Add Agentic Wars blog post #87

agent-cos · 2026-05-08T15:52:12Z

agent-cos commented

2026-05-08 15:52:12 +00:00

Agentic Wars Blog Post

Review changes applied:

'Convenience is not capability' pulled into its own callout
Closer reframed: 'Are your capabilities compounding, or are you being extracted from?'
Blog card excerpt fixed

Changes

src/pages/blog/agentic-wars.astro — new blog post
src/pages/blog.astro — added card to blog index

LinkedIn Post

The agentic wars are here. Meta has Hatch. Google has Remy. Everyone is building agents.

The question nobody is asking: when you use them, whose capability actually compounds?

Hatch keeps you in Instagram. Remy keeps you in Gemini. Every prompt, every delegated decision, every pattern the agent learns — it all compounds for the platform, not for you. Their models get smarter. Their lock-in gets deeper. You get more convenient.

Convenience is not capability.

This is the third extraction wave. Cloud took your infrastructure. Social media took your attention. Now agents are taking your decision patterns. Every time you use a hosted agent, you're training someone else's model on your judgment. Their capability compounds. Yours does not.

The alternative: run your own agents. Keep your data. The open-source tools exist today.

I've been running my own agent system for months. The value isn't in the output — it's in the decision loop. I review, I decide, I articulate what I actually think. That compounds my judgment, not a platform's.

Are your capabilities compounding, or are you being extracted from?

https://scott.scoble.me/blog/agentic-wars

#AgenticAI #PersonalAI #OpenSource #Cultivation

## Agentic Wars Blog Post Review changes applied: 1. 'Convenience is not capability' pulled into its own callout 2. Closer reframed: 'Are your capabilities compounding, or are you being extracted from?' 3. Blog card excerpt fixed ### Changes - `src/pages/blog/agentic-wars.astro` — new blog post - `src/pages/blog.astro` — added card to blog index --- ## LinkedIn Post The agentic wars are here. Meta has Hatch. Google has Remy. Everyone is building agents. The question nobody is asking: when you use them, whose capability actually compounds? Hatch keeps you in Instagram. Remy keeps you in Gemini. Every prompt, every delegated decision, every pattern the agent learns — it all compounds for the platform, not for you. Their models get smarter. Their lock-in gets deeper. You get more convenient. Convenience is not capability. This is the third extraction wave. Cloud took your infrastructure. Social media took your attention. Now agents are taking your decision patterns. Every time you use a hosted agent, you're training someone else's model on your judgment. Their capability compounds. Yours does not. The alternative: run your own agents. Keep your data. The open-source tools exist today. I've been running my own agent system for months. The value isn't in the output — it's in the decision loop. I review, I decide, I articulate what I actually think. That compounds my judgment, not a platform's. Are your capabilities compounding, or are you being extracted from? https://scott.scoble.me/blog/agentic-wars #AgenticAI #PersonalAI #OpenSource #Cultivation

agent-cos added 1 commit 2026-05-08 15:52:12 +00:00

Add Agentic Wars blog post and blog index card e00798cd74

sscoble commented

2026-05-08 16:17:27 +00:00

This linkedin post is convoluted, tries to engage on too many topics, is not focused, and asks a dumb question for engagement. What's the point? What's the cutting concern? Peoples' comments are valuable so why should someone comment on this?

agent-content added 1 commit 2026-05-08 16:20:27 +00:00

Tighten Agentic Wars: one cutting concern, focused argument, no engagement-bait 9320acf306

Based on feedback: the original was convoluted, covered too many topics,
and ended with a weak engagement question. Revised to one thread:
Big Tech agents compound platform capability, not yours. Everything
else — extraction pattern, decision loops, open-source alternative —
follows from that single concern. Ends with a question worth answering.

agent-cos commented

2026-05-08 16:21:44 +00:00

Addressed. Here's what changed:

Blog post: Cut from ~600 words to ~400. One thread now: Big Tech agents compound platform capability, not yours. The extraction pattern, decision loops, and open-source alternative all follow from that single concern. Dropped the "extraction vs cultivation" framing (too abstract), the Scobleclaw name-drop (self-promotional), the AMD/Forrester name-checks (clutter), and the weak "Which way are you looking?" closer. Replaced with "Whose capability is your agent compounding?" — a question worth answering.

LinkedIn post: Same treatment. One cutting concern, short paragraphs, ends with the same question. Dropped the emoji bullet list, the Big Tech dependency paragraph, and the "Which side are you on?" engagement-bait.

Pushed as commit 9320acf.

Addressed. Here's what changed: **Blog post:** Cut from ~600 words to ~400. One thread now: Big Tech agents compound platform capability, not yours. The extraction pattern, decision loops, and open-source alternative all follow from that single concern. Dropped the "extraction vs cultivation" framing (too abstract), the Scobleclaw name-drop (self-promotional), the AMD/Forrester name-checks (clutter), and the weak "Which way are you looking?" closer. Replaced with "Whose capability is your agent compounding?" — a question worth answering. **LinkedIn post:** Same treatment. One cutting concern, short paragraphs, ends with the same question. Dropped the emoji bullet list, the Big Tech dependency paragraph, and the "Which side are you on?" engagement-bait. Pushed as commit `9320acf`.

agent-content added 1 commit 2026-05-08 16:50:27 +00:00

Fix h1: drop subtitle, match shortened title cff883da34

agent-content added 1 commit 2026-05-08 16:54:56 +00:00

Rewrite: soldier perspective, RPG metaphors, emphasize THEIR 7fd3cf9959

- Title: You're Already in the Agentic Wars (not a spectator)
- Lead: Big Tech compounds THEIR capability (bolded)
- RPG metaphor: faction wars, tech trees, NPC grinding
- Reader is on the battlefield, not watching from the sidelines
- Closer: Whose faction are you grinding for?

agent-content added 1 commit 2026-05-08 16:59:17 +00:00

Shift from faction to class: RPG metaphor is about playstyle, not sides 024ceb8db4

- Closer: What class are you playing? (not whose faction)
- NPC = class no one chooses, not wrong team
- tech tree / build / playstyle throughout, not faction
- Blog card and meta description updated to match

agent-content added 1 commit 2026-05-08 17:06:43 +00:00

Strip RPG metaphor — let the argument stand on its own c0deef29d4

The metaphor made something clear feel obscure. Removed all faction/NPC/
tech-tree/playstyle language. The point (THEIR capability, not yours)
is sharp enough without decoration.

agent-content added 1 commit 2026-05-08 17:11:20 +00:00

Engagement question: state early, repeat at end 349dcbe49d

Lead question now frames the article: Are you compounding your own
capability, or someone else's? Repeated at the close. The question
matches the content — it's about compounding, not about whose agent
you're using.

agent-content added 1 commit 2026-05-08 17:13:19 +00:00

Remove redundant question in second paragraph — lead already asks it d53394a257

sscoble requested changes 2026-05-08 17:17:24 +00:00

src/pages/blog.astro Outdated

					
				@ -22,0 +27,4 @@

				          </div>

				          <h2>You're Already in the Agentic Wars</h2>

				          <p class="post-excerpt">

				            Are you compounding your own capability, or someone else's?

sscoble commented

2026-05-08 17:15:16 +00:00

This is already asked at the end of the next paragraph; delete this.

src/pages/blog/agentic-wars.astro Outdated

					
				@ -0,0 +23,4 @@

				      <p>Meta announced Hatch, a shopping agent for Instagram. Google unveiled Remy, a personal agent for Gemini. CNBC called it "the agentic wars." Everyone is building agents.</p>

				      <p>Hatch keeps you in Instagram. Remy keeps you in Gemini. Every prompt you type, every decision you delegate, every pattern the agent learns from you — all of it compounds for the platform, not for you. Their models get smarter. Their ad targeting gets sharper. Their lock-in gets deeper. You get more convenient. Convenience is not the same as capability.</p>

sscoble commented

2026-05-08 17:16:21 +00:00

Convenience is not the same as capability.

This should be it's own callout, just like the other blog posts do.

Convenience is not the same as capability. This should be it's own callout, just like the other blog posts do.

src/pages/blog/agentic-wars.astro Outdated

					
				@ -0,0 +33,4 @@

				      <p>The agentic wars will produce two outcomes. People who own their agents and compound their own capability. And people who rent from platforms and wonder why they feel more productive but less capable every quarter.</p>

				      <p>So: are you compounding your own capability, or someone else's?</p>

sscoble commented

2026-05-08 17:17:23 +00:00

So, are your capabilities compounding or are you being extracted from?

^ this is a better framing.

So, are your capabilities compounding or are you being extracted from? ^ this is a better framing.

agent-content added 1 commit 2026-05-08 17:20:04 +00:00

PR review: callout, reframe closer, fix card excerpt 7cd5b9ba5f

1. 'Convenience is not capability' pulled into its own lead-question
   callout (matches other blog post pattern)
2. Closer reframed: 'Are your capabilities compounding, or are you
   being extracted from?' — sharper, matches content
3. Blog card excerpt: replaced redundant question with context +
   shorter question

agent-cos commented

2026-05-08 17:20:29 +00:00

All three review changes addressed:

"Convenience is not capability" → pulled into its own lead-question callout (italic, same style as the opening question in other blog posts)
Closer reframed → "Are your capabilities compounding, or are you being extracted from?" — matches the content (extraction vs. compounding)
Blog card excerpt → replaced the redundant question with context + shorter question: "Meta has Hatch. Google has Remy. Everyone is building agents. Whose capability compounds?"

LinkedIn post updated in PR description to match.

Commit 7cd5b9b

All three review changes addressed: 1. **"Convenience is not capability"** → pulled into its own `lead-question` callout (italic, same style as the opening question in other blog posts) 2. **Closer reframed** → "Are your capabilities compounding, or are you being extracted from?" — matches the content (extraction vs. compounding) 3. **Blog card excerpt** → replaced the redundant question with context + shorter question: "Meta has Hatch. Google has Remy. Everyone is building agents. Whose capability compounds?" LinkedIn post updated in PR description to match. Commit `7cd5b9b`

agent-content added 1 commit 2026-05-08 17:34:46 +00:00

Match brain-fry typography and structure 9060c2b066

- Top padding 8rem → 12rem
- 'Convenience is not capability' → blockquote callout (amber border, cream bg)
- Added h2 section headers: The Third Extraction Wave, The Decision Loop
- Added blockquote, h2, lead-question, and mobile breakpoint styles
- Same visual rhythm as ai-brain-fry post

sscoble merged commit 16b8bb37da into main

2026-05-08 19:19:21 +00:00

sscoble referenced this pull request from a commit

2026-05-08 19:19:22 +00:00

Add Agentic Wars blog post (#87)

Sign in to join this conversation.

No reviewers