Notices by scribe (scribe@mastodon.sdf.org)

@rek Would definitely be interested!

@evan Possibly the toughest poll I've ever had to do.

> new post: 'Things I’m Playing: Cryptmaster'

https://6days.exmosis.net/2026/02/16/things-im-playing-cryptmaster/

Why how who what is this a sensible default?

Life as a blur of triple-dot menus, as a form of art.

When everything feels like the movies / yeah, you bleed just to know you’re alive

> new post: 'stoicism in th’ face of th’onslaughts'

https://drpfd.exmosis.net/stoicism-in-th-face-of-thonslaughts/

Awesome, Ardingly reservoir looks fully refreshed finally.

Wonder if the drought order will be lifted before June.

#water

> new post: 'fragments 2026-01-26'

https://6days.exmosis.net/2026/01/26/fragments-2026-01-26/

Your LLM model name is your inability to teach people, plus the contents of your .ssh folder.

> new post: 'fragments 2026-01-20'

https://6days.exmosis.net/2026/01/20/fragments-2026-01-20/

Make your own AI-powered website by simply prompting people for their name on load, and then printing out your static content at random intervals.

"Privacy-Led Marketing™"

> new post: 'inwards, upwards'

https://drpfd.exmosis.net/inwards-upwards/

If you come to depend on a tool because you think it is omniscient, you're assuming that the Abstract is more real than the current, specific context. And any solutions built in top will default to the same, generic abstraction.

This is opposite to the hacker mindset, which seeks out specifics in order to arrive at non-obvious (but not ungraceful) next steps and solutions.

My specialist skill is knowing when and how to ask a human, rather than an AI.

Weirdly not a flippant comment. I'm trusted to pick up random in-place and often bespokely configured tech quickly. Existing team members, docs and git/bash history have specifics and memory that LLMs will obviously lack.

> new post: 'Things I’m Playing: Mundaun'

https://6days.exmosis.net/2026/01/08/things-im-playing-mundaun/

Please just stop using Evri. Please.

@firefoxwebdevs In the context of granular control they're quite good though? ie I can accept/delete them per domain as I like. I voted for the middle option as the only granular option, but granularity can be more fine-grained than that, is what I mean. And privacy is all about control over the granularity. Time-limited access tokens with very specific permissions are maybe a better analogy.

@firefoxwebdevs I think it's less about use cases and more about general trust, as privacy often boils down to. If you're talking about an AI "kill switch", you're talking about trust in what's been defined as "AI", and trust in the browser developer as a whole.

Once definitions are murky, there's an area open for ongoing redefinition. One way to adopt a "private by default" approach is to follow what cookies do, for instance, and allow users to allow limits to the extents of permissions.