I'd like to add to this that I feel that it's a fine balance to be made, but that right now the limits were from a time where on top of the events system and layers system, Construct 3 did not have much to offer. So limiting that heavily was good, but now that the engine is so much more capable, I think that holding back on more extensibility features like Plugins, Timelines and JavaScript is already a major incentive to subscribe.
The balance weighs between limiting too little and having people develop a large portion of any game while off subscription. But on the other hand, limiting too much and having new users not invest any time into trying what real development feels like with C3 and also potentially making the engine look like a Toy.
I say this because it's truly hard for me to explain how powerful c3s event system is when seeing only the demos and 25 events. 25 events and 2 layers is very fine for maybe making a mario clone, but its hard to demonstrate that the engine is more than that with those limits.
I think atleast a few hundred events and just not having a limit on layers would be a good move because any game within a 1 year development scope is already forced to subscribe for a year. So no matter how simple a game is, for publishing they would subscribe for a year in minimum already. And for extended projects, for example 5 years, a 500 event limit or 1000 event limit would be hit by the first month or so, so as long as it isn't 1 year of free development, it's no revenue lost for the most part.
I think this conversation we had on discord, we also were discussing how Global Game Jam and game jams in general would be less in favour of C3 at the moment because while other engines have a comprehensive free engine for people to be learning, tinkering with, prototyping and generally using all year round, C3 only is fully available during select jams and so it's less likely people will be atleast a certain familiarity with the engine when that time of the year comes.
As with OP: I am and have been subscribed to Construct 3 for years so it's not about wanting more for free, but over the years I've encountered a few situations where people have responded to me trying to construct 3 as too large an investment to "try". And I think this would be improved with a free version that atleast gets them to a point they can see the benefits, and speed of C3 development.