I read it as "... if there's syntax cleaning/prettier when the markup gets petrified?" 🤣

By doing so, they are helping improve large models and, hence, their competitors as well. As delivery velocity becomes less of a differentiating factor (throw copyrights into the mix), I suspect some sort of closed model will be the way to protect IP and offer unique value. No 🔮 just thoughts 😀

I think the combination of push towards the local models and smaller dedicated models along with improvements in hardware and learning techniques (under the guise of cost, security, and autonomy) will likely catalyze such a shift. We will see 😀

You are correct if we use AI-enabled editors powered by non-local AI-engines. The key is to use the content with local models/engines (think closed source) or with fully open/public models/engines (think LGPL).

With "AI" companies slurping up and using content without attribution or compensation while needing more content and validation to become better, I think controlled sharing of code may be the way forward.

🤣🤣🤣

🤣🤣🤣

💯 @wired.com had been doing a kickass job of covering news that is usually covered by MSM. I have been recommending Wired in my circle for some time now 😃