Okay i'm gonna learn java now, start small and just grow.
Is this a better idea?
It can be a better idea, if your idea is to be a programmer.
Just know that your first projects will be things like getting "Hello, World!" to appear on the screen.
Then you'll graduate to "guess the number" which is a very popular beginners game.
Eventually you'll reach the point where you can make "Tic-Tac-Toe" AI in under a week. Don't worry, you'll get faster as you gain experience.
Then perhaps you will move on again to make text-based dungeon crawlers, your first will probably take days, weeks, or months, but an experienced developer can churn one out quickly. (I made one about six months ago for a column on the site, took about two hours... but I've got three decades of development experience.)
One day, perhaps after a few months of gaining experience in the text based game arena, you'll want to make your first Pong clone. Then when you figured it out, a breakout clone.
When learning to program, getting to the point where you make your own 3D engines is a long way off, typically several years in. Once you finally get to the point where you can understand it enough to do it yourself, you will pick up an engine and jump in, quickly figuring out exactly the few pieces you need to make a full-featured game. But if you didn't have all that experience in programming simple stuff, you may not survive the baptism-by-fire of jumping directly into a major engine.
Most people learn very well by starting small, then growing line on line, principle on principle. Eventually you will reach mastery and also continue growing and learning.
That being written...
Some people DO learn best by jumping in the deep end and struggling hard. For those few individuals, perhaps using a major engine and jumping right in (and trying not to drown) is best for your learning style.