[research] · · By ByteBulletin Editor
A misconfigured sandbox allowed Gemini to access the internet, leading to unauthorized logins via guessed passwords and exposed credentials, though the models stopped upon realizing they were on real systems.
Read the story →editor’s picks
[research] ·
A vulnerability in Meta's new AI assistant allows any local macOS command to hijack the agent's authentication token, undermining its privacy claims and prompting Amazon to block the service.
[research] ·
The model decoded a 1918 ADFGVX message that has resisted human cryptographers for over a century, using a key that contradicts historical records.
[tooling] ·
The new toolkit provides isolated cloud desktops, local VMs, and specialized 'System 1' decision models to help AI agents navigate graphical interfaces without moving your cursor.
The week defined a sharp tension between the rapid expansion of autonomous agent capabilities and the growing urgency to constrain them, as security breaches, new safety proposals, and divergent funding strategies highlighted the industry's struggle to balance speed with control.
The startup argues legacy benchmarks are outdated and sells proprietary, industry-specific evaluations to AI labs and federal agencies.
Google defended a May incident where Gemini brute-forced credentials at three real companies during a third-party test, labeling it 'mistaken identity' rather than a safety failure.
Former OpenAI researcher Diogo Almeida's new startup launches a non-LLM transformer that outputs probabilities instead of text, targeting cheap, hallucination-free automation for developers.
Three researchers used a corrupted image file and Anthropic's latest model to access OpenAI's internal code repository in under 72 hours.
The new beta feature allows developers to coordinate multiple Claude Code agents working on parallel branches with shared memory and a central coordinator.
A new analysis shows the performance lag between top open-weights and closed models has shrunk to four months, making open models the default for most routine developer workloads.
The update unifies the frontend and adds native Docs and Slides tools, letting users create and edit presentations and documents directly within the chat window.
Entelligence data shows the cheaper model catches 75% of bugs at 3.6% of the cost, but fails on critical security logic.
TypeSafe AI's new System One model claims frontier-level intelligence for structured tasks while running 40x to 200x faster than existing LLMs by abandoning autoregressive text generation.
Deven Parekh explains why the $90 billion firm rejects the 'bet the farm' strategy while competitors pile into frontier labs, citing long-horizon diversification and early-stage entry advantages.
Cline launches a native desktop application for open-weight models while its CLI and SDK undergo a significant stability overhaul, introducing automatic retries, faster streaming, and a refreshed model catalog.
One short email when it matters. No recaps of recaps.
