Research - Maginative

Research

Thinking Machines Claims 30x Cost Cut for Training AI Models

Chris McKay• October 28, 2025 • 2 min read

Thinking Machines published a detailed recipe for “on-policy distillation,” showing how per-token grading from a teacher model can push small students to strong math and assistant performance at a fraction of RL’s cost, with reproducible code via its Tinker SDK.

OpenAI Research

AI Hallucinations Are a Test-Taking Problem, Says OpenAI

Chris McKay• September 5, 2025 • 3 min read

OpenAI researchers argue that language models hallucinate because current training and evaluation methods statistically reward guessing over expressing uncertainty.

Microsoft Healthcare Research

Microsoft's MAI-DxO Crushes Doctors at Medical Diagnosis while Cutting Costs

Chris McKay• June 30, 2025 • 3 min read

Microsoft researchers developed an AI system that achieved 85.5% accuracy on challenging medical cases versus 20% for human physicians, while reducing estimated diagnostic costs in a controlled study.

Google Biotech & Health Research

DeepMind Launches AlphaGenome to Better Understand Gene Regulation

Chris McKay• June 26, 2025 • 3 min read

DeepMind’s AlphaGenome AI model decodes how mutations affect non-coding DNA, potentially transforming our understanding of disease.

Startups Research

Databricks Co-founder Launches Laude Institute, Stakes His Own $100 Million

Chris McKay• June 24, 2025 • 2 min read

Databricks and Perplexity co-founder Andy Konwinski is putting $100 million of personal funds behind Laude Institute, a grant-style venture that’s kicking off with a five-year, $15 million commitment to an AI Systems Lab at UC Berkeley.

Google Research

Google's DeepMind Unveils AlphaEvolve, an AI System that Designs and Optimizes Algorithms

Chris McKay• May 14, 2025 • 3 min read

Google DeepMind’s AlphaEvolve is an AI agent that evolves algorithms with LLMs and automated evaluation, already improving infrastructure, math problems, and its own AI training systems.

OpenAI Research Education AI Literacy

AI in Higher Ed: 7 Major Takeaways from OpenAI’s ChatGPT Student Usage Report

Chris McKay• February 20, 2025 • 3 min read

More than one-third of college-aged students in the US use ChatGPT, with a significant portion relying on it for learning. OpenAI’s new report reveals how this impacts workforce readiness, economic competitiveness, and the role of AI in higher education.

Microsoft Media & Entertainment Research

Microsoft’s Muse AI Model Can Generate Video Game Environments in Real-Time

Chris McKay• February 19, 2025 • 2 min read

Microsoft has unveiled Muse, a generative AI model designed to create video game environments that respond to player actions, potentially transforming game development and preservation.

Google Research

Google Unveils AI Co-Scientist to Accelerate Scientific Discovery

Chris McKay• February 19, 2025 • 2 min read

Google has introduced an AI co-scientist built on Gemini 2.0, designed to assist researchers in generating novel hypotheses, synthesizing literature, and formulating experimental plans, with early access available through the Trusted Tester Program.

NVIDIA Research Biotech & Health

NVIDIA and Arc Institute Introduce Evo 2, A State of the Art Foundation Model for Biology

Chris McKay• February 19, 2025 • 2 min read

The Arc Institute and NVIDIA unveil Evo 2, the largest AI model for biology ever, trained on 9.3 trillion DNA base pairs from over 128,000 species to advance synthetic biology, medicine, and genome design.

OpenAI Research

OpenAI's New Benchmark Tests AI Models Against Real-World Software Engineering Tasks

Chris McKay• February 18, 2025 • 2 min read

OpenAI has launched SWE-Lancer, a benchmark evaluating AI models on over 1,400 real-world software engineering tasks sourced from Upwork, collectively worth $1 million in payouts.

Adobe Research China

Adobe's New AI, TransPixar, Adds Transparency to Generated Videos

Chris McKay• January 9, 2025 • 2 min read

TransPixar generates videos with RGBA channels, integrating transparency for effects like smoke, fire, and glass reflections.

Apple Research

Apple's Ferret-UI is an AI that Can Understand and Navigate Mobile Interfaces

Chris McKay• September 16, 2024 • 3 min read

Ferret-UI can discuss specific parts of the interface, infer the overall function of a screen, and even reason about how a user might interact with it.

Robotics Research

Living Skin Makes Robots More Human-Like

Chris McKay• July 1, 2024 • 2 min read

To show how versatile their technique is, University of Tokyo researchers made two prototypes: a 3D facial mold fully covered in living skin, and a simple robotic face that can smile.

Research

PhysDreamer Brings Realistic, Interactive Dynamics to AI-Generated Video

Chris McKay• April 22, 2024 • 1 min read

The key idea behind PhysDreamer is to generate a plausible video of the object in motion and then optimize the material properties to match this synthesized motion.

Microsoft Research

Microsoft's VASA can Create a Realistic Talking Head Video from a Single Photo

Chris McKay• April 18, 2024 • 1 min read

VASA demonstrates superior performance compared to existing methods, delivering high-quality video frames with precise lip-audio synchronization and a diverse array of lifelike facial dynamics.

Apple Research

Apple Introduces MobileCLIP, a State-of-the-Art Image-Text Model for Mobile Devices

Chris McKay• April 4, 2024 • 3 min read

Apple's multi-modal reinforced training demonstrates 10-1000x improved learning efficiency compared to standard CLIP training.

Anthropic AI Safety Research

Anthropic Shares Research on Technique to Exploit Long Context Windows to Jailbreak Large Language Models

Chris McKay• April 2, 2024 • 3 min read

Many-shot jailbreaking works by prompting the model with a large number of fictitious question-answer pairs that depict the AI assistant providing harmful or dangerous responses.

Google Research

Google DeepMind's New Research Shows LLMs Can Outperform Humans in Fact-Checking

Chris McKay• March 29, 2024 • 2 min read

The research introduces a novel method for evaluating long-form factuality in LLMs, demonstrating that bigger models are more factual and that LLMs are 20 times cheaper than human annotators.

NVIDIA Research

NVIDIA's LATTE3D Can Generate 3D from Text Prompts in Seconds

Chris McKay• March 21, 2024 • 1 min read

Like a virtual 3D printer, LATTE3D turns text prompts into 3D representations of objects and animals within a second.

Google Research

Meet SIMA: Google's New AI that Can Play Video Games with You

Chris McKay• March 13, 2024 • 3 min read

The research is building towards more general AI systems and agents that can understand and safely carry out a wide range of tasks in a way that is helpful to people online and in the real world.

Research

Tencent Proposes Technique to use LLMs to Make Diffusion Models More Accurate

Chris McKay• March 11, 2024 • 2 min read

By enabling diffusion models to more faithfully visualize the content of natural language prompts, this approach opens up exciting possibilities for creative applications and beyond.

$Orca-Math Shows the Potential of Specialized Small Language Models$

Microsoft Research

Orca-Math Shows the Potential of Specialized Small Language Models

Chris McKay• March 5, 2024 • 3 min read

The research shows the value of smaller, specialized models in specific domains, where they can match or even surpass the performance of much larger models.

Research Alibaba

Meet EMO, Alibaba's New AI That Can Make Any Photo Talk or Sing

Chris McKay• February 28, 2024 • 2 min read

By enabling the generation of highly expressive and lifelike videos from a single reference image and audio input, EMO opens new avenues in entertainment, telepresence, and beyond.