Summary of "Free AI Tools: Tested by 2 Techies | E44"
Overview
Two hosts, Niha and Nidhi, review, compare, and demo practical AI tools across four use cases:
- AI for thinking (conversational LLMs)
- AI for creating (images, video, music, audio)
- AI for building (no-code / code-assisted)
- AI content workflows
The episode was recorded in March 2026. The hosts report hands‑on testing and give pricing/usage notes throughout.
AI‑for‑thinking (conversational LLMs)
ChatGPT (OpenAI)
- Best all‑round conversational partner: browsing, image generation, data analysis, code generation, and strong reasoning.
- Free tier is capable; ChatGPT Plus ~ $20/month for newer models and higher usage.
- Good for brainstorming and iterative back‑and‑forth conversation.
Claude (Anthropic)
- Strongest writing quality, follows nuanced instructions, and offers good tone control.
- Very large context window (~200k tokens) — can ingest very long documents.
- Product variants: Claude Co‑work (tool integrations) and Claude Code (coding product).
- Free tier available; paid tiers (Pro ~ $20/mo; enterprise tiers for larger customers).
Gemini (Google)
- Tight integration with Google Workspace (Docs, Drive, Gmail, Slides, Meet).
- Advantageous for real‑time or fresh information via Google Search integration.
- Best choice if you want AI embedded in existing Google workflows.
Perplexity
- Search + answer experience with inline citations to sources — ideal for research and verifying claims.
- Free tier; Pro ~ $20/mo.
Practical tips for conversational AI
- Use “projects” / memory features to store context: templates, brand guidelines, past notes so the model remembers across sessions.
- Save prompt templates (e.g., social-post formats, episode briefs) to speed workflows.
- Many power users use multiple models depending on the task (writing vs research vs workspace integration).
AI‑for‑creating (images, video, music, audio)
Canva (Magic Studio)
- Text‑driven image/video/presentation generation; great beginner experience.
- Some AI features in the free plan; Pro ~ $13/mo.
Advanced image & video tools
- Examples: Runway, Sora, Cling AI, Google tools (referred to as View / Nano Banana).
- These produce higher‑quality or more professional content; better for advanced video/image creation.
Suno (music generation)
- Generates full structured songs (vocals, instrumentals, chorus/bridge) from text prompts.
- Suno v5 sounds close to human, according to testing.
- Free credits available; paid plans for heavier use. Ranked among top AI music generators (2025).
NotebookLM (Google / powered by Gemini)
- Upload docs, PDFs, audio, videos — produces summaries, study guides, FAQs, and visual briefs.
- Unique features: audio overviews (two AI voices discussing content) and “Discover” visual brief generation.
- Uses only supplied sources, reducing hallucination risk.
- Useful for students, podcasters, product teams (creates training quizzes, briefs). Supports 80+ languages. Free tool.
AI‑for‑building (no‑code / white‑coding / developer assistance)
White‑coding
White‑coding: describe apps in plain English and the AI writes full code, databases, and UI — enables non‑coders to build real products/MVPs. (term popularized by Andrej Karpathy)
- Used by startups and prototypes; not always enterprise‑ready but effective for side projects and MVPs.
Lovable (no‑code builder)
- Natural‑language site/app creation; fast prototyping and design control via prompts.
- Limitations: daily/free credits can run out quickly — paid plans recommended for heavy use.
Replit (browser‑based coding)
- Write and run code in the browser; now with mobile support.
- AI features & agents can generate, test, and fix code; good for non‑technical builders who want more control.
Claude Code (Anthropic)
- Developer‑focused: reads entire codebase, edits across files, plans implementations, and integrates with workflows.
- Part of Anthropic’s ecosystem alongside Claude chat and Co‑work.
- Reported rapid commercial uptake by reviewers.
Practical constraints
- Generated code often needs review; models can loop or struggle on complex enterprise features.
- Best suited for prototypes, MVPs, and small business apps.
Key product / feature comparisons & reviewer takeaways
- ChatGPT: best all‑rounder for conversation and multi‑feature usage.
- Claude: best writer and deep‑context thinker (excellent for long documents).
- Gemini: best when embedded Google integration and fresh web data matter.
- Perplexity: best research tool for verifiable answers and citations.
- Suno: one of the top AI music generators (2025 reviews).
- NotebookLM: underrated for digesting large corpora and producing audio/visual study outputs.
- Lovable / Replit / Claude Code: top options for no‑code and AI‑assisted code building.
Jargon buster (five terms explained)
- Agents: AI systems that act autonomously on your behalf (can browse, fill forms, book, run workflows). They range from simple question‑answer helpers to fully autonomous workflows (examples: OpenAI operator, Replit agents).
- RAG (Retrieval‑Augmented Generation): an “open‑book” approach — the model retrieves relevant external/doc sources at query time, then generates answers to reduce hallucination.
- White‑coding: building apps using natural‑language prompts where AI writes the code and infrastructure.
- Fine‑tuning: specializing a general model on your data to change its behavior (for example, a medical specialization).
- MCP (Model Context Protocol): a standard/connector (analogy: USB) that lets models interact with external tools (CRM, calendar, Slack) and take actions.
Best practices & recommendations
- Start with free tiers; try one problem/tool at a time to avoid tool overload.
- Use multiple models depending on use case (writing vs research vs workspace).
- Invest upfront: create projects, upload context, and save templates — the model returns more value over time.
- For building: clear prompting and detailed specifications matter — AI is powerful but requires good instruction.
- Expect limitations: models can hallucinate, mispronounce, and require iteration and code review.
Next episodes
- Part 3: AI ethics (bias, privacy, deepfakes, ownership of AI content, job impacts).
- Part 4: Guidance for parents and the next generation.
Main speakers / sources
- Hosts:
- Niha — computer science engineer + MBA, 20+ years in tech/leadership
- Nidhi — engineer + business degree, 15+ years in product/tech
- Companies / tools discussed: OpenAI (ChatGPT), Anthropic (Claude, Claude Code), Google (Gemini, NotebookLM), Perplexity, Canva (Magic Studio), Runway, Sora, Cling AI, Nano Banana, Suno, Lovable, Replit.
Category
Technology
Share this summary
Is the summary off?
If you think the summary is inaccurate, you can reprocess it with the latest model.
Preparing reprocess...