#pybeach — Public Fediverse posts

On Saturday I attended PyBeach, my local Python conference. Here's a summary to give you a taste of it: https://bernat.tech/posts/pybeach-2026-recap/. If you're interested in Python, I recommend joining next year! #pybeach

Slowly trying to catch up on 650 posts from mastodon like an animal. Seeing #PyBeach in my timeline, I wish it didn’t collide with my racing again this year. Someday…

RE: https://mastodon.social/@RonJeffries/117502908774333109

Yesterday at #pybeach, I mentioned that a relatively common AI tooling error is reversing negation... here's an example 😅

(Thanks for sharing it, @RonJeffries !)

#PyBeach notes

And just like that, #pybeach is over. What a lovely conference and what a lovely group of people. I feel lucky to be able to call them my friends. ❤️

(this was in my drafts, but didn't go out!) Final (non-lightning) talk of the day, @MaggieFero explains the importance of good documentation practice, regardless of who is reading it. #pybeach #pybeach2026

Today's top ten tag trends:

10: #sabafoto
9: #ScreenshotSaturday
8: #zeitumstellung
7: #lhp3
6: #hochwasser
5: #Cats
4: #CatsOfMastodon
3: #pybeach
2: #Caturday
1: #tvTalkShowASong

#trendytoots 📈

@phildini is it "3 (or whatever) ways to put something on the internet" because you didn't want to write "3 or 4" after this week? #pybeach

tl;dr: continue writing useful docs well. 🫠

@MaggieFero #PyBeach

@MaggieFero documenting anti-recommended settings is far more risky in the age of LLMs. The chance of negation means that AIs might recommend the settings you don’t want used.

#pybeach

@MaggieFero the reading attention span of users has only decreased, and keeping this in mind only compounds with LLM context limits.

The LLM doesn’t have an “attention span”, but if the doc is too big it can’t fit in context and won’t be tokenized into a representation that returns authoritative results.

#pybeach

@MaggieFero

For the age of AI, it’s more important than ever to write docs for a specific audience, as well as match queries well using GEO or AEO techniques

(Generative engine / answer engine optimization)

News items perform well, mixed content appears less authoritative, content length impacts relevance

#pybeach

"generative engine optimisation (GEO); answer enging optimisation (AEO)" - @MaggieFero
"Old McDonald had a farm (EIEIO)" - Me

#PyBeach

@MaggieFero AI greatly improves discovery

- secrets in commits
- creds in commits
- live keys in common crawl

Credential abuse is still 13% of breach causes. Security by obscurity is already failing, AI makes it even worse

#pybeach

@itsthejoker @MaggieFero be the tooter you wish to etc etc #pybeach

@MaggieFero docs can impact security!

Preset auth material (default passwords), anti-recommended settings, and revealing potential vulnerabilities are all potentially impactful and regularly show up in docs

#pybeach

Up next at #pybeach, @MaggieFero is covering “Technical Documentation in the Age of LLMs”

Things may not be as different as you think!

Also this talk will cover most of what we currently call AIs or LLMs

“oh shit, it’s a certificate key!” #pybeach

In closing his #pybeach talk, @andrew informs us all that mountain rescue in the US is always free, and they'd rather be called sooner than later if you think you might need it.

As a security professional, please also notify us sooner than later if something feels weird, whether you clicked a link that felt wrong in retrospect or a process is suddenly taking an unexpected length of time. Most of us wanna help!

"Clear leadership is critical": "rescue is not a democracy", advice must be considered but you need a clearly accountable individual who can make the call. You can't do everyone's thing every time. - @andrew @ #pybeach

extremely important point from @andrew: safety culture has *two* parts:

1) blamelessness
2) everyone can meaningfully say "stop"

tech organizations tend to do reasonably-well at (1) and utterly lack (2)

#pybeach