Reposting/aggregating from Discord so it's not buried forever.

Right now, there are very few uses for GTP that gates can't do for a fraction of the price. GTP probably needs a buff, yes, but in the meantime I (and others) think gates should probably be made a bit less awesome. We have two suggestions for this:

• Require a position to be given when creating the gate iota, not when teleporting entities using it. This prevents gates from being a strictly better GTP, while still allowing permanent teleportation networks, item collectors for stationary builds, and such.
• Add another type of gate iota, similar to a true name, that always teleports to wherever the creator is. It should probably have the same protections against writing as a true name does. This allows many of the cool utility-style hexes to still work (eg. remote railguns, item retrieval, wireless charging), but only if the destination is a player, generally the one casting the hex.

Personally dislike since it removes a lot of the flexibility gates have for wild stuff using dynamic positioning.

Also posting from the discord pretty much so it isn't buried there:

• Add an "Anchor" type iota that gate can close to as a fixed position out of ambit (possibly dimension teleportation as well?)

A potential nerf to make gate less overpowered compared to GTP could also be to significantly reduce the range also? Currently GTP has a range of 2^15-1 and so does gate, perhaps reduce the range to 2^14-1 or even 2^13-1 (the first one is already a reduction of 16k blocks of range)
This or scale the gate close cost with range and entity count to match gtp at range but still be better at close range?

come up with a solution to this that I'm happy with

• Changed Gates to now have 3 types:
  • Location Anchored, which will always teleport the marked entities to the same location every time that gate is closed.
  • Entity Anchored, which will teleport the marked entities to the location of the anchor entity (plus an offset defined when the gate is created)
  • Drifting, which has the same behaviour as current gates, with an additional cost per distance gated per entity.