I’m personally partial to @tangled.org for now, though alternatives will be considered. It’s just sad that npm is owned by MS so github is pretty much the only platform you can publish OICD verified packages from.

Except that runs the first-run information every single time… (Don’t get me started 😉 )

None of the attributes that differentiate wsocial are actually good things: id collection? Publishing PII? If you want to know what the best thing you could do is then I’ll tell you: shit down W and donate all funds to actually valuable projects like eurosky and support them.

And let’s be clear: I know you’re trying to make a better world by getting europe away from its dependency on US big-tech, but the answer cannot be creating something with the politics of Elon Musk and the accountability of Mark Zuckerberg.

They are trying to get you to do one of two things: act ethically and transparently while doing basic due diligence (which this data-leak was definitely not) or to shit down and go away. You are getting pushback because tou are doubling down on bad behaviour when being called out.

Sorry everyone I have been fighting codex all day and needed to vent. I will be cancelling my openai subscription shortly to preserve my mental health, not with Ai, but with the absolutely horrendously tooling around it.

<rant>and also how can a company not realise that their first-run introduction is no good if it shows up on every run. And who ever thought having more than a single page of first-run content was acceptable? who actually reads that stuff</rant>

<rant>why on earth does vscode insist on having AI agents in an additional sidebar? and there is no way to just have them be a regular tab. Like does nobody ever test this on. Machine without an ultra-wide display? And even then, who says I want my window to be full width?</rant>

Why not rclone? Because davsync keeps the last seen state so it can do a 3-way compare and get better conflict handling when both sides could have changed.

There is something to that, but remember that every crossing of the js/cpp boundary also incurs a cost. So it’s about very careful tradeoff decisions. And there are few that I trust to make these decisions more carefully that mcollina.

There is actually ongoing conversations on how to do that and also gain performance over the current http stack.

But as for being independently and long-term useful I’m a bit sceptical. So I’d rather have the ecosystem keep experimenting and pushing than sustaining an experiment that has run its course.

To be honest I always respected Deno and Bin as things trying to push the boundaries. Experiments if you will. They pushed the ecosystem forward and that’s great.