Bonsai 27B is a multi-billion parameter model small enough to fit into a smartphone. It’s useful for dev and research work, ...

Lanterns HBO pilot packed five real science concepts into 57 minutes: photon-molecule physics, brain-computer interface ...

Glimmer stakes out different ground: a dense model with native vision input, trained end-to-end around the agent loop, ...

Earlier this week, the AI startup Liquid, formed in by former MIT computer scientists, debuted LFM2.5-2.6B, a new open-weight ...

The biggest AI developments, decoded. 12 Aug 2026.

1Password researchers found LLM-generated fixes often overlooked root causes, architectural context, and long-term security ...

Conduit, a non-invasive brain-computer interface startup, has recruited OpenAI alignment researcher Naomi Bashkansky, arguing ...

NVIDIA Nemotron Lightning delivers 4x throughput for AI agents. This 30B parameter model handles repetitive execution tasks with high speed and low cost.

Encrypted AI reasoning vulnerability in Anthropic, OpenAI, and Google APIs let researchers decode 315,320 published session ...

Meta open-sourced Muse Glimmer, a 30B agentic model that runs entirely on a laptop or consumer GPU, free to download under Apache 2.0.

AI's real fight isn't about the smartest model — it's about who controls scarce compute, data and distribution. Preemption vs. proliferation, explained.

Inference has settled into engineering budgets as a recurring operating expense, sitting beside cloud storage and observability on the invoice. Estimates cited ...