Today in AI

In partnership with

• News roundup

• Today’s big story

• Research spotlight

• This week on the podcast

News roundup

The top stories in AI today.

NEW LAUNCHES

The latest features, products & partnerships in AI

• Fast.ai and Answer.AI merge, unveil AI code education platform

• CapCut’s ‘Commerce Pro’ streamlines content creation for e-commerce

• Swarovski unveils the world’s first AI-enabled binoculars

• Google’s new ‘Vids’ tool will create video presentations for you

• Microsoft just updated a 38-year-old software with AI, and the results are amazing

• Bagel Explores: training-time and inference-time techniques to advance AI reasoning (Sponsored)

GOVERNMENT

Press releases, regulation, defense & politics.

• Canada invests $2.4B to accelerate AI adoption nationwide

• ChatGPT blocked more than 250,000 political deepfakes, OpenAI reports

• AI startups anticipate looser regulations under Trump

• What Trump’s return as President means for tech

• OpenAI has won a legal battle against publishers, but the war will continue

• Anthropic, the AI safety poster child, is going into the defense industry

AI MODELS

Deployment, research, training & infrastructure

• Gary Marcus: AI models are reaching a performance plateau

• OpenCoder is a new code-focused LLM that is truly open

• FrontierMath: How to determine advanced math capabilities in LLMs

• How new AI models are compressing videos without reducing quality

• Google may accelerate its Gemini 2 AI model release timeline

• New OpenAI model accelerates media generation by 50X

• How Roboflow saved 74 years of developer time with Meta’s SAM model

• AI expert Bruce Schneier on why society needs ‘public AI models’

IMPLEMENTATION

Announcements, strategies & programs

• 3 key LLM compression tactics to boost AI performance

• AI SDRs fail for 83% of companies, but change is coming

• The top use cases for AI marketing

• Essential considerations when choosing an AI model for your business

• Companies are using AI to screen job candidates — here are the AI tools that might outsmart them

• ‘Multimodal RAG’ is all the rage — here’s what it is and how to get started

What’s happening in AI right now

The content wars enter a new phase

OpenAI's recent court victory, dismissing a copyright lawsuit from publishers Raw Story and AlterNet reveals an uncomfortable truth: we're using industrial-era legal frameworks to regulate information-age innovations. While publishers argue about content ownership, a big question is how AI is upending traditional notions of value creation.

The case at hand

The lawsuit centered on claims that OpenAI violated digital copyright protections by using publishers' content to train its models. Judge Colleen McMahon ruled that the publishers lacked legal standing to bring the claim, citing insufficient proof that ChatGPT was trained on their material or that such training caused harm.

Media economics breaking down?

The fundamental rules and frameworks that governed content businesses are unraveling. In the traditional model, value came from scarcity - creating and distributing unique content that others couldn't easily replicate. The marginal cost of producing quality content was high, and distribution channels were limited.

AI inverts these economics. Value emerges not from individual pieces of content but from the patterns discovered across vast datasets. The marginal cost of "understanding" new content approaches zero as once an AI model is trained, it can process and analyze new content at virtually no additional cost.

This explains why conventional damage claims struggle in court. How do you calculate harm when the value derived isn't from copying specific content but from discovering patterns across billions of data points? Traditional copyright law never contemplated this type of use.

Network effects at scale 

The dynamics become even more interesting at scale. While traditional content businesses face diminishing returns as they grow, AI systems often exhibit opposite characteristics. More data generally leads to better performance which attracts more users, generating more data in a powerful flywheel effect.

Beyond simple patterns 

What makes this transition particularly challenging is that we're not simply seeing a shift from one clear business model to another. Instead, we're entering a period where multiple new models may emerge and compete. Some organizations will focus on becoming AI infrastructure providers, others on specialized applications, and still others on curating and validating training data.

The most important question isn't who owns what content - it's who can best capture value in a world where understanding patterns across all content matters more than owning any specific piece.

We publish daily research, playbooks, and deep industry data breakdowns. Learn More Here

The Bagel team just published new research on two complementary methods for advancing AI reasoning: training-time and inference-time techniques.

Training-Time Enhancements involve refining model structures. Methods like Parameters Efficient Fine-Tuning (PEFT) optimize learning efficiency by targeting specific neural pathways within fixed frameworks, while approaches like WizardMath’s 3-step reasoning improve structured reasoning.

Inference-Time Enhancements focus on on-the-fly thinking, using Chain of Thought prompting to activate logical processing without extra training. Techniques like Self-Consistency for validation and Program of Thought for coding tasks enable precise, stepwise reasoning.

The findings suggest combining both approaches: training-time builds foundational reasoning, while inference-time optimizes its application. Bagel Network supports this advancement through open-source infrastructure, empowering a community-driven AI evolution though monetizable open source AI.

About Bagel: Bagel is an AI & cryptography research lab, building the world's first monetization layer for open-source AI.

This week on the podcast

Can’t get enough of our newsletter? Check out our podcast Future-Proof.

In this episode, the hosts Anthony Batt and Shane Robinson with guest Joe Veroneau from Conveyor discuss outsmarting paperwork. Conveyor is a company that helps automate security reviews and document sharing between companies. They use AI technology, specifically language models, to automate the process of filling out security questionnaires. This saves customers a significant amount of time and improves the quality of their responses.