<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/stop-building-ai-wrappers-build-agents-mgd2riqzz" -->

---
title: Stop Building AI Wrappers  Build Agents | daily.dev
description: A podcast interview recorded at Google I/O Connect Bangalore with Prashant, Google Cloud&#x27;s developer adoption lead for APAC. Key topics include the difference...
canonical: https://daily.dev/posts/stop-building-ai-wrappers-build-agents-mgd2riqzz
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: Stop Building AI Wrappers  Build Agents | daily.dev
og:description: A podcast interview recorded at Google I/O Connect Bangalore with Prashant, Google Cloud&#x27;s developer adoption lead for APAC. Key topics include the difference...
og:url: https://daily.dev/posts/stop-building-ai-wrappers-build-agents-mgd2riqzz
og:image: https://api.daily.dev/og/posts/MGD2RIqzZ.png
og:image:alt: Stop Building AI Wrappers  Build Agents
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Stop Building AI Wrappers  Build Agents

**[Telusko](https://daily.dev/sources/telusko)** · 19 min read · 0 upvotes · 0 comments

## Summary

A podcast interview recorded at Google I/O Connect Bangalore with Prashant, Google Cloud's developer adoption lead for APAC. Key topics include the difference between MCP and REST APIs (MCP is a stateful protocol designed for LLM interactions vs stateless REST), why developers should build goal-oriented agents instead of simple AI wrappers, when to use local models like Gemma 4 vs cloud-hosted frontier models, and practical advice for traditional developers looking to upskill in AI by building personal projects rather than following structured learning paths.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.youtube.com/watch?v=JOXpBgb4Wk8>

## Questions this post answers

### What is the difference between MCP and a REST API?

MCP is a protocol designed specifically for how LLMs and models interact with services, while REST APIs are stateless and built around resources and verbs like GET and POST. MCP started as a stateful protocol built to handle the chatty, back-and-forth nature of long-running agent conversations efficiently, rather than requiring repeated overhead like traditional REST calls.

_Developers deciding how to expose services to AI agents can follow protocol comparisons like this one on daily.dev._

### Can Gemma 4 E2B or E4B models run locally on a mobile phone for AI tasks?

Yes, Gemma 4's E2B and E4B variants are small models with 2 billion and 4 billion parameters respectively, and they can run on mobile devices. They retain the reasoning, tool-calling, and thinking capabilities distilled from the Gemini 3 architecture, but lack the broad world knowledge of larger hosted models like Gemini 3.1 Pro or 3.5 Flash, making them well suited for tasks like text summarization or local RAG over documents rather than deep reasoning.

_Builders weighing local versus cloud models for agent projects can track model capability trade-offs on daily.dev._

### What programming language do AI coding tools default to when building a new project?

Most AI coding assistants tend to default to React for the frontend and Python for the backend when given a use case without a specified language, based on observed patterns. However, the actual choice can vary by tool and use case; one described example had an AI agent choose Rust for a screen-and-face recording application because it needed hardware-level access.

_Developers letting AI pick their stack can compare real-world language defaults on daily.dev._

## Similar posts on daily.dev

- [Bridging the Gap - How We Made AI Agents 10x Developers in Our Organization](https://daily.dev/posts/bridging-the-gap---how-we-made-ai-agents-10x-developers-in-our-organization-9xldqiaaq) · DEV · 4 upvotes · 0 comments
- [AI Agents Are Growing an Infrastructure Layer](https://daily.dev/posts/ai-agents-are-growing-an-infrastructure-layer-jk8bjcch7) · Medium · 0 upvotes · 0 comments
- [Build Better AI Agents: 5 Developer Tips from the Agent Bake-Off](https://daily.dev/posts/build-better-ai-agents-5-developer-tips-from-the-agent-bake-off-r4rpn57rq) · Google Developers · 0 upvotes · 0 comments
- [Episode \#266: Dangers of Automatically Converting a REST API to MCP – The Real Python Podcast](https://daily.dev/posts/episode-266-dangers-of-automatically-converting-a-rest-api-to-mcp-the-real-python-podcast-wbndjr0i3) · Real Python · 1 upvotes · 0 comments

---

Tags: [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#mcp](https://daily.dev/tags/mcp), [#google-gemini](https://daily.dev/tags/google-gemini), [#gemma](https://daily.dev/tags/gemma)

[View this post on daily.dev](https://daily.dev/posts/stop-building-ai-wrappers-build-agents-mgd2riqzz)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"Stop Building AI Wrappers  Build Agents","url":"https://daily.dev/posts/stop-building-ai-wrappers-build-agents-mgd2riqzz","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/stop-building-ai-wrappers-build-agents-mgd2riqzz"},"datePublished":"2026-07-22T11:37:31.210Z","dateModified":"2026-09-13T18:52:32.293Z","description":"A podcast interview recorded at Google I/O Connect Bangalore with Prashant, Google Cloud's developer adoption lead for APAC. Key topics include the difference...","image":"https://i.ytimg.com/vi/JOXpBgb4Wk8/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/JOXpBgb4Wk8/sddefault.jpg","isAccessibleForFree":true,"articleSection":"Telusko","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Telusko","logo":"https://media.daily.dev/image/upload/s--xVEBUipN--/f_auto/v1728072944/logos/telusko","url":"https://daily.dev/sources/telusko"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/stop-building-ai-wrappers-build-agents-mgd2riqzz","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,ai-agents,mcp,google-gemini,gemma","timeRequired":"PT19M","video":{"@type":"VideoObject","name":"Stop Building AI Wrappers  Build Agents","description":"A podcast interview recorded at Google I/O Connect Bangalore with Prashant, Google Cloud's developer adoption lead for APAC. Key topics include the difference...","thumbnailUrl":"https://i.ytimg.com/vi/JOXpBgb4Wk8/sddefault.jpg","uploadDate":"2026-07-22T11:37:31.210Z","duration":"PT19M","url":"https://api.daily.dev/r/MGD2RIqzZ","embedUrl":"https://www.youtube.com/embed/JOXpBgb4Wk8"}}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Telusko","item":"https://daily.dev/sources/telusko"},{"@type":"ListItem","position":3,"name":"Stop Building AI Wrappers  Build Agents"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/stop-building-ai-wrappers-build-agents-mgd2riqzz#faq","mainEntity":[{"@type":"Question","name":"What is the difference between MCP and a REST API?","acceptedAnswer":{"@type":"Answer","text":"MCP is a protocol designed specifically for how LLMs and models interact with services, while REST APIs are stateless and built around resources and verbs like GET and POST. MCP started as a stateful protocol built to handle the chatty, back-and-forth nature of long-running agent conversations efficiently, rather than requiring repeated overhead like traditional REST calls. Developers deciding how to expose services to AI agents can follow protocol comparisons like this one on daily.dev."}},{"@type":"Question","name":"Can Gemma 4 E2B or E4B models run locally on a mobile phone for AI tasks?","acceptedAnswer":{"@type":"Answer","text":"Yes, Gemma 4's E2B and E4B variants are small models with 2 billion and 4 billion parameters respectively, and they can run on mobile devices. They retain the reasoning, tool-calling, and thinking capabilities distilled from the Gemini 3 architecture, but lack the broad world knowledge of larger hosted models like Gemini 3.1 Pro or 3.5 Flash, making them well suited for tasks like text summarization or local RAG over documents rather than deep reasoning. Builders weighing local versus cloud models for agent projects can track model capability trade-offs on daily.dev."}},{"@type":"Question","name":"What programming language do AI coding tools default to when building a new project?","acceptedAnswer":{"@type":"Answer","text":"Most AI coding assistants tend to default to React for the frontend and Python for the backend when given a use case without a specified language, based on observed patterns. However, the actual choice can vary by tool and use case; one described example had an AI agent choose Rust for a screen-and-face recording application because it needed hardware-level access. Developers letting AI pick their stack can compare real-world language defaults on daily.dev."}}]}
```

