This morning, I decided to pay approximately $22 for Claude AI. What did I want to be able to do? Create a Claude Project with my past presentations, custom instructions (which I had ChatGPT generate from a Project over there) with this goal in mind:

Create a new workshop presentation aligned to a new topic, but that captured elements of my previous work and blended them into a new creation.

Claude did the work flawlessly. Let me say that again. The first draft was perfect and so well-designed, I was able to make some minor additions and take off and use it.

I have tried a lot of different Gen AI tools at the $20 per month level, and Claude 4.6 just blew them all out of the water. While Google Gemini, NotebookLM are incredible, Claude 4.6 just became my number one choice. The problem with NotebookLM creations? PDFs that are watermarked. Same for images.

Vibe-Coding

Long ago, I started my vibe-coding journey on Claude. It seemed like it was always gasping for air. Today, what it generated was mind-blowing and amazing. This went so far beyond my prompt, it was like having an expert at my shoulder:

What’s really amazing about this vibe-coding, one-pager of HTML code is that there’s a LOT more to it below what you see in the screenshot. Not only did it vibe-code beautifully, it blended it core components of my previous presentations in a series right into it. This is work that would have taken hours, if not days, and as for vibe-coding, well, that would have been nigh impossible for me to do.

Claude did it ALL with my content in less than an hour. This is a huge time-savings, but also, my brain won’t be agonizing over all the layering in of ideas that I usually suffer through. While Claude isn’t perfect yet (hey, it’s Canva integration, although better than ChatGPT IMHO, needs some work).

But for my day job, Claude is my new go-to. I will need to investigate the context window to see how comparable it is…ugh, context window may not be up to snuff but it was enough for my project this morning. I’ll have to see how it shakes out over the next few months.

Model	Context Window (Chat Users)	1M Token Access
Claude Sonnet 4.6	200K tokens	Beta, API only (Tier 4)
Gemini 3 Pro	1M tokens	Yes, standard
ChatGPT Plus (GPT-5.2)	~196K–400K tokens*	No (Pro tier only)

*Varies by variant (Instant/Thinking/Pro) and tier. Verify at platform.openai.com.

Re-Evaluating My Gen AI Investments

As much as I enjoy and use Google ecosystem, Claude’s results today made me sit up and take notice. Here’s my ranking of the tools now:

• Primary AI Tool: Claude 4.6 Sonnet/Opus (and I relied mostly on Opus)
• Secondary, Backup: BoodleBox Unlimited – All-purpose, general AI tool and bot maker that I use for image generation.
• Nice to have: Google Gemini Pro with NotebookLM
• On the Cutting Room Floor: ChatGPT Plus

I’ve said it before, if I have to cut, I will probably do away with the last two to save money.