Oh boy, another AI CEO feigning concern about the development of AI in order to convince governments around the world to implement their regulatory capture schemes.

Yes, AI models will be used maliciously by companies and governments across the world, just like every other advancement in technology has been. Instead of creating an artificial barrier for new players to the field with no money while helping governments write the rules on what’s acceptable in a model to block out others, we should be educating people on the potential dangers and biases of AI models and punishing actual malicious uses. I know it sounds like some “free market will punish bad actors” bullshit and I hate to tread down that path too heavily, but if the alternative is to allow early birds with big money to be the ones who write the rules governing themselves so that they can lock out everyone else, I’d much rather take the option that allows greater free access to this field that has so much potential to help humanity as well.

Is the prevailing advice still not to bother with self-hosting email?

Unfortunately, yes. Here’s a fantastic blog post from someone who gave up on self-hosting last year due to the vast number of problems with self-hosting. The post also does a great job at pointing out why the death of self-hosting email is so bad in general, but still regrettably concludes that it isn’t worth it in any scenario right now.

parses HTML with regex

This is really what brings the meme together, it’s just so… chef’s kiss

Any idea why Tesla went against that to a shatter proof glass?

Because Elon Musk is a manchild who thought it was cool and forced his engineers to entertain his idea? The entire design of the Cybertruck just screams “someone high up had a shitty idea and told engineering they had a month to figure out how to make it real without any regard for rules or regulations”.