Category intelligence

AI News Briefing — April 13, 2026

8 current items analyzed and ranked.

Executive synthesis

AI News Summary

Anthropic dominated headlines by withholding its frontier model Mythos on cybersecurity grounds, prompting a response from the US Treasury Secretary and UK lawmakers—though skeptics call it an investor-courting publicity move.

MiniMax open-sourced M2.7, a self-evolving MoE agent model posting strong scores on SWE-Pro (56.22%) and Terminal Bench 2 (57.0%)
Meta AI and KAUST proposed Neural Computers, a novel paradigm where neural networks replace traditional computing stacks
Liquid AI released LFM2.5-VL-450M, a compact vision-language model achieving sub-250ms edge inference with bounding box prediction

On the policy front, OpenAI launched a multi-pronged influence campaign including a policy paper, podcast acquisition, and DC office as public disapproval rises. Meanwhile, bipartisan US datacenter protests threaten to slow the White House's AI infrastructure plans.

Key Themes

AI Safety & Frontier Model Governance · 1Open Source Model Releases · 2AI Policy, Lobbying & Public Backlash · 3Novel AI Research & Architectures · 2

Primary evidence

Top Ranked Signals

News AI (artificial intelligence) | The Guardian Apr 12

‘Too powerful for the public’: Inside Anthropic’s bid to win the AI publicity war

By Aisha Down

82 score

AI Analysis

Continuing our coverage of Claude Mythos, Anthropic announced it created a frontier AI model called Mythos that it deemed too powerful to release publicly, citing catastrophic cybersecurity risks. The US Treasury Secretary summoned major bank heads over the model, while skeptics suggest the withholding is a publicity stunt to attract investors.

The firm says it withheld an AI model on cybersecurity grounds but sceptics say this was hype to lure investmentThis week, the AI company Anthropic said it had created an AI model so powerful that, out of a sense of overwhelming responsibility, it was not going to release it to the public.The US treasury secretary, Scott Bessent, summoned the heads of major banks for a chat about the model, Mythos. The Reform UK MP Danny Kruger wrote a letter to the government urging it to “engage with AI firm A

AI SafetyFrontier ModelsAnthropicAI HypeCybersecurity

News MarkTechPost Apr 12

MiniMax Just Open Sourced MiniMax M2.7: A Self-Evolving Agent Model that Scores 56.22% on SWE-Pro and 57.0% on Terminal Bench 2

By Asif Razzaq

75 score

AI Analysis

First spotted on Reddit yesterday, now getting detailed coverage, MiniMax open-sourced M2.7, a Mixture-of-Experts agent model scoring 56.22% on SWE-Pro and 57.0% on Terminal Bench 2. Notably, it is MiniMax's first model to actively participate in its own development cycle, marking a shift toward self-evolving AI systems.

MiniMax has officially open-sourced MiniMax M2.7, making the model weights publicly available on Hugging Face. Originally announced on March 18, 2026, MiniMax M2.7 is the MiniMax’s most capable open-source model to date — and its first model to actively participate in its own development cycle, a meaningful shift in how large language models are built and iterated. What is MiniMax M2.7? MiniMax M2.7 is part of MiniMax’s M2-series of Mixture-of-Experts (MoE) models. MoE is an ar

Open SourceAgentic AIModel ReleaseSoftware EngineeringMoE Architecture

News MarkTechPost Apr 12

Meta AI and KAUST Researchers Propose Neural Computers That Fold Computation, Memory, and I/O Into One Learned Model

By Michal Sutter

68 score

AI Analysis

As first covered in Research on April 10 and Social yesterday, Meta AI and KAUST researchers proposed Neural Computers (NCs), a framework where a neural network itself acts as the entire computer rather than a layer on top of one. They demonstrated two working video-based prototypes handling CLI and GUI runtime primitives.

Researchers from Meta AI and the King Abdullah University of Science and Technology (KAUST) have introduced Neural Computers (NCs) — a proposed machine form in which a neural network itself acts as the running computer, rather than as a layer sitting on top of one. The research team presents both a theoretical framework and two working video-based prototypes that demonstrate early runtime primitives in command-line interface (CLI) and graphical user interface (GUI) settings. arxiv.org/

AI ResearchMeta AINovel ArchitecturesComputing Paradigms

News MarkTechPost Apr 12

Liquid AI Releases LFM2.5-VL-450M: a 450M-Parameter Vision-Language Model with Bounding Box Prediction, Multilingual Support, and Sub-250ms Edge Inference

By Asif Razzaq

62 score

AI Analysis

Liquid AI released LFM2.5-VL-450M, a 450M-parameter vision-language model designed for edge devices with sub-250ms inference. The update adds bounding box prediction, multilingual support, and function calling while running on hardware from NVIDIA Jetson to smartphone SoCs.

Liquid AI just released LFM2.5-VL-450M, an updated version of its earlier LFM2-VL-450M vision-language model. The new release introduces bounding box prediction, improved instruction following, expanded multilingual understanding, and function calling support — all within a 450M-parameter footprint designed to run directly on edge hardware ranging from embedded AI modules like NVIDIA Jetson Orin, to mini-PC APUs like AMD Ryzen AI Max+ 395, to flagship phone SoCs like the Snapdragon 8 Elite insid

Edge AIVision-Language ModelsModel ReleaseOpen SourceSmall Models

News AI (artificial intelligence) | The Guardian Apr 12

AI companies know they have an image problem. Will funding policy papers and thinktanks dig them out?

By Nick Robins-Early

60 score

AI Analysis

OpenAI published a policy paper calling for reimagining the social contract, acquired the tech podcast TBPN, and announced plans for a Washington DC office with a dedicated nonprofit workshop. This is part of a broader industry effort to reshape public perception as polls show growing disapproval of AI.

The aggressive effort by major players aims to reshape the narrative as polls show increasing public disapproval of AIOpenAI made a surprise announcement this week – not an update to ChatGPT or another multibillion-dollar datacenter – but a policy paper that called for a reimagining of the social contract based around “a slate of people-first ideas”. It’s the latest move in an aggressive effort by the major AI players to reshape the narrative around their industry, as polls show public disapprov

AI PolicyOpenAIPublic PerceptionLobbyingAI Governance

News AI (artificial intelligence) | The Guardian Apr 12

The Guardian view on AI politics: US datacentre protests are a warning to big tech | Editorial

By Editorial

58 score

AI Analysis

Building on the datacenter moratorium bill noted in Research two days ago, Bipartisan opposition to AI datacenter construction is growing across the US, with Texas Republicans and liberal Californians joining forces against unregulated expansion. The protests pose a direct challenge to the White House's AI infrastructure rollout plan.

In both Republican and Democratic states, scepticism and hostility towards an unregulated construction boom is growingWhen blue-collar Trump voters and Maga-friendly midwest states join the same cause as Bernie Sanders and liberal California teachers, something novel is afoot. Last month it was the turn of the Republican party in Texas to express forthright opposition to the construction of datacentres for artificial intelligence, pending adequate environmental safeguards for local communities.

AI InfrastructurePoliticsEnvironmentPublic OppositionDatacenter

News MarkTechPost Apr 12

A Coding Implementation of MolmoAct for Depth-Aware Spatial Reasoning, Visual Trajectory Tracing, and Robotic Action Prediction

By Michal Sutter

35 score

AI Analysis

A step-by-step coding tutorial demonstrating MolmoAct for depth-aware spatial reasoning, visual trajectory tracing, and robotic action prediction from natural language instructions. The tutorial covers model setup, multi-view image input, and inference pipelines for robotics tasks.

In this tutorial, we walk through MolmoAct step by step and build a practical understanding of how action-reasoning models can reason in space from visual observations. We set up the environment, load the model, prepare multi-view image inputs, and explore how MolmoAct produces depth-aware reasoning, visual traces, and actionable robot outputs from natural language instructions. As we move through the workflow, we run inference and also examine how the model parses actions, visualizes trajectori

RoboticsTutorialsComputer VisionSpatial Reasoning

News AI (artificial intelligence) | The Guardian Apr 12

Is AI the greatest art heist in history?

By Molly Crabapple

28 score

AI Analysis

A Guardian opinion piece argues generative AI constitutes a large-scale plundering of the art world, noting that artists were early to predict the negative consequences now widely evident in 2026. The piece frames AI's use of training data as a form of theft.

New technologies of reproduction are plundering the art world – and getting away with itIn 2026, its easy to see why generative AI is bad. The internet has nicknamed its excretions “slop”. The CEOs of AI companies prance about on stage like supervillains, bragging that their products will eliminate vast swathes of work. Generative AI requires sacrificing the world’s water to feed its hideous data centres. Around the globe, chatbots induce schizophrenic delusions and urge teens to kill themselve