Sunday Rundown #122: Vibe Coding & "Bulls(hit)-eye"
Sunday Bonus #82: Guide to infographic styles for Nano Banana Pro
Happy Sunday, friends!
Welcome back to the weekly look at generative AI that covers the following:
Sunday Rundown (free): this week’s AI news + a fun AI fail.
Sunday Bonus (paid): an exclusive segment for my paid subscribers.
In case you missed it, here’s this week’s Thursday deep dive:
Note: If you’re consistently missing out on my emails, remember to check your “Promotions” tab and mark whytryai@substack.com as a “Safe Sender.”
Let’s get to it.
🗞️ AI news
Here are this week’s AI developments.
👩💻 AI releases
New stuff you can try right now:
Adobe launched Acrobat, Express, and Photoshop for free inside ChatGPT, so you can create and edit images and documents using natural language.
Alibaba news:
Qwen Code v0.3.0 is a major update with streaming JSON input/output, easy English/Chinese switching, and a bunch of stability fixes.
Qwen3-Omni-Flash is now better at audio-visual conversations, multilingual compliance, and human-like speech with fully customizable system prompts.
Cursor launched a Visual Editor in the Cursor Browser so you can make visual UI edits, drag-and-drop elements, adjust sliders, tweak styles, etc.
DiffSynth-Studio dropped Qwen-Image-i2L, which turns one image into a custom LoRA (low-rank adaptation) to copy its style.
Google integrated Gemini 3 into its experimental design tool Stitch, so you can generate high-quality UI and “stitch” screens together into working prototypes.
Mistral AI open-sourced a coding model Devstral 2 and a Mistral Vibe CLI command-line tool that automates coding tasks from natural language prompts.
OpenAI is rolling out GPT-5.2, with strong performance in coding, science, vision, tool use, and long-context understanding for professional knowledge work.
Orchids launched a coding IDE that can see your screen and hear your voice instructions as it builds apps from simple requests.
Z.ai open-sourced GLM-4.6V, a multimodal model with native tool use that can process image and video inputs.
🔬 AI research
Cool stuff you might get to try one day:
Google is working on a GenTabs feature in Google Labs that turns your open tabs and chats into interactive web apps. (Join the US-only waitlist.)
OpenAI is rumored to be testing new Image-2 models spotted on LM Arena, showing early boosts in detail and realism.
Runway announced GWM-1, a trio of general world models built on its Gen-4.5 video model and designed to simulate and interact with the real world.
📖 AI resources
Helpful AI tools and stuff that teaches you about AI:
“ChatGPT Foundations for Teachers” [COURSE]—free Coursera offer by Open AI.
“The state of enterprise AI” [PDF REPORT]—findings from a survey of 9,000 workers in 100 enterprises by OpenAI.
🔀 AI random
Other notable AI stories of the week:
Anthropic donated its Model Context Protocol (MCP) to the new Agentic AI Foundation as an open standard for connecting AI models with tools and data.
Disney and OpenAI signed an agreement that officially lets Sora 2 users create videos with 200+ characters (Disney, Marvel, Pixar, and Star Wars).
McDonald’s pissed off the entire Internet with a lazy AI-slop ad that should have probably been the true AI fail of the week.
🤦♂️ AI fail of the week
Tom’s friends have turned white lies into pure gaslighting.
Send me your AI fail for a chance to be featured in an upcoming Sunday Rundown.
💰Sunday Bonus #82: 11 Nano Banana Pro infographic styles and how to use them
Google’s Nano Banana Pro can do a lot of things. Just two weeks ago, I shared more than 100 possible use cases for it.
One thing Nano Banana Pro does exceptionally well is creating infographics. Simply asking for “an infographic about [BLANK]” will typically spit out a polished result.
But you can get way more control by specifying the visual aesthetic/style of the infographic. They all have their strengths and are individually well-suited for certain topics.
Below, I’ll share 11 infographic styles, a comparison table with copy-paste starter prompts, and Nano Banana Pro outputs that showcase each style.