-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
318 lines (306 loc) · 18.7 KB
/
Copy pathindex.html
File metadata and controls
318 lines (306 loc) · 18.7 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Collaboration Layer for Democratizing Artificial Intelligence</title>
<meta name="description" content="A lightweight collaboration layer that sits between the human and the expanding ecosystem of AI agents — and Coco, an open-source proactive co-assistant.">
<meta property="og:title" content="Collaboration Layer for Democratizing Artificial Intelligence">
<meta property="og:description" content="The next frontier is not building more agents, but building a collaboration layer between humans and the AI agents they work with.">
<meta property="og:type" content="website">
<meta property="og:image" content="assets/collaboration-layer.svg">
<link rel="icon" type="image/png" href="assets/pet.png">
<link rel="apple-touch-icon" href="assets/pet.png">
<link rel="stylesheet" href="style.css">
</head>
<body>
<!-- ==================== Full-screen Hero ==================== -->
<header class="hero" id="top">
<!-- TODO: add the hero background video, e.g.:
<video class="hero-video" autoplay loop muted playsinline poster="assets/hero-poster.jpg">
<source src="assets/hero.webm" type="video/webm">
</video>
Until then the deep-blue gradient below serves as the backdrop. -->
<div class="hero-overlay"></div>
<div class="hero-content">
<h1>Collaboration Layer for Democratizing Artificial Intelligence</h1>
<p class="hero-subtitle">
The next frontier is not building more agents, but building a collaboration layer between humans and the AI agents they work with.
</p>
<div class="hero-buttons">
<a class="btn btn-primary" href="https://github.com/collaborative-agents/coco" target="_blank" rel="noopener">
<svg viewBox="0 0 16 16" aria-hidden="true"><path d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82a7.5 7.5 0 0 1 2-.27c.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.01 8.01 0 0 0 16 8c0-4.42-3.58-8-8-8Z"/></svg>
Coco
</a>
<a class="btn btn-secondary" href="https://github.com/collaborative-agents/coco#demos" target="_blank" rel="noopener">
<svg viewBox="0 0 16 16" aria-hidden="true"><path d="M6 4.5v7l6-3.5-6-3.5Z"/></svg>
Demo
</a>
<a class="btn btn-secondary" href="#contact">
<svg viewBox="0 0 16 16" aria-hidden="true"><path d="M2 3.5h12a.5.5 0 0 1 .5.5v8a.5.5 0 0 1-.5.5H2a.5.5 0 0 1-.5-.5V4a.5.5 0 0 1 .5-.5Zm.5 1.4 5.5 4.1 5.5-4.1" stroke="currentColor" stroke-width="1.2" fill="none" stroke-linecap="round" stroke-linejoin="round"/></svg>
Contact
</a>
</div>
</div>
<a class="scroll-indicator" href="#overview" aria-label="Scroll to content">
<span>Scroll</span>
<svg viewBox="0 0 16 16" aria-hidden="true"><path d="M2.5 5.5 8 11l5.5-5.5" stroke="currentColor" stroke-width="1.8" fill="none" stroke-linecap="round" stroke-linejoin="round"/></svg>
</a>
</header>
<!-- ==================== Overview ==================== -->
<section id="overview">
<div class="container">
<h1 class="post-title">Collaboration Layer for Democratizing Artificial Intelligence</h1>
<p class="post-author">Yijia Shao · July 2026</p>
<p>
For most of computing history, the relationship between humans and machines was unambiguous:
the human decided, and the computer executed. Yet people have long envisioned computers not as
mere tools but as genuine counterparts, “serving not just secretaries, but cooperative partners
and coaches” (<a href="https://www.sciencedirect.com/science/article/pii/S1071581996900737" target="_blank" rel="noopener">Nass, Fogg & Moon, 1996</a>).
That vision, articulated thirty years ago, becomes more tangible today: recent advances in large
language models have produced systems trained on more text than any human could read in a
lifetime, and AI agents that now match or exceed human experts in certain domains.
</p>
<p>
This shift in what machines <em>can do</em> demands a corresponding shift in how we design systems
for humans and AI to <em>work together</em>. Progress to date has largely benefited those who were
already capable - strong engineers, experienced researchers, and the few who know how to get the
most out of these tools. The deeper aspiration is to extend that leverage to everyone: not just to
make experts faster, but to help more people achieve what they did not know they could attempt.
</p>
</div>
</section>
<!-- ==================== Pain Points ==================== -->
<section id="pain-points">
<div class="container">
<h2>Pain Points With AI Agents Today</h2>
<p>
Despite rapid advances in model capability, working effectively with AI agents remains harder
than it should be. Over the last two years, we have observed the following recurring pain points
in our own research and beyond.
</p>
<ul class="pain-list">
<li>
<strong>AI support is siloed within apps, blind to the broader context of a user's work.</strong>
Today's AI agents are largely app-bound, confined to a browser page, a mailbox, or a specific
folder. As tasks grow more complex and span multiple tools, the user is left manually bridging
the context gap: re-explaining background, reconstructing history, and stitching together
outputs that no single agent ever sees in full
(<a href="https://arxiv.org/abs/2505.10831" target="_blank" rel="noopener">Shaikh et al., 2025</a>).
</li>
<li>
<strong>Collaboration outcomes depend heavily on the human, not just the agent.</strong>
In a study evaluating human-agent collaboration across various real-world occupations, when we
decouple human and agent CollabSkill, we find that variance across humans is significantly
larger than variance across agents. Users in the top 25% of CollabSkill see human-agent
collaboration outperform solo agents in 74% of cases; for those in the bottom 25%, the win
rate is only 27%
(<a href="https://arxiv.org/abs/2606.09833" target="_blank" rel="noopener">Shao et al., 2026</a>).
</li>
<li>
<strong>Many people simply don't know what AI can help them with.</strong>
When Open Claw went viral earlier this year, the
<a href="https://reddit.com/r/LocalLLM/comments/1qp0jhl/i_used_clawdbot_now_moltbot_and_here_are_some/" target="_blank" rel="noopener">top comment</a>
on a community thread read plainly: <em>“My biggest issue is idk what to really use it
for.”</em> Our own observations echo this. In May 2026, we ran a
<a href="https://www.youtube.com/live/klL5ev8OPzk?si=9K94aJ8dFooPaZDa" target="_blank" rel="noopener">live stream</a>
teaching people how to use AI agents and offered Claude Pro subscriptions to participants who
stayed. When we followed up a month later, a recurring theme was that people had used AI far
less than they expected.
</li>
</ul>
</div>
</section>
<!-- ==================== Collaboration Layer ==================== -->
<section id="collaboration-layer">
<div class="container">
<h2>A Lightweight Collaboration Layer</h2>
<p>
What would it take to bridge that gap systematically? We argue that the next frontier is not
building more agents, but building a <em>collaboration layer</em> — a lightweight coordination
substrate that sits between the human and the expanding ecosystem of AI agents they might work with.
</p>
</div>
<div class="container-wide">
<figure class="borderless">
<img src="assets/collaboration-layer.svg" alt="Diagram of the collaboration layer sitting between the user's workspace and the ecosystem of AI agents">
<figcaption>
The collaboration layer bridges the user's workspace and existing AI tools, democratizing AI to empower people across domains and skill levels through proactive support, context understanding, smart orchestration, and managing memory and preferences.
</figcaption>
</figure>
</div>
<div class="container">
<p>
By sitting between the user's workspace and existing AI tools, the collaboration layer ingests
signals from a broader stream of the user's work and activity over time. From these signals, it
builds a unified understanding of what the user is trying to achieve, and that understanding
drives four core capabilities: <em>proactive support</em>, which anticipates needs and surfaces
suggestions before the user has to ask; <em>context understanding</em>, which unifies signals
across apps, tasks, and time into a coherent picture of intent; <em>smart orchestration</em>,
which selects the right downstream agent, scopes the work appropriately, or nudges the user to
work on certain parts themselves; and <em>memory and preferences</em>, which maintains a
persistent, personalized model of how a particular user works. The collaboration layer is deliberately decoupled from heavy execution workload to ensure it can
be adaptive, lightweight, and private.
</p>
</div>
</section>
<!-- ==================== Coco ==================== -->
<section id="coco">
<div class="container">
<h2>Coco: A First Step</h2>
<p>
As a first step toward this vision, we introduce
<a href="https://github.com/collaborative-agents/coco" target="_blank" rel="noopener"><strong>Coco</strong></a>
(Proactive Co-Assistant Through Continuous Context
Observation), an open-source desktop application that focuses on the proactive
support component of the collaboration layer. Coco quietly observes the user's computer-use
context — browsing, reading, and editing activity across applications — and steps in with targeted
assistance at the right moment: a ready-to-send prompt when it detects repetitive work, a
synthesized summary inferred from browsing intent, or concrete questions to help a user navigate
unfamiliar territory.
</p>
<p>
It runs fully on the user's machine with no telemetry or cloud storage of personal data,
demonstrating that meaningful proactive assistance does not require sacrificing privacy. Coco does
not execute tasks itself, but bridges the gap between a user's unspoken needs and the agents best
positioned to address them.
</p>
</div>
<div class="container-wide">
<figure class="borderless">
<img src="assets/coco-overview.svg" alt="Overview of Coco: continuous context observation driving proactive assistance">
<figcaption>
Coco implements the Computer Use Behavior Observation Protocol, which specifies how behavioral signals from the user's desktop are captured, classified, and forwarded to downstream agents.
It uses a VLM to decide when to offer proactive suggestions.
</figcaption>
</figure>
</div>
</section>
<!-- ==================== Scaffolding ==================== -->
<section id="scaffolding">
<div class="container">
<h2>Scaffolding Humans to Move Up the Stack</h2>
<p>
As AI agents grow more capable, a narrow focus on existing tasks and efficiency risks making
humans feel increasingly peripheral. But if we believe AI is as transformative as the invention of
the computer, the right measure of success is not only how much work agents can do for people, but
also how much more people can do because of them. Agents with broad knowledge and general
reasoning ability are uniquely positioned to help people venture beyond their current
competencies: to attempt problems they have never framed, acquire skills they never imagined, and
explore fields adjacent to their own. Realizing this potential, however, requires scaffolding that
actively draws people into new territory rather than simply completing tasks on their behalf.
</p>
<p>
Every person operates with a horizon of familiar domains and a much larger territory of unknown
unknowns. In our earlier work on
<a href="https://github.com/stanford-oval/storm" target="_blank" rel="noopener">building deep research agents</a>,
we found that one promising direction for scaffolding users into the unknown is to proactively
engage them in the agent's work rather than simply presenting search outcomes
(<a href="https://arxiv.org/abs/2408.15232" target="_blank" rel="noopener">Jiang et al., 2024</a>).
The collaboration layer extends this insight into everyday computer use. For instance, when Coco
observes that a user is working on an interdisciplinary project at the intersection of AI and
hardware but appears to have limited background in the hardware side, it proactively surfaces key
concepts and questions worth exploring.
</p>
</div>
<div class="container-wide">
<figure>
<img src="assets/teaser_searching_unknown_unknown.webp" alt="Coco proactively surfacing key concepts and questions worth exploring during interdisciplinary work">
<figcaption>
The user wants to run Nemotron-3-Nano-Omni locally but doesn't know what hardware specs matter. Coco senses the struggle and suggests concrete questions that are at the core of the space.
</figcaption>
</figure>
</div>
<div class="container">
<p>
This also reconnects with an early vision of agents that reduce work and information overload
(<a href="https://dl.acm.org/doi/pdf/10.1145/176789.176792" target="_blank" rel="noopener">Maes, 1994</a>).
Ironically, as our pain points illustrate, today's AI agents can exacerbate information overload,
particularly for non-technical users. The collaboration layer addresses this missing piece of
infrastructure. It works toward an AI future in which the benefits of capable AI systems are not
gated on technical sophistication or any particular agent provider — one in which the relationship
between humans and their AI tools is genuinely collaborative rather than merely convenient.
</p>
</div>
</section>
<!-- ==================== Limitations ==================== -->
<section id="limitations">
<div class="container">
<h2>Limitations and What We Are Working On</h2>
<p>
In this piece, we introduce a vision for a collaboration layer that sits between the user's
workspace and the broader ecosystem of AI tools, with the goal of democratizing AI to empower
people across domains and skill levels. We have also open-sourced Coco, a first instantiation of
that vision focused on proactive co-assistance. The initial release carries several limitations
that we are actively working to address.
</p>
<ol class="pain-list">
<li>
<strong>On-device inference.</strong>
While Coco runs fully on the user's computer, the observation component currently makes calls
to a vision-language model, adding latency, a dependency on network connectivity, and ongoing
API costs. We are working with hardware experts to make Coco on-device native.
</li>
<li>
<strong>Personalization and relevance grounding.</strong>
Proactive assistance is only valuable when it surfaces the right thing at the right moment.
Defining ground truth for what is “useful” versus “noise” is inherently
subjective (<a href="https://erichorvitz.com/chi99horvitz.pdf" target="_blank" rel="noopener">Horvitz, 1999</a>).
We are developing personalization mechanisms that learn from user feedback without
requiring explicit annotation.
</li>
<li>
<strong>Intelligent orchestration.</strong>
The current delegation logic is based on parametric knowledge in the language model and does not
account for the actual capabilities of available agents or the preferences of a specific user.
We are exploring intelligent orchestration approaches.
</li>
</ol>
</div>
</section>
<!-- ==================== Citation ==================== -->
<section id="citation">
<div class="container">
<h2>Citation</h2>
<p>If you find this work useful, please consider citing:</p>
<div class="bibtex" id="bibtex-block">@misc{shao2026collaborationlayer,
title = {Collaboration Layer for Democratizing Artificial Intelligence},
author = {Shao, Yijia},
year = {2026},
url = {https://collaborative-agents.github.io}
}<button class="copy-btn" onclick="copyBibtex(this)">Copy</button></div>
</div>
</section>
<!-- ==================== Contact ==================== -->
<section id="contact">
<div class="container">
<h2>Contact</h2>
<p>
We would love to hear your thoughts, feedback, and ideas. Reach out at
<a href="mailto:shaoyj@stanford.edu">shaoyj@stanford.edu</a>, or open an issue on the
<a href="https://github.com/collaborative-agents/coco" target="_blank" rel="noopener">Coco GitHub repository</a>.
</p>
</div>
</section>
<!-- ==================== Footer ==================== -->
<footer>
<div class="ack">
<div class="ack-title">Acknowledgements</div>
We thank Yihan Wang, Haowen Wang, Boom Iamphongsai, Zora Wang, Omar Shaikh, Yanzhe Zhang, Caleb Ziems,
Song Bian, Huizi Mao, Junxian Guo, Zhen Dong, Jindong Jiang, Shang Yang, Song Han, and Diyi Yang for their support and
feedback at various stages of this post.
</div>
<div>© 2026 Collaborative Agents · <a href="https://github.com/collaborative-agents/coco" target="_blank" rel="noopener">GitHub</a></div>
</footer>
<script>
function copyBibtex(btn) {
const block = document.getElementById('bibtex-block');
const text = block.childNodes[0].textContent.trim();
navigator.clipboard.writeText(text).then(() => {
btn.textContent = 'Copied!';
setTimeout(() => { btn.textContent = 'Copy'; }, 1500);
});
}
</script>
</body>
</html>