Ep 513: OpenAI’s new open model, Copilot updates, Perplexity going after Siri & more AI News That Matters
🎯 Summary
Podcast Episode Summary: Ep 513: OpenAI’s new open model, Copilot updates, Perplexity going after Siri & more AI News That Matters
This episode of the Everyday AI Show provides a rapid-fire weekly roundup of significant developments across the AI landscape, focusing on model accessibility, enterprise integration, legal challenges, and governmental policy shifts. The main narrative arc moves from updates on major foundational models (OpenAI, Google) to enterprise tool enhancements (Microsoft, Adobe) and concludes with critical cautionary tales and policy news.
1. Focus Area
The discussion is centered on Generative AI News and Applications, covering:
- Model Deployment & Accessibility: Updates on OpenAI’s lightweight research model and API releases.
- Enterprise AI Integration: Major feature rollouts for Microsoft 365 Copilot and Adobe Firefly’s multi-model strategy.
- Regulatory & Legal Issues: The DOJ’s antitrust case against Google and a high-profile case of AI hallucination misuse in legal filings.
- AI in Education Policy: A new US executive order focusing on expanding AI literacy in schools.
2. Key Technical Insights
- OpenAI’s Lightweight Research Model: OpenAI introduced a “lightweight deep research mode” powered by the GPT-4 Mini model, distinct from the full GPT-4, aimed at reducing resource load. Free users get 5 queries/month, while paid tiers receive higher caps before switching to the lightweight version.
- Adobe’s Multi-Model Platform: Adobe is integrating competitor models like OpenAI’s GPT Image Gen and Google’s Imagen 3 into Firefly, signaling a shift toward platform aggregation rather than purely proprietary model dominance in creative tools.
- GPT Image One API Launch: OpenAI released API access for its image generation model (powered by GPT Image One), allowing developers to build new applications with features like simultaneous image generation and C2PA watermarking for provenance tracking.
3. Business/Investment Angle
- Platform Aggregation as Strategy: Adobe’s move to incorporate rival models suggests that for creative suites, ecosystem integration and user workflow continuity (via shared credit systems) are becoming more valuable than strictly owning the underlying model IP.
- Copilot’s Enterprise Deepening: Microsoft’s Wave 2 Copilot updates, focusing on AI agents (research, skill discovery) and Copilot Notebooks, emphasize embedding AI deeply into existing enterprise workflows (M365, Jira, Miro) to drive tangible ROI.
- Search Dominance Under Threat: The DOJ’s potential forced sale of Chrome highlights the massive commercial value tied to default search engine placement, a battleground where Google currently leverages its $20B/year deal with Apple.
4. Notable Companies/People
- OpenAI: Released the lightweight GPT-4 Mini research mode and the GPT Image One API.
- Google: Facing a potential breakup of Chrome by the DOJ; Gemini AI chatbot has 350M monthly users.
- Microsoft: Launched significant updates to Microsoft 365 Copilot, including new AI agents and Copilot Notebooks.
- Adobe: Expanded Firefly to support competitor models (OpenAI, Google), causing a stock jump.
- Mike Lindell (MyPillow CEO): Highlighted as the subject of a major legal case involving attorneys submitting court filings filled with AI hallucinations.
- US Government (DOJ/White House): Involved in the antitrust case against Google and the issuance of an executive order on AI education.
5. Future Implications
The industry is moving toward specialized model tiers (lightweight vs. full-power) to manage costs and accessibility. Furthermore, the integration of competitor models into established platforms (Adobe) suggests a future where AI tool choice is abstracted away from the end-user, focusing instead on workflow integration. The recurring legal issues surrounding AI hallucinations underscore the urgent need for mandatory human-in-the-loop verification, especially in high-stakes fields like law. Finally, the US government is signaling a serious, albeit potentially under-resourced, commitment to AI literacy in K-12 education.
6. Target Audience
This episode is highly valuable for AI/ML Professionals, Technology Executives, Product Managers, and Business Leaders who need to stay current on strategic shifts, enterprise adoption trends, and regulatory risks associated with leading AI technologies.
🏢 Companies Mentioned
💬 Key Insights
"Yeah, so DeepSeek and Llama are not truly open source; people don't understand that. And it doesn't look like this model from OpenAI will be either, but it could be open weights. So, a big difference."
"OpenAI plans to use a highly permissive license with minimal restrictions on usage and commercialization, avoiding criticisms faced by models like Meta's Llama and Google's Gemma and DeepSeek."
"OpenAI could release its first 'open' language model since GPT-2, targeting an early summer launch."
"A year ago, Perplexity was on top in terms of being able to do this deep research... but they're not—they don't have a top-three product in that same category that they essentially helped create."
"the US education system, let me say this right now, is in huge trouble. It isn't huge trouble because we've essentially been ignoring artificial intelligence for years, banning it in classrooms across the US when we should have been doing exactly this all along: a national effort to push AI literacy K through 12."
"large language models rarely hallucinate if number one, you know what you're doing, and you have a knowledgeable human in the world who's using the right model at the right time for the right purposes, is going through a basic prompt engineering process."