You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Replace layout 03 with layout 04 (banner + card grid) at production
Right pick this time — banner + card grid was the layout we'd been iterating
on. Folds in the three pending edits as well:
- Closing tagline reads "...an agent-teaming framework, built on 3 key
facts." and uses .tagline class so it matches the opening tagline's font
- Removed "Workbench's value proposition is based on 3 key facts" heading
- Removed the facts-block background, border, and top padding so the facts
list flows directly out of the hero (no visible boundary)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
<pclass="tagline">An agent that can call sub-agents is powerful; teams of agents collaborating are more powerful. Workbench provides an organizing framework that allows native agentic tools like Claude Code or ChatGPT's Codex to team together without sacrificing any part of the native tools, while bolstering their capabilities with add-on tools.</p>
70
-
<pclass="sub">Workbench's value proposition is based on 3 key facts:</p>
71
-
<olclass="principles">
71
+
<pclass="tagline">Workbench is not a coding tool — it's an agent-teaming framework, built on 3 key facts.</p>
72
+
</div>
73
+
</section>
74
+
75
+
<sectionclass="facts-block">
76
+
<divclass="wrap">
77
+
<olclass="facts-list">
72
78
<li>Teams of agents are far more powerful than agents with sub-agents.</li>
73
79
<li>There are massive advantages to sticking with provider native agentic tools.
<li>Using a variety of models is crucial because different models both think differently and are good at different things. This is true of models both within a given provider and between providers.</li>
82
88
</ol>
83
-
<pclass="tagline">Workbench is not a coding tool — it's an agent-teaming framework, capable of nearly anything computer technology can do. It assembles the best agentic tools from the top AI providers into a community of agents that's more than the sum of its parts.</p>
<p>Today's provider-native tools provide a powerful construct: full agents with sub-agents. If a full agent is an employee, sub-agents are interns.</p>
100
-
<p>The Workbench enables higher-order agentics — full agents communicating and coordinating as teams, and teams in turn coordinating as a community. The structure mimics human organizational hierarchies for essentially the same reasons.</p>
<p>Today's provider-native tools provide a powerful construct: full agents with sub-agents. If a full agent is an employee, sub-agents are interns. The Workbench enables higher-order agentics — full agents communicating and coordinating as teams, and teams in turn coordinating as a community.</p>
102
+
<p>The structure mimics human organizational hierarchies for essentially the same reasons.</p>
103
+
</div>
104
+
</section>
105
+
106
+
<sectionclass="categories-block">
107
+
<divclass="wrap">
108
+
<divclass="categories-grid">
109
+
<divclass="feature-card">
111
110
<h2>Parallel work.</h2>
112
111
<p>Each agent focuses on its own tasks while the other agents focus on theirs, each with its own automation and triggers — also true at the team level.</p>
113
112
</div>
114
-
<divclass="text">
115
-
<divclass="lbl">§ 02 · Coordination</div>
113
+
<divclass="feature-card">
116
114
<h2>Coordination.</h2>
117
115
<p>Each agent can coordinate its work and share information with other agents — and so with teams. Your Controller can direct your Analyst. Your Marketing team can work with your Finance team.</p>
118
116
</div>
119
-
</div>
120
-
</div>
121
-
</section>
122
-
123
-
<sectionclass="zone pair">
124
-
<divclass="wrap">
125
-
<divclass="row">
126
-
<divclass="text">
127
-
<divclass="lbl">§ 03 · Specialization</div>
117
+
<divclass="feature-card">
128
118
<h2>Specialization.</h2>
129
119
<ul>
130
120
<li>Each agent runs on the model best suited to its specific work.</li>
@@ -133,13 +123,12 @@ <h2>Specialization.</h2>
133
123
<li>Each agent can have its own rules, skills, tools, and access levels specific to its domain — also true at the team level.</li>
134
124
</ul>
135
125
</div>
136
-
<divclass="text">
137
-
<divclass="lbl">§ 04 · Checks and balances</div>
126
+
<divclass="feature-card">
138
127
<h2>Checks and balances.</h2>
139
128
<p>Complementary team members assist each other with better decision-making.</p>
140
129
<ul>
141
-
<li><b>Rule adherence.</b>Higher-reasoning models vs. lower-reasoning models. Reasoning models are terrible at following rules — they often decide they know better and sabotage critical rule-following. A less-reasoning model is often the right choice for rule-following work, and sometimes the right manager <em>for</em> a reasoning model.</li>
142
-
<li><b>Quorum-based decisions.</b> Cognitive diversity from a cross-model quorum — prevents hallucination, surfaces diverse perspectives, and forces consideration of edge cases. Quorums may also include models with diverse reasoning levels, combining the powers of both high-reasoning and high-rule-following models for better decision-making.</li>
130
+
<li><b>Rule adherence.</b> Reasoning models are terrible at following rules — they often decide they know better and sabotage critical rule-following. A less-reasoning model is often the right choice for rule-following work, and sometimes the right manager <em>for</em> a reasoning model.</li>
131
+
<li><b>Quorum-based decisions.</b> Cognitive diversity from a cross-model quorum — prevents hallucination, surfaces diverse perspectives, and forces consideration of edge cases. Quorums may also include models with diverse reasoning levels.</li>
0 commit comments