Copilot goes cheap as Microsoft phases out OpenAI and Anthropic models to cut costs
Microsoft is swapping OpenAI and Anthropic models for its own MAI models in Copilot to reduce costs, potentially impacting performance for users.
- Microsoft is replacing OpenAI and Anthropic models with its own MAI models in Copilot to cut costs.
- Tens of thousands of weekly queries are already running through the new in-house models.
- The shift may reduce performance for users despite lower costs.
- AI chief Mustafa Suleyman aims to eliminate external model expenses entirely.
Microsoft is transitioning its Copilot AI assistant from models developed by OpenAI and Anthropic to its own in-house MAI models. The shift aims to significantly reduce costs, with tens of thousands of weekly queries already running through the new models. AI chief Mustafa Suleyman has stated the goal is to "ultimately eliminate" the expense of using external models entirely.
The move is part of a broader strategy to make Copilot more affordable for customers, though it may come at the cost of performance. Users could experience differences in response quality or capabilities compared to the previous models. The transition is already underway in products like Excel and Outlook, signaling a long-term shift in Microsoft's AI infrastructure strategy.
Source: Copilot goes cheap as Microsoft phases out OpenAI and Anthropic models to cut costs. Read the full piece at the source.
Developers using Copilot may need to adapt to changes in model behavior and performance.
Businesses relying on Copilot could see cost savings but potential trade-offs in AI response quality.
Investors should monitor Microsoft's AI cost-cutting strategy and its impact on Copilot's market position.
Users may experience changes in Copilot's performance as Microsoft shifts to in-house AI models.
- MAI models
- Microsoft's in-house AI models designed to replace third-party models like those from OpenAI and Anthropic.
Apple closes in on Nvidia as investors treat less AI capex spending as an advantage - CNBC
Government AI can’t scale — and it’s not the models - Federal News Network
Palantir CEO Alex Karp is paranoid about Anthropic and OpenAI. That doesn't mean he has nothing to fear. - Fortune
Anthropic Is Closing the Doors To Chinese Engineers - Yahoo Finance
Elon Musk-Sam Altman judge sends OpenAI battle to mediator - The Mercury News
sqlite-utils 4.0, now with database schema migrations
SQLite-Utils 4.0 introduces database schema migrations and nested transactions, marking its first major update since 2020.
Discord admits AI moderation bug wrongfully banned users over harmless images
Discord’s AI moderation system wrongly banned over 8,000 users in two months due to a bug that misclassified harmless images like spreadsheets and chessboards as harmful content.
Meta is launching its first in-house AI image generator for consumers and advertisers - qz.com
Meta is rolling out its first in-house AI image generator, targeting both consumers and advertisers.
Meta AI app adds image generation, also powering WhatsApp and Instagram features - 9to5Mac
Meta’s AI app now generates images and powers visual features in WhatsApp and Instagram, expanding its multimodal capabilities.
AI ToolsBuilding a physical API
A developer built a Rust-based system that converts raw CAN bus traffic from a robot into structured telemetry via a physical API, enabling real-time monitoring and control.
AI ToolsI connected ChatGPT to my bank, and it's my go-to finance app now - here's how (and why)
OpenAI’s ChatGPT Finances integrates banking data to let users view balances, investments, and debts in one place and query AI for financial insights.