Thread by @BobBodily

The Internet Computer Protocol $ICP continues to improve.

@lastmjs and I frequently discuss building on the #IC: problems, pain points, bottlenecks

In our past few conversations, we've been running out of problems to talk about.

Here's what I'm most excited about for 2023: 🧵

Increasing the binary size limit.

Right now all of your canister code has to fit in 2MB. This is pretty low, especially if you are doing canister dev in Python.

They are working on increasing it which will enable standard library in Python, external dependencies, etc.

Decentralizing node providers and boundary nodes.

The sooner anyone can buy a rig and become a node provider the better. This is a common reason why people complain about $ICP centralization.

This is currently being worked on, and I can't wait to forever remove this concern.

Hardening and further securing the NNS.

Let's make it very hard to change canister code, censor canisters, and change protocol tokenomics.

This will increase trust in the protocol and give us safeguards while governing the protocol.

Cross-canister query calls.

All cross-canister calls right now (one canister calling another canister directly) are update calls. Being able to do cross-canister queries will be absolutely amazing and will enable SO many new things. Will be within same subnet first.

Improvements to tECDSA speed.

At 1 txn/s/subnet we can't really handle any kind of meaningful activity at scale. I think Dfinity can get it up to 10 txns/s/subnet, at which point combining it with some roll-up or txn bundling will enable tECDSA at scale I believe.

Image caching at boundary nodes

I would love to be able to cache assets in a granular and configurable way at the boundary node layer. Would be extremely convenient to not have to worry about caching as an application developer.

Unsure on progress, but hope it comes in 2023.

Threshold Key Derivation

This will add incredible privacy to the IC by enabling canisters to perform threshold key derivation. This means even node providers cannot access your canister data.

Once implemented, the #IC could host the most secure and private data in the world!

Increased message size or better tooling

Right now all messages are limited to 2MB, which is limiting in many cases.

Having a general solution at the protocol level to handle messages larger than 2MB would be amazing in abstracting away chunking/streaming from #IC app devs.

Modular robust DAO tooling.

With SNS, @CigDao_, @DSCVR1, we have a great start to DAO tooling. I expect this to increase substantially in 2023.

$ICP has a major advantage over other chains in being able to control a full stack app with a DAO. Will 2023 be the year of the DAO?

Now for the crazy wish-it-could-happen item: sending $ICP in an update call.

If you could call a canister ".with_icp(10)" it would be the simplest payment flow of all time. This would dramatically simplify payment flows on $ICP & enable app devs to do way more with ICP payments

I have been building canisters in Kybra (using Python) over the past month, and I am impressed at how capable the Internet Computer is right now.

I can basically do whatever I want in Python on the #IC, which is absolutely insane.

And it's just going to get better.

Thanks to everyone at @dfinity for continuing to make incredible progress at the protocol layer. 2023 is going to be absolutely massive for #ICP 💥