Activity for bdbch.com

Sadly bugfixes don't help with PMF or reaching new customers directly. (I would argue that having a stable product is the best way to win over an unstable rival platform)

“Software vendors have long been KPI-oriented, and making things more stable doesn’t always have a direct effect on the numbers. It doesn’t look exciting in presentations: This quarter, we won’t be releasing any new features, and we have n…”

on ptrchm.com

Nice one - right now I mostly see it wasted on unnecessary demos. It may be a noisey minority but I believe if AI is used correctly, products could actually become better and more stable.

“but we still don’t use them to build better software”

on ptrchm.com

Important distinction: The developer will need to understand: A. If the LLM can be trusted with identification of the bug B. The bugfix is not causing new regressions Unit tests help - but they're also just helpful if a human knows what is actually being tested - otherwise who…

“LLMs can be really good at squashing bugs if given the chance.”

on ptrchm.com

Happened to the best of us - and it's infuriating.

“I opened Slack on macOS, the icon kept bouncing in the dock for a few seconds. I got impatient, switched to Ghostty, and started typing. Just then, the Slack window appeared, stole focus from Ghostty and the git pull command was sent to th…”

on ptrchm.com

Very good post about AI psychosis, things getting shittier day by day but the whole world is clapping and being enthusiastic because the golden dawn will come soon!

on ptrchm.com

Uh, finally stacked PRs on Github. ✨

Listening to some of the music I enjoyed as a late teen/early adult in the early 2010s and holy damn, I heard to the most depressing music I could imagine.

Just thinking: - Obsidian Sync on your own PDS - Notion alts - Discord alts - Family/Team planning apps - Figma / Figjam alts - Social media planning apps - AI tools for keeping memories, plans, configurations at one place - CRMs - Photo Libraries - Client / Freelancer tools (with sifa integration)

Dominik Biedebach@bdbch.com

I wonder what will happen when private data finally hits PDS. I mean people could use their PDS for all kinds of crazy stuff! 😄

I had a crashout over coding with LLM and had to write a post about my weekend experiment. Still quite new to blogging, don't be to harsh pls. 🥲 bdbch.com/blog/breakin...

Breaking out of the LLM cycle

How I stopped using LLM tools for everything and started writing code again.

bdbch.com

Read article

This sounds worrying. However I wonder what stops the right-wing from hosting their own PDS and Relays.

The Verge@theverge.com

Bluesky's new CEO Toni Schneider on the platform's reputation for being a liberal bubble: "Yes, we definitely want that to change. It is already changing. It certainly wasn’t designed to attract one specific group of people."

Personal opinion and recommendation: Avoid large conferences like WeAreDevelopers and embrace + enjoy smaller ones like @localfirstconf.com. I'm currently watching localfirstconf talks and all of them are more interesting than anything I've seen on WAD by a large magnitude!

On one side people run 24/7 AI agents to do the most mundane and unnecessary shit while on the other side the Rhine and Danube river are completely dried up rn. Dystopian as fuck.

I think a lot of OSS maintainers will resonate with this. The temptation to just throw an Agent on LLM-generated PRs is big but should not be the norm.

“He described waking up to thirty PRs every morning, each one pulled overnight by someone's AI, and needing to make snap judgment calls on every single one. The temptation to delegate the review itself to an AI was enormous. But, as he put …”

on pydantic.dev

Very important point - AI subtracts away small moments we had, not only alone but between each other. Small learnings or teachings, small wins, etc. Now usually are rewarded with a "This looks good, let's do X next".

“You and the machine, going back and forth, refining and prompting and reviewing. The natural moments where you'd turn to a colleague to ask a question, to rubber-duck a problem, to share the small victory of something finally clicking. Tho…”

on pydantic.dev

Not sure what to think about this. Technically it's right. Frontier AI becomes more and more dangerous - on the other hand, the people signing this were full-in for replacing "everything" with their AI models a few months ago and made fun because Europe was regulating. www.pacingthefrontier.com

So Opencode just silently dropped their Zero Data Retention info because of the new DeepSeek v4 Flash update I assume? opencode.ai/docs/go/#pri... What a shame, I'd rather use the old one and keep ZDR instead of now basically not being able to trust the providers anymore.

Good morning my absolute favorite social media platform on the internet. 👋