Quick, nuke the server.

My non-AI brownies came out MUCH better and I even used fewer ingredients. Human verification, at least for now, still has some place.

In retrospect, the AI model didn't seem to consider that just because I have an ingredient doesn't mean I need to use it, even if some worse recipe does.

I'm sure my prompts played some role in this, but the fact remains that it was far less work for me to use current non-AI tools to find a recipe that other humans liked.

AI also made me miss out on the serendipity of discovering that I have an ingredient I'd forgotten about (because the website gave me a recipe that was one ingredient shy in my results). It turned out to be highly rated, used fewer ingredients, and called for a close alternative I thought I didn't have but did.

I don't have a big grand point here except to say that AI isn't the magic bullet everyone says it is. It's cool, but there's still times when it just doesn't get humanity. We're messy, forgetful, and care about how our food tastes. Yes, AI gave me brownies, but they were off... Kind of like the double thumbs thing, except with food texture. And it was probably more expensive than just using a dumb search tool that has existed for years.

I've had the same experience on many websites looking for things like engine coolant. Instead of entering my vehicle attributes in ten seconds and getting a product, I'm forced to go back and forth with a chat bot that frustrates me as it takes much longer or fucks up. We're reinventing wheels too early in many cases.

Anyway, that's my brownie story. Thanks for reading.

For anyone interested in saving compute and reading my AI summary:

Matt Shumer, the founder of Hyperwrite, wrote an essay titled "Something Big Is Happening" warning that artificial intelligence (AI) has reached a pivotal moment where it can now substitute for human expertise in technical work. He argues that the world is on the cusp of a transformation that will disrupt nearly all white-collar professions, including law, finance, medicine, and design.

Shumer draws a parallel to the early days of the COVID-19 pandemic, stating that the world is in the "this seems overblown" phase, but soon AI will have a significant impact. He highlights that AI systems can now autonomously write, test, and refine complex software applications, making intelligent decisions and exhibiting "judgment" and "taste."

The essay was sparked by the release of OpenAI's GPT-5.3 Codex and Anthropic's Opus 4.6, which marked a shift in AI's capabilities. Shumer emphasizes that AI is not just executing tasks, but contributing to its own development, with OpenAI confirming that GPT-5.3-Codex was instrumental in debugging, deploying, and evaluating its own training.

The essay has garnered over 40 million views and widespread attention, with some critics arguing that it may overstate the immediacy and universality of job disruption. However, Shumer's intention was to convey the urgency of the moment and encourage those outside the tech industry to take notice of AI's current capabilities.

The essay has sparked a broader conversation about AI's accelerating pace, the need for societal adaptation, and the importance of AI literacy, strategic thinking, and resilience in the face of rapid change. Key points include:

* AI has reached a pivotal moment where it can substitute for human expertise in technical work
* AI will disrupt nearly all white-collar professions
* The world is in the "this seems overblown" phase, but soon AI will have a significant impact
* AI systems can now autonomously write, test, and refine complex software applications
* AI is contributing to its own development, exhibiting "judgment" and "taste"
* The essay has sparked a broader conversation about AI's accelerating pace and the need for societal adaptation.