HeyJaredWim Sweldens
Articles by Wim Sweldens on navigating and quantifying the jagged edge between humans and AI
- Indexed issues, last 90 days
- 5
- Latest publication
- Sep 29, 2026
- Audience
- Checking…
- Earliest in this view
- Jul 12, 2026
Latest issues
Responsible AI in Filmmaking (opens the original)
Read excerpt
Last week, DreamWorks co-founder posted an essay on X titled “The World is Changing: AI For Creativity.” It caught our attention because we are developing a feature film about Christina Mirabilis, a 12th century Belgian mystic. We share a passion for the film’s themes of liberation, female empowerment, and mental health. Kathy is a Hollywood film editor, with credits including Genius: Einstein (National Geographic) and The Morning Show (Apple TV). Wim is a technologist, a JPEG2000 contributor,
Let AI make decisions … up to a point. (opens the original)
Read excerpt
Something changed in AI over the past six months. We used to prompt AI by telling it what to do and how to do it. These days, we tell the AI what the goal is and how to measure it, and then let the agent work out the best method to achieve that goal. This shift allows increasingly autonomous agentic workflows.The deeper change is where the optimization loop lives. It used to be entirely human: we set the objective, wrote the prompt, read the answer, judged it, adjusted the prompt, and iterated.
Contain AI. Contain Yourself. (opens the original)
Read excerpt
In the last three weeks we have seen 2.5 AI containment stories. Two significant ones from OpenAI (July 21) and Anthropic (July 30), and a smaller Meta one (August 5) that felt like a me too. The media quickly jumped on a narrative of AI circumventing safeguards, escaping, and going rogue. My father, 87 years old and one of the most digitally capable great-grandfathers, asked me if he should worry. The stories fueled both the AI cheerleaders (”look how smart AI is”) and the AI fearmongers (”look
AI is the new UI (opens the original)
Read excerpt
A few weeks ago I connected Claude to Slack, Google Drive, Notion, and Gmail through the Model Context Protocol (MCP), an open standard that lets an AI system see what a piece of software can do and call it directly. Since then I rarely open those four apps myself. I mostly go back to fix something the agent missed, or to work around a limitation, like an attachments MCP can’t handle yet. That will get fixed soon. This is bigger than personal productivity: the very idea of an app having its own
Realigning AIQ (opens the original)
Read excerpt
<img alt="" class="sizing-normal" height="216.04945054945054" src="https://substackcdn.com/image/fetch/$s_!J1FJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0e57c480-6843-46dd-bcec-c2d2cf058a41_680x360.svg" wi
Publishing over time
Last 90 days. Choose a month to open its work.
Recurring subjects
Named in the text we hold. One piece can cover several.
Audience
No verified audience measurement yet.
About this data
Counts cover the work we have indexed. Tone needs enough text and a confident classification. Excerpts and episode notes are not full articles or transcripts.
Identity or attribution wrong? Suggest a correction.