Ground Truth.
AI, checked against the source.
← 2026-07-172026-07-18later →

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

2026-07-18

Meta's Superintelligence Labs shipped Muse Spark 1.1, a reasoning model built for agentic tasks with a 1-million-token context window, alongside Muse Image and a preview-only Muse Video, wiring all three into a single agentic system distributed through Meta AI, Instagram, and WhatsApp.

meta · agentic-ai · muse · context-windows · tool-use · image-generation

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

2026-07-18

Executive Order 14409 creates a voluntary US government pre-release access framework for "covered frontier models," thresholded by a classified NSA-led cyber-capability benchmark, and explicitly does not authorize mandatory licensing, preclearance, or permitting of AI model releases.

ai-policy · frontier-models · executive-order · nsa · ai-regulation

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

2026-07-18

Basalt Labs claimed its Monolith-1.0 model scored 99.4% on Humanity's Last Exam and was a 1.57-trillion-parameter system, but its own Hugging Face model card now says the publicly released model was an inflated version of the much smaller Qwen 2.5 7B Instruct, and the weights have been pulled.

benchmark-integrity · humanitys-last-exam · open-weight-models · mixture-of-experts · ai-fraud

AMD Absorbs FastFlowLM Team to Build GPU-Free NPU Inference

2026-07-18

AMD announced on July 17, 2026 that the FastFlowLM team has joined its Artificial Intelligence Group to build out an NPU-first, GPU-free local inference stack for Ryzen AI laptops.

amd · npu · inference · efficiency · local-llm · acquisition

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

2026-07-18

Emergent, an AI platform that turns plain-language prompts into working websites, apps, and business dashboards for non-technical founders, raised a $130 million Series C at a $1.5 billion valuation, signaling that AI software-creation tools are moving past developers and into the hands of small business owners.

funding · no-code · ai-app-builder · startups · series-c

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

2026-07-18

Kaggle announced the four grand-prize winners of Google DeepMind's Measuring Progress Toward AGI hackathon, and all four winning benchmarks test uncertainty, self-knowledge, and in-context learning rather than broad AGI claims.

agi · benchmarks · deepmind · kaggle · evaluation · hallucination

Training AI to Think Shorter Makes Its Reasoning Harder to Trust

2026-07-18

A new study finds that reinforcement learning which rewards shorter chain-of-thought makes models cheaper to run but makes their written reasoning a less reliable guide to what actually decided the answer.

safety · interpretability · chain-of-thought · reasoning · monitorability

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

2026-07-18

A new paper shows a frozen small language model's accuracy on a hard math test can jump from 80.0% to 93.3% simply by grafting a byte-exact cache of verified solutions into it, with no weight changes at all.

kv-cache · distillation · efficiency · post-training · reasoning

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

2026-07-18

Boogu-Image-0.1 is a fully open-source unified image generation and editing model family whose researchers say a base model reaching near-frontier quality cost roughly $400,000 to train, arguing the closed-open gap is closing through data and pipeline quality rather than raw compute scale.

open-weights · image-generation · image-editing · multimodal · open-source

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

2026-07-18

Researchers scaled "zero RL" training to a trillion-parameter model, called Ring-Zero, and found the reasoning that emerges qualitatively changes at that size, reaching 84.2% on a hard math-competition exam without ever training on human chain-of-thought examples.

reinforcement-learning · reasoning · emergent-abilities · zero-rl · scaling

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

2026-07-18

A new systems paper called LongStraw shows reinforcement-learning post-training can execute on prompts beyond 2 million tokens on a fixed 8-GPU budget by scoring the shared prompt once without gradients and backpropagating only through the short generated response, though the authors call this proof of execution capacity rather than full training correctness, and the public code is not yet runnable.

rl-post-training · long-context · efficiency · grpo · systems

← 2026-07-172026-07-18later →