filter

Florian's Filter — W15 2026

Florian Hollender

12 Apr 2026 — 1 min read

I have been thinking a lot lately about the gap between what AI promises and what it actually delivers in practice. We are seeing a fascinating tension between massive corporate narratives and the more transparent, open-source movements emerging underneath them. This week's selection touches on that friction, covering everything from security vulnerabilities to the way we visually represent technology.

Articles

Better Images of AI
How we visualize something shapes how we think about it. This initiative provides a repository of more realistic and less misleading images to help diversify our mental models of AI.
Using AI to Contribute to Open Source | VisiData
As open source projects face a flood of bot-generated pull requests, this is a sensible proposal for categorizing and flagging different levels of AI support in writing software.

Videos

Nvidia Just Open-Sourced What OpenAI Wants You to Pay Consultants For.
Interesting thoughts on the business adoption of the latest innovations from major LLM providers and the obstacles currently standing in the way.
AI Whistleblower: We Are Being Gaslit By AI Companies, They’re Hiding The Truth! - Karen Hao
An interesting interview with author Karen Hao, who has been shining a critical light on the "Empire of AI" for years now.
McKinsey's AI Lie Explains What's Happening to Work
A critical view of the "agents instead of analysts" trope currently circulating within consulting firms.
OWASP's Top 10 Ways to Attack LLMs: AI Vulnerabilities Exposed
IBM is producing very good explainers lately. This video breaks down the primary ways to attack Large Language Models.
Little Tikes Cozy Coupe Review
I have a thing for good car reviews and this is my favorite for the beginning of April.

Florian's Filter — W16 2026

I have been spending a lot of time lately thinking about where automation ends and true agency begins. It is easy to get lost in the hype of generative models, but the real interest lies in how these technologies integrate into our workflows and even our philosophical outlooks. This week&

The e-bike that wouldn‘t

Oftentimes I encounter problems that I never imagined one could have. Like: „my bike doesn‘t turn on“. Yeah. So my awesome e-gravel-bike that I got this summer seems to have an issue. And if I still had hair… I would be pulling it out over this. Trigger warning: the

Awesome problems?

When I updated this website I thought about the theme that I want to follow here for the next time. Although this is my publicly private place, I still like to have some general direction, a common thread that stitches the content together. The phrasing "awesome problems" is

B2B Customers. Know them.

There are many wrong ways to know your customers. With Consumers, statistics can mislead. Averages. You might ask for opinion, for comparative preference. Not for their money. There are many more wrong ways to know your B2B customers. * Through the words of enterprise sales. * Through the numbers of corporate strategy.