Google has announced a major pivot for its generative AI platform, set to retire Gemini’s “Gems” in favor of a new system called “Skills.” This shift arrives as tech giants and AI developers race to define what customization and extensibility mean for large language models (LLMs) in production. The Gemini update signals more than a brand tweak — it marks a deeper attempt to standardize third-party plugin ecosystems, closing the gap with OpenAI’s GPTs and prompting urgent strategic questions for developers, enterprises, and startups building on top of generative AI.
- Google replaces Gemini’s “Gems” with “Skills,” shifting focus to developer-built abilities and standardized integrations
- The move aims to catch up to OpenAI’s wildly successful GPTs, reshaping the generative AI competition
- Developers and AI teams face new constraints, migration tasks, and emerging monetization opportunities as the ecosystem resets
- Broader implications reach from user experience to AI assistant interoperability and developer toolchains
Key Takeaways
Google’s retirement of Gemini’s Gems in favor of Skills is not a cosmetic change — it ushers in a different paradigm for AI extensibility, shifting from lightweight, user-defined behaviors to a robust framework for developer-created capabilities. As Skills attract tighter controls and likely tap into broader APIs, Google positions itself to rival OpenAI’s GPT Store strategy and redefine commercial opportunities for the AI developer ecosystem.
“Google’s pivot from Gems to Skills signals that the generative AI platform wars are entering a new phase—where extensibility, developer monetization, and ecosystem power matter more than mere model size.”
A Closer Look at Google’s New “Skills” Ecosystem
Google’s Gemini originally introduced “Gems” as custom, user-provided persona tweaks that let individuals personalize chatbot behaviors. In practice, Gems functioned similarly to GPTs: users defined a persona or utility (“Personal Trainer,” “Travel Planner”), and Gemini would respond accordingly. Adoption, however, lagged behind OpenAI’s GPTs, which gained traction on the back of an active, growing third-party marketplace.
With the debut of “Skills,” Google refocuses on purpose-built extensions that developers can deploy for specialized tasks—think authentication, integrations with enterprise SaaS, or real-time data lookups. By ditching the casual, user-facing persona model, the company aims to empower developers to bring more complex capabilities into Gemini, both for consumer and enterprise use cases.
“By moving beyond simple persona customization, Skills open the gates to richer AI automation—linking assistants, APIs, and real-world data in ways Gems couldn’t achieve.”
Impact for Developers and Startups Building on LLMs
This transition puts developers front and center. The new Skills system will require migration from existing Gems, presenting both technical hurdles and major opportunities. Google has made clear that its new platform will allow a new wave of third-party abilities, more closely echoing OpenAI’s plugin and GPT integration models. Reportedly, Gemini users and developers must move their content and integrations before the July cutoff, or risk loss of their AI automation setups.
For startups and enterprise AI teams, this shakeup delivers a double-edged sword: on one side, better developer tooling and potential access to Google’s distribution; on the other, the threat of ecosystem churn and the need to rebuild existing solutions for the new environment. Migrations of this scale often generate friction, but they also level the playing field for emerging players to compete where first-movers previously dominated with Gems.
“For AI professionals and startups, the move to Skills is both a hurdle and an onramp—the end of DIY customizations, but the start of a competitive developer ecosystem.”
Commercialization and the AI Platform Power Play
The move to Skills reflects the intensifying competition among major LLM platforms to anchor their own developer marketplaces. OpenAI’s GPT Store, launched in early 2024, rapidly built traction with over 3 million GPTs and opened new revenue streams for AI builders. Google’s shift is widely seen as a step toward a comparable, possibly even interoperable, marketplace for Skills—where developers could commercialize their creations and enterprises could deploy compliant, organization-specific integrations at scale.
Amazon and Microsoft have pursued parallel strategies, with Amazon Bedrock and Azure AI leveraging hosted model “skills” and marketplace APIs as selling points in their cloud ecosystems. The AI arms race is no longer model-versus-model, but platform-versus-platform—where capability libraries, developer monetization, and deep integrations shape what LLMs can do in real-world scenarios.
“The race is on for AI platforms not only to deliver better models, but to capture developer mindshare by offering the richest, most extensible ecosystems.”
User Experience and AI Assistant Interoperability
Beyond developers, end users will notice the shift as Gemini’s new Skills bring more robust workflows and deeply integrated actions. Unlike Gems, which often lived as isolated behaviors, Skills are positioned to interact with broader software solutions, data sources, and enterprise workflows. This move could accelerate AI assistant interoperability—enabling scenarios where Skills interact seamlessly across devices, apps, and even competing LLM platforms if standards emerge.
What Developers Should Do Now
- Audit existing Gems for valuable automations that need migration
- Explore early-access documentation on Skills APIs and developer programs
- Assess where new monetization or user acquisition opportunities may emerge as Google opens Skills distribution
- Monitor for interoperability standards among LLM platforms, which could boost Skills’ visibility outside the Google ecosystem
The Road Ahead for Generative AI Ecosystems
Google’s overhaul of Gemini’s customization layer from Gems to Skills marks a strategic pivot that will shape how developers create, monetize, and deploy new AI-powered capabilities. As generative AI shifts from experimentation to critical infrastructure, the ability to build reliable, extensible, and secure integrations becomes essential. Expect the Skills model to drive a new wave of competition—not just between LLM architectures, but among the ecosystems and marketplaces built around them.
The coming months will reveal whether Skills can energize developer participation at Gemini’s scale, or if first-mover advantage keeps OpenAI’s GPTs in the spotlight. Either way, the interplay between AI platforms, developer ecosystems, and real-world adoption will determine whose assistant powers tomorrow’s work and life.
Source: TechCrunch
