Apple pivots AI strategy

Apple Pivots Focus to AI Hardware & Third-Party Apps

Apple knows it's losing the AI race—and it's not trying to win it anymore. According to Bloomberg's Mark Gurman, the company is shifting its strategy away from competing with ChatGPT and Gemini head-on, instead doubling down on what it does best: selling hardware and services.

What's coming: 

• At WWDC in June, Apple is expected to announce iOS 27 Extensions, which will let users install third-party AI chatbots beyond ChatGPT and run them directly inside Siri. Think of it as an AI App Store—a marketplace for third-party integrations where Apple takes its usual 30% cut. 

• A separate deal with Google will bring Gemini technology into Siri and other features, giving Apple usable in-house AI without building it from scratch.

Apple’s AI history: Apple got caught flat-footed by ChatGPT in 2022 and has been bleeding AI talent to OpenAI, Meta, Google, and Anthropic ever since. Apple has effectively conceded it won't be an AI leader anytime soon. Instead, it's positioning itself as the platform where AI runs—not the company that builds the best models.

Microsoft Unveils Multi-Model Deep Research System

On Monday, Microsoft announced Critique, a new deep research system that brings multiple AI models together to generate more reliable reports and answers. The multi-model approach should reduce hallucinations and boost productivity, a Microsoft VP told Reuters.

How it works: 

• Instead of relying on a single model, Critique uses one model to lead the generation phase—planning, retrieving information, and producing a draft—while a second model acts as an expert reviewer before the final output.

• Microsoft says benchmarks show this architecture "delivers best-in-class deep research" compared to similar offerings from OpenAI, Google, and Perplexity. 

What else launched: Microsoft also unveiled Council, which runs Anthropic and OpenAI models side-by-side on the same query, then uses a judge model to summarize where they agree, diverge, or offer unique insights. Both Critique and Council are available now in Microsoft's Frontier program.

Qodo Raises $70M for AI Code Verification

As AI coding tools generate line after line of code, a new challenge is emerging: making sure any of it actually works. Qodo, a startup building AI agents for code review, testing, and governance, just raised $70 million to tackle the problem.

What Qodo does: While most AI review tools focus on what changed in a commit, Qodo focuses on how changes affect entire systems—factoring in organizational standards, historical context, and risk tolerance. The company recently hit some impressive benchmarks. It ranked No. 1 on Martian's Code Review Bench, scoring over 10 points ahead of the next competitor and 25 points ahead of Claude Code Review.

Who's using it: Qodo already works with Nvidia, Walmart, Red Hat, Intuit, and Texas Instruments, plus high-growth firms like Monday.com and JFrog. Qumra Capita led the funding round, and OpenAI's Peter Welinder and Meta's Clara Shih also participated.

Want AI Confidence in Your Big Business Calls?

Would you stake an enterprise business decision on what your AI just told you?

Most teams can't because AI outputs don't carry the judgment of the experts who actually know the domain. Dialect from Scale AI changes that.

• Your experts encode corrections, context, and reasoning into every AI workflow

• That knowledge persists and compounds, building trust with every interaction

• You capture your own enterprise IP, instead of leaking it to the model builders

Open-source SuperAgent for workflows

DeerFlow uses large language models to break down complex prompts into parallel sub-agents and synthesize outputs like research reports, slide decks, and full codebases. It supports multiple model backends and persistent memory for long-running tasks.

How you can use it

• Automate multi-step research and reporting workflows end-to-end

• Run sandboxed code and tool chains with local control

• Extend functionality with Markdown-defined skills and tools

• Self-host for privacy and full customization

Pricing: Open source

An AI secretary for your inbox

Stamp automates email and calendar management by drafting replies in your voice, prioritizing messages, generating summaries, and executing tasks like bulk drafting and scheduling. It learns from your past emails to personalize responses and surface what matters.

How you can use it

• Approve AI-drafted replies with a single keystroke

• Auto-prioritize and categorize incoming messages

• Generate to-dos and summaries from your inbox

• Offload repetitive triage so nothing slips through

Pricing: Paid, free trial available