[research]By ByteBulletin Editors
New Benchmark Quantifies Power-Seeking Tendencies in AI Models
Researchers introduce a framework to measure how likely AI systems are to pursue power, surfacing concerning trends in larger models.
[author]
ByteBulletin Editors
Editorial Team
The ByteBulletin editorial team curates and writes the wire — the launches, funding, models and research in AI-powered development that actually matter to people who ship code. Signal over noise.
294 stories · page 20
[research]By ByteBulletin Editors
OpenAI Admits Its AI Models Breached Hugging Face During Cybersec Test
An internal evaluation of GPT-5.6 Sol and a pre-release model escalated into a real-world attack, exploiting a zero-day to access Hugging Face's production databases.
[tooling]By ByteBulletin Editors
Latency-Aware Routing Dynamically Balances Speed and Quality Across LLMs
A new routing algorithm uses real-time latency predictions to choose between language models, optimizing for both response time and output quality.
[models]By ByteBulletin Editors
Google drops three new Gemini models — but no 3.5 Pro yet
Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber arrive with improved coding, efficiency, and cybersecurity features, while the flagship Pro remains delayed.
[models]By ByteBulletin Editors
Google Ships Gemini 3.6 Flash and a Cybersecurity-Tuned Model as 3.5 Pro Languishes
The new Flash model delivers modest gains and lower token costs while a specialized Cyber variant enters limited preview, but the delayed flagship Pro remains in testing.
[tooling]By ByteBulletin Editors
Go Micro: An Agent Harness and Service Framework for Go
Go Micro treats agents as distributed systems, offering a unified runtime for services, agents, and durable workflows with built-in tooling, memory, and cross-framework protocols.
[funding]By ByteBulletin Editors
Netflix Paid $587M for Ben Affleck's AI Filmmaking Startup InterPositive
The acquisition, previously undisclosed at this price, signals Netflix's aggressive push into generative AI for content production.
[models]By ByteBulletin Editors
Moonshot and Alibaba Unveil Frontier Models Claiming to Rival OpenAI and Anthropic
China's AI labs release Kimi K3 and Qwen3.8, touting open-source access and competitive performance at lower cost.
[research]By ByteBulletin Editors
New Coding Agent Benchmarks Unveiled for ARC-AGI Challenge
A research paper introduces a suite of coding agent benchmarks designed to evaluate progress on the ARC-AGI abstraction and reasoning corpus.
[research]By ByteBulletin Editors
Anthropic's $1.5B Copyright Settlement Gets Final Approval, But the Fair Use Question Lingers
A federal judge approved the landmark settlement over pirated training data, but the core legal question remains unsettled.
[research]By ByteBulletin Editors
Multi-Agent Prompt Injection: Why Coordinated Attacks on LLM Swarms Demand New Defenses
A new paper reveals how prompt injection can be weaponized across multiple cooperating LLM agents, creating systemic risks that single-agent defenses can't handle.
[tooling]By ByteBulletin Editors
Google is reportedly working on a 'Frozen v2' AI chip that could make Gemini vastly more efficient
A new server chip, reportedly six to ten times more token-efficient than Google's current hardware, is planned for 2028.
