Shiny Weekly: March 2-9, 2025

Hey folks, quick note from me. First the new site is live here and this email was researched, written and compiled by a new tool called Manus as a test. Its insane. Full breakdown and tutorial coming soon.

So here we go.

The AI world's been on fire this week with one major trend stealing the spotlight: AI agents are taking over! We're way beyond simple chatbots now - these new AI systems can actually do stuff on their own.

This week, Meta and Salesforce rolled out cool new AI agents for businesses, while OpenAI's planning to charge a whopping $20K monthly for their high-end research agents. GitHub's made their GPT-4o Copilot smarter at writing code, and Google's pushing boundaries with new Gemini features.

For tech fans like us, this isn't just another incremental update - it's a game-changer for how we'll work with technology. We're shifting from "AI assistants" to "AI agents" that can handle complex tasks independently.

In this first issue, we'll break down what's happening, why it matters, and what's coming next. Whether you're coding, running a business, or just love cutting-edge tech, we've got you covered.

Let's dive in!

Meta and Salesforce Introduce AI Agents for Small Businesses

GitHub Upgrades GPT-4o Copilot with Enhanced Code Completion

Good news for developers! GitHub just upgraded its GPT-4o Copilot preview model, making it way better at suggesting code. This is a big deal for anyone who codes.

The new version builds on last month's release but uses reinforcement learning to get smarter about code suggestions. Basically, it learns from how developers interact with it and keeps getting better over time.

You can try it now under the same "gpt-4o-copilot" name, with full release expected by mid-April 2025. It works with popular coding environments like VS Code and JetBrains IDEs.

If you're using Copilot Business or Enterprise, your admin will need to enable this in the policy settings. They've also rolled out personal instructions for Copilot Chat, letting you tell it your preferred languages, coding style, and standards.

Tools like this are changing how coding works - less time on routine tasks means more time for creative problem-solving.

Source: GitHub Blog

OpenAI's Specialized AI Agents Could Cost $20,000 Per Month

OpenAI is jumping into the high-end AI market with a jaw-dropping price tag. According to The Information, they're planning to launch specialized "agent" products with one premium option costing a whopping $20,000 per month!

This top-tier agent will target "PhD-level research" and goes way beyond what consumer AI can do. The eye-popping price reflects both the advanced capabilities and OpenAI's financial situation - they reportedly lost about $5 billion last year running their services.

Meanwhile, there's drama in the background. A judge recently denied Elon Musk's attempt to stop OpenAI's transition to a for-profit company, though the court might still hold a trial on whether this move violates OpenAI's original mission. Musk claims Sam Altman and the team abandoned their nonprofit roots.

In related news, Scale AI (another big AI player) is under investigation by the Labor Department for potential issues with worker classification and compensation.

These stories highlight the growing pains AI companies face as they try to balance innovation with sustainable business models. Will enterprises actually pay these premium prices? That's the real test of how valuable these advanced AI capabilities really are.

Google Advances AI Capabilities with New Gemini Features

OpenAI is jumping into the high-end AI market with a jaw-dropping price tag. According to The Information, they're planning to launch specialized "agent" products with one premium option costing a whopping $20,000 per month!

This top-tier agent will target "PhD-level research" and goes way beyond what consumer AI can do. The eye-popping price reflects both the advanced capabilities and OpenAI's financial situation - they reportedly lost about $5 billion last year running their services.

Meanwhile, there's drama in the background. A judge recently denied Elon Musk's attempt to stop OpenAI's transition to a for-profit company, though the court might still hold a trial on whether this move violates OpenAI's original mission. Musk claims Sam Altman and the team abandoned their nonprofit roots.

In related news, Scale AI (another big AI player) is under investigation by the Labor Department for potential issues with worker classification and compensation.

These stories highlight the growing pains AI companies face as they try to balance innovation with sustainable business models. Will enterprises actually pay these premium prices? That's the real test of how valuable these advanced AI capabilities really are.

AI Models Face Off in Super Mario Bros. Benchmark

Here's something fun - researchers at UCSD's Hao AI Lab decided to test AI models by making them play Super Mario Bros! It's a clever way to see how different AIs handle complex environments that need quick thinking and adaptation.

Here's something fun - researchers at UCSD's Hao AI Lab decided to test AI models by making them play Super Mario Bros! It's a clever way to see how different AIs handle complex environments that need quick thinking and adaptation.

The results? Anthropic's Claude 3.7 crushed it in the gaming tests, while OpenAI's GPT-4o struggled with the same challenges. This shows some interesting differences in how these AI systems handle real-time decision making.

Games like Mario are actually great testing grounds for AI because they combine visual understanding, planning, and quick reactions - much like real-world situations. To succeed, the AI needs to develop strategies, adapt to surprises, and keep focused on goals over time.

This research shows why we need diverse ways to test AI beyond just generating text and images. If we want AI agents that can operate in complex environments, we need to evaluate them in interactive challenges that better reflect the real world.

Source: TechCrunch

Wrapping Up

Looking at all the AI news this week, one thing's clear: we're moving from AI that just answers questions to AI that actually does stuff for you. From Meta and Salesforce's business agents to GitHub's smarter code helper and OpenAI's high-end research assistants, everyone's betting on AI agents as the next big thing.

This shift brings some amazing opportunities and challenges. Businesses can boost productivity and engage customers in new ways. Developers can code faster and focus on the creative parts of their work. And for everyone watching this space, there are big questions about pricing, jobs, and regulations as these technologies move from experiments to everyday tools.

In future issues, we'll keep tracking not just what these systems can do technically, but what they mean for how we work and create. The move to AI agents is one of the biggest tech shifts happening right now, and it's going to change how we interact with technology in pretty much every area of life.

We'd love to hear what you think about this newsletter and what topics you want us to cover next time.

Have a good one.

Thumbnail Credit - vittorio on X.