Tuesday, April 28, 2026

TL;DR

Today’s briefing covers 687 items from 13 sources. Top stories include: OpenAI available at FedRAMP Moderate; Announcing our partnership with the Republic of Korea; Open source package with 1 million monthly downloads stole user credentials.

Top Stories

OpenAI available at FedRAMP Moderate

OpenAI

OpenAI is available at FedRAMP Moderate authorization for ChatGPT Enterprise and the OpenAI API, enabling secure AI adoption for U.S. federal agencies.

Announcing our partnership with the Republic of Korea

DeepMind

Google DeepMind and Korea partner to accelerate scientific breakthroughs using frontier AI models

Open source package with 1 million monthly downloads stole user credentials

Ars Technica AI

If you’re one of millions using element-data, it’s time to check for compromise.

DeepMind’s David Silver just raised $1.1B to build an AI that learns without human data

TechCrunch AI

Ineffable Intelligence, a British AI lab founded a mere few months ago by former DeepMind researcher David Silver, has raised $1.1 billion in funding at a valuation of $5.1 billion.

Microsoft and OpenAI’s famed AGI agreement is dead

The Verge AI

OpenAI and Microsoft’s partnership-turned-situationship just got even less committed. And a clause about artificial general intelligence, which has for years dictated the future of their deal, has officially been dropped. On Monday morning, Microsoft announced a handful of big changes to its long-st

Research Corner

Paper Source
An Intelligent Fault Diagnosis Method for General Aviation Aircraft Based on … arXiv AI
PExA: Parallel Exploration Agent for Complex Text-to-SQL arXiv AI
The Power of Power Law: Asymmetry Enables Compositional Reasoning arXiv AI
On the Existence of an Inverse Solution for Preference-Based Reductions in Ar… arXiv AI
Towards Causally Interpretable Wi-Fi CSI-Based Human Activity Recognition wit… arXiv AI
FormalScience: Scalable Human-in-the-Loop Autoformalisation of Science with A… arXiv AI
A Systematic Approach for Large Language Models Debugging arXiv AI
A Decoupled Human-in-the-Loop System for Controlled Autonomy in Agentic Workf… arXiv AI
Don’t Make the LLM Read the Graph: Make the Graph Think arXiv AI
Analytica: Soft Propositional Reasoning for Robust and Scalable LLM-Driven An… arXiv AI

Industry Updates

News & Analysis

Deep Dive: Today’s AI Landscape

Synthesized from multiple research sources

Key AI announcements on April 28, 2026, include IBM’s launch of Bob, an AI development partner for enterprise software lifecycles, and recaps of recent breakthroughs like Anthropic’s Claude Mythos 5 and Google DeepMind’s Gemini 3.1 models. [1][2]

Major Model Releases and Industry Shifts (Last 24 Hours)

  • Anthropic released Claude Mythos 5, a 10-trillion parameter model, amid reports of SpaceX acquiring xAI for $250 billion.[1]
  • Google DeepMind launched Gemini 3.1 suite, with the Ultra variant scoring 94.3% on GPQA Diamond benchmark; Gemini 3.1 Flash-Lite offers 2.5x faster responses and 45% better output speed.[1]

Sources:

In April 2026, AI startups dominated venture funding with 1,314 total deals across all sectors, 764 (58%) involving AI or machine learning companies, reflecting AI’s role as the primary focus of VC activity.[1]

Major trends include accelerated seed-stage funding for AI (98 of 137 seed deals, or 72%), with generative AI seed rounds averaging $2.1M and Series A rounds averaging $18.5M—3.5x higher than non-AI peers ($5.2M seed, $12.1M Series A).[1] Infrastructure for AI (e.g., software/hardware enabling models) captured 145 deals, alongside 280 in general-purpose LLMs and generative tools, showing investor bets on both applications and enablers.[1]

Top AI Startups by Funding and Valuation (as of April 2026)

Sources:


This newsletter is automatically generated by PAI using RSS aggregation and AI research tools. Sources include arXiv, HuggingFace, OpenAI, Google AI, MIT News, VentureBeat, and more.

Have feedback? Reply to this post or reach out on X/Twitter.