My current AI stack for each use:
General → Gemini 2.5 Flash
in Gemini App
Coding → Gemini 2.5 Pro & 3.7 Sonnet
in Gemini App & Replit
Frontend → DeepSeek V3
in DeepSeek Chat
Image → GPT-4o
in ChatGPT
Video → Veo 2
in AI Studio
Search → Gemini & Sonar
in Google AI Mode & Perplexity
Report gen → Gemini 2.5 Pro
in Gemini App Deep Research
Spreadsheets → Gemini AI Formula
in Google Sheets
Custom bots → GPTs & Gems
in ChatGPT & Gemini App
Automations → custom AI scripts
in Apps Script & Cloudflare
Private (laptop) → Mistral Small 24B
in LM Studio
Private (mobile) → Gemma 3 4B
in PocketPal
I just built an MCP-powered RAG app!
- It runs inside your Cursor IDE.
- Can process any complex, real-world doc.
- Ingest a new doc using its path in the Agent chat.
Here's everything you need to build this:
Introducing https://t.co/7dJrzFur6Y
The fastest way to concatenate any website into the single file standard for AI.
Just add https://t.co/CVta8FqXbv in front of any url.
@gammaapp Your website creation solution is great, but not feasible to use because the page loading speed is still very slow. If you improve the loading on mobile to under 1.5s, it would be great.
We made an AI Agent that can create Google Ads for you right from ChatGPT*
- Researches your company (searches internet)
- Does keyword research (with live data from Keyword Planner)
- Analyzes keywords & generates the full ad copy
Comment / dm and I’ll give you access.
(It’s our own chat interface, which uses Claude under the hood)
I generated a developer portfolio website using https://t.co/VpcykR8F2o with my Anthropic key. Model used is Claude Sonnet.
Initial generation is good and comparable to https://t.co/6tSuA3jo6N
Tokens used
input 26,288
output 6,218
approx cost $0.17 for this generation
Raw video
AI Coding Crash Course:
(Everything you need to know to get started)
You can build anything with AI code, If you know how to talk to AI models.
You need to learn:
- Development Documentation
- What AI tools to use
- AI prompting
That's it. Don't overcomplicate it.
Let's expand: