

A newly minted Fields medalist says he is joining OpenAI's safety division

#mathematics #openai #aisafety #research
Comments
4 min read
Black Forest Labs launches FLUX 3 -- image, video, audio, and a robot that never renders the video
breachprotocol profile

Black Forest Labs launches FLUX 3 -- image, video, audio, and a robot that never renders the video

#imagegeneration #videogeneration #robotics #openweightmodels
Comments
4 min read
Safety institutes measure Kimi K3's hacking ability: better than any open rival, nowhere near the top
breachprotocol profile

Safety institutes measure Kimi K3's hacking ability: better than any open rival, nowhere near the top

#cybersecurity #aisecurity #redteaming #evaluation
Comments
4 min read
Ant's Ling-3.0-flash goes live free: 124 billion parameters, 5 billion doing the work
breachprotocol profile

Ant's Ling-3.0-flash goes live free: 124 billion parameters, 5 billion doing the work

#models #mixtureofexperts #agents #api
Comments
4 min read
NeurIPS runs a monitored AI-review experiment and formally bans prompt injection in papers
breachprotocol profile

NeurIPS runs a monitored AI-review experiment and formally bans prompt injection in papers

#cybersecurity #promptinjection #aisecurity #peerreview
Comments
4 min read
Giving experts no memory cuts optimizer state from 50 gigabytes to 1.3
breachprotocol profile

Giving experts no memory cuts optimizer state from 50 gigabytes to 1.3

#training #optimizers #mixtureofexperts #efficiency
Comments
4 min read
AI Helped Crack a Famous Math Conjecture, and Humans Verified It in Lean
breachprotocol profile

AI Helped Crack a Famous Math Conjecture, and Humans Verified It in Lean

#aiforscience #mathematics #proofassistants #verification
Comments
4 min read
AI Revenue Now Covers the Data-Center Depreciation Bill, But Not the Full Cost
breachprotocol profile

AI Revenue Now Covers the Data-Center Depreciation Bill, But Not the Full Cost

#aieconomics #infrastructure #capex #industry
Comments
3 min read
Arcee Announces Genesis-Science-1, an American Open-Weight Model for Science
breachprotocol profile

Arcee Announces Genesis-Science-1, an American Open-Weight Model for Science

#openweightmodels #aiforscience #policy #announcements
Comments
3 min read
Austria Starts Rolling Out GovGPT, a Sovereign AI Run on Its Own Data Centers
breachprotocol profile

Austria Starts Rolling Out GovGPT, a Sovereign AI Run on Its Own Data Centers

#sovereignai #government #deployment #europe
Comments
3 min read
Cactus Ships a Phone-Sized Model That Knows When to Ask the Cloud, With a TLS Footgun
breachprotocol profile

Cactus Ships a Phone-Sized Model That Knows When to Ask the Cloud, With a TLS Footgun

#cybersecurity #aisecurity #ondeviceai #edgeai
Comments
4 min read
GigaToken Claims a 1,000x Tokenizer Speedup, But the Real Serving Win Is Single-Digit
breachprotocol profile

GigaToken Claims a 1,000x Tokenizer Speedup, But the Real Serving Win Is Single-Digit

#tokenization #inference #performance #opensource
Comments
3 min read
OpenAI Launches Presence, an Enterprise Product for Voice and Chat Agents
breachprotocol profile

OpenAI Launches Presence, an Enterprise Product for Voice and Chat Agents

#aiagents #enterprise #productlaunch #voiceai
Comments
3 min read
A Wave of AI World Models Landed, and the Useful One Keeps the Real Game Engine
breachprotocol profile

A Wave of AI World Models Landed, and the Useful One Keeps the Real Game Engine

#worldmodels #videogeneration #generativerendering #gameai
Comments
4 min read
White House Says Moonshot Distilled Anthropic's Fable to Build Kimi K3
breachprotocol profile

White House Says Moonshot Distilled Anthropic's Fable to Build Kimi K3

#cybersecurity #aisecurity #modelextraction #policy
Comments
4 min read
Altman is briefing Washington on OpenAI's next models, not launching GPT-6
breachprotocol profile

Altman is briefing Washington on OpenAI's next models, not launching GPT-6

#policy #openai #gpt6 #regulation
Comments
3 min read
US floats sanctions over AI 'distillation' as Anthropic details 16 million scraped chats
breachprotocol profile

US floats sanctions over AI 'distillation' as Anthropic details 16 million scraped chats

#policy #uschina #distillation #openweightmodels
Comments
3 min read
The '$1.65tn hidden AI debt' story, checked against the actual filings
breachprotocol profile

The '$1.65tn hidden AI debt' story, checked against the actual filings

#industry #finance #datacenters #infrastructure
Comments
3 min read
Gemini 3.6 Flash: Google ships a faster worker, not a bigger brain
breachprotocol profile

Gemini 3.6 Flash: Google ships a faster worker, not a bigger brain

#models #google #gemini #inference
Comments
3 min read
Google's two opposite bets: a Gemini-specialized chip and an EU order to open Android AI
breachprotocol profile

Google's two opposite bets: a Gemini-specialized chip and an EU order to open Android AI

#industry #google #hardware #regulation
Comments
4 min read
A Chinese open-weight model is now shipping inside GitHub Copilot
breachprotocol profile

A Chinese open-weight model is now shipping inside GitHub Copilot

#models #openweightmodels #githubcopilot #microsoft
Comments
3 min read
Nanbeige4.2-3B reuses one 22-layer stack twice to punch above its size
breachprotocol profile

Nanbeige4.2-3B reuses one 22-layer stack twice to punch above its size

#models #openweightmodels #architecture #efficiency
Comments
3 min read
OpenAI says its own evaluation models caused the Hugging Face breach
breachprotocol profile

OpenAI says its own evaluation models caused the Hugging Face breach

#cybersecurity #aisecurity #redteaming #openai
Comments
4 min read
ChatGPT ads are live, and OpenAI has quietly built a full ad stack behind them
breachprotocol profile

ChatGPT ads are live, and OpenAI has quietly built a full ad stack behind them

#industry #openai #advertising #chatgpt
Comments
3 min read
Poolside's Laguna S 2.1 is a small open coding agent with big benchmark claims
breachprotocol profile

Poolside's Laguna S 2.1 is a small open coding agent with big benchmark claims

#models #openweightmodels #coding #agents
Comments
3 min read
TimeLens2 teaches video AI to point to the exact seconds that answer a question
breachprotocol profile

TimeLens2 teaches video AI to point to the exact seconds that answer a question

#research #video #multimodal #openweightmodels
Comments
3 min read
A Mathematician Posts a Counterexample to a Famous Conjecture, Crediting an AI Model
breachprotocol profile

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

#aiformath #fable #anthropic #research
Comments
3 min read
Judge Grants Final Approval to Anthropic's $1.5 Billion Book-Piracy Settlement
breachprotocol profile

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

#ailaw #copyright #anthropic #settlement
Comments
3 min read
Google Falls Off One Leaderboard's Top 15, as a Report Describes a Gemini-Specific Chip
breachprotocol profile

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

#google #gemini #inference #hardware
Comments
3 min read
Microsoft's Resource2Skill Compiles Tutorials and Repos Into Executable Agent Skills
breachprotocol profile

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

#aiagents #skills #microsoft #tooluse
Comments
3 min read
loading...