Alternatively, you can use a typical, full Python kernel in JupyterLab by accessing it differently. The other options at the Try Jupyter page are served via the MyBinder service. You can tell by hovering over the other tiles and seeing ‘mybinder.org’ at the start of the URL. To save resources the two most popular were changed to JupyterLite to still give a sense of the interface but take demand off the service which runs on remote machines and therefore costs money. You can though opt in the full Python kernel backed offerings.
To do that:
Go here and press ‘launch binder’. You’ll get a temporary session on a remote computer without logging in. Using configuration files you can even control what is already installed, see here. Keep in mind the sessions are temporary and will time out with inactivity. And so be sure to download anything useful you make back to your local machine promptly. You can drag-and-drop from your local machine into the left side browser panel when you start a new session to pick up again later.