Now that spacex has been saddled with twitter/grok, I wouldn't say its particularly healthy either, but I am no expert. SpaceX alone was a genuinely innovative company, and still is a leader in space, but I dunno if it will continue now, or if it will just become another meme stock.

This flying car thing is a joke

Also a late joke, Top Gear made a Vauxhall Astra fly decades ago.

https://m.youtube.com/watch?v=5eRdzSWv8NY

Similar stunt too, stripped it out, filled with helium.

Not all of his fan base are stock holders, there are others who are just stupid :/

I still go to reddit when bored at work, but to be honest, the comments are all basically the same. There is only so many times you can read the same reductive comments before you get bored.

It'd be nice for there to be more activity, but I'm happy that its at least humans (mostly) here.

Do it, the fediverse needs a tesla

I think for his supporters, keeping all the companies going builds into his mythos. If he lets any of them fail, suddenly he isnt on a perfect game anymore, and his valuation tanks. So he has to keep the circus going...

The people idolising him are too stupid to realise that he has failed a bunch of companies at this point.

So, one of those companies must be fucked?

For shits and giggles, I had a go at writing a script that could rewrite all the commits in a branch with LLM generated commits. Couldn't work out how to do the commit changing, tried cherry picking, but it doesnt work well for conflicts (had no idea that was possible). But, I can report timings at least.

On my M1 16GB macbook air, each commit was summarised in about 10s. So its a plausible approach if you just do all your commits, and then run a script to fix up the messages, go get a glass of water and you'll come back to automatic messages.

If it's of interest to you, I can share the script, its a starting point, but I am pretty ashamed of it :D

All the admins are listed on the instances themselves.

You realise most (all?) fediverse instances aren't operated by countries? They are just random groups of people. There are no legal agreements or blocs tying servers together.

I've seen and reported plenty of spammers, most don't post that often. They create a few posts and the go silent or get banned. Blocking after N posts still leaves N spam posts to cleanup.

Every post is federated to every other instance. My home instance could purge my posts, but all other instances would still have a copy that can't be deleted by my admins. So the modlog is resilient by virtue of being distributed copies. If the post was dropped upon submission, then nothing is seen, but thats hardly a silent censorship either, I can immediately see that it wasn't created.

I dunno, but I think you need a much stronger selling point to get people on board with this, it just seems at odds with reality.

Ideally, you'd want to get feedback from admins rather than general users. The moderation burden is basically hidden from us.

Given I am not European, any server that required EUID would be off-limits for me. But unless you plan on building it into activitypub itself, anyone who wants to use a non-EUID account can just spin up their own instance and use that completely unattested.

Post/comment limits per account would either be too low, and annoying for some users, or too high, and completely ineffective for combatting spam posts. Spammers will typically only send 1-3 posts per day.

I'm not sure what your bringing to the table for censorship either, the modlog is public, and federated, how does one get silently censored?

You're not wrong though, even 1GB is quite lot for writing commit messages, I gave it a quick go, and it takes around 15-30s to analyse a very small diff. Kinda silly for git messages.

Its small in the sense it'll run on basically any GPU from the last decade, rather than high end or specialised GPUs. And there is a 1GB gemma4 model, which is meant for CPU/phones.

https://victorzhou.com/blog/better-profanity-detection-with-scikit-learn/

That might be of interest to you then? Its a small model trained for profanity detection, might work for your use case. Dunno if it can be used for generation though.

Otherwise, there were 1GB gemma models, you can run them on CPU.