my website at https://anirudh.fi
blog: atproto and the ownership of identity
+63
+63
pages/blog/identity.md
+63
pages/blog/identity.md
···
1
+
---
2
+
template:
3
+
slug: identity
4
+
title: atproto and ownership of identity
5
+
subtitle: The new age of social-enabled apps
6
+
date: 2025-01-18
7
+
draft: false
8
+
---
9
+
10
+
[atproto](https://atproto.com) is very exciting to me as it's the
11
+
perfect abstraction between the identity and user data layer, and the
12
+
application layer. Compare that to the fediverse and some striking
13
+
differences become apparent.
14
+
15
+
On the fediverse, your application -- Mastodon, Pleroma, WriteFreely,
16
+
whatever -- and your user account are tied together. Your presence on say
17
+
fosstodon.org isn't the same as what you'd use on Lemmy. This is
18
+
partially due to both services implementing entirely different schemas
19
+
of the ActivityPub spec[^1], and due to how AP addressing works: so
20
+
@user@fosstodon.org is fundamentally distinct from @user@lemmy.ml.
21
+
22
+
[^1]: Or in case of the Big M, doing things mostly their own way.
23
+
24
+
atproto solves this using Personal Data Servers (PDS)[^2] and
25
+
domain-based identities. This now allows for two levels of ownership:
26
+
1. **Ownership of identity**: Use your own domain and now that's your
27
+
account across all of atproto.
28
+
2. **Ownership of data**: Run your own PDS and store all of your data
29
+
yourself.
30
+
31
+
[^2]: [atproto for distributed systems
32
+
engineers](https://atproto.com/articles/atproto-for-distsys-engineers) is recommended reading.
33
+
34
+
Thanks to this, users can re-use the same [DID](https://atproto.com/guides/identity)
35
+
across other apps built on atproto. Consequently, new social apps have
36
+
their two biggest problems solved for free:
37
+
38
+
1. The need for a new account (for users), and
39
+
2. The social graph.
40
+
41
+
This paves the wave for all kinds of new "social-enabled" services to
42
+
emerge: forums, long-form writing, and potentially even more complex
43
+
ones like code forges and more -- all sharing the same account.
44
+
45
+
Further, the separation of the app and user layers now allows for
46
+
building "apps" that are viable businesses. The app layer can be a
47
+
monetized service much like Bluesky's supposed "premium" model that's in
48
+
the works. This is a good thing -- a financially viable open network is
49
+
one that sticks around longer.
50
+
51
+
There's also signs of early VC interest in atproto.
52
+
[skyseed.fund](https://skyseed.fund/) is a fund focused solely on
53
+
backing atproto projects. I predict this is the first of many. Given
54
+
that building on atproto is so much easier than building a traditional
55
+
social app from ground up, startups here can be small and scrappy
56
+
without needing much seed capital to take off. Bluesky already having
57
+
done the hard part of acquiring its 27M strong userbase, as of this
58
+
writing, is the icing on the cake.
59
+
60
+
So yes, bottom line, I think atproto has a promising future. There's a
61
+
ton of cool stuff being built atop it already and as the network and
62
+
protocol improve, I predict a new age of social apps with user-owned
63
+
identity at its core.