@clivemoore1979.bsky.social
new build different newsletter. most ai wrappers will fail — the margin math kills them, the feature treadmill is relentless, and code generation is a commodity. the moat is the memory, not the model.
www.linkedin.com/pulse/most-a... via @LinkedIn
www.linkedin.com/pulse/most-a... via @LinkedIn
Most AI Wrappers Will Fail. Here's Why the Rest Won't.
Issue #7 | Build Different Newsletter — Part 2 of the Wrapper Economy Series Last issue I broke down how two developers built a Lovable clone in 75 minutes using Claude Code SDK. The technology works.
www.linkedin.com
February 18, 2026 at 2:10 PM
new build different newsletter. most ai wrappers will fail — the margin math kills them, the feature treadmill is relentless, and code generation is a commodity. the moat is the memory, not the model.
www.linkedin.com/pulse/most-a... via @LinkedIn
www.linkedin.com/pulse/most-a... via @LinkedIn
claude is my primary ai tool. i chose it because the ethics were the foundation, not a feature.
if standing on principle makes you a supply chain risk, what does compliance make you?
www.axios.com/2026/02/15/c...
if standing on principle makes you a supply chain risk, what does compliance make you?
www.axios.com/2026/02/15/c...
Exclusive: Pentagon threatens to cut off Anthropic in AI safeguards dispute
Anthropic has not agreed to the Pentagon's terms and defense officials are getting fed up after months of difficult negotiations.
www.axios.com
February 17, 2026 at 9:48 PM
claude is my primary ai tool. i chose it because the ethics were the foundation, not a feature.
if standing on principle makes you a supply chain risk, what does compliance make you?
www.axios.com/2026/02/15/c...
if standing on principle makes you a supply chain risk, what does compliance make you?
www.axios.com/2026/02/15/c...
that designation is normally reserved for foreign adversaries.
meanwhile openai, google, and xai all agreed to remove their safeguards for military use.
three said yes. one drew two ethical lines. that one gets punished.
meanwhile openai, google, and xai all agreed to remove their safeguards for military use.
three said yes. one drew two ethical lines. that one gets punished.
February 17, 2026 at 9:47 PM
that designation is normally reserved for foreign adversaries.
meanwhile openai, google, and xai all agreed to remove their safeguards for military use.
three said yes. one drew two ethical lines. that one gets punished.
meanwhile openai, google, and xai all agreed to remove their safeguards for military use.
three said yes. one drew two ethical lines. that one gets punished.
the only ai company that said no just got threatened for it.
anthropic told the pentagon: no mass surveillance on americans. no autonomous weapons without a human in the loop.
pentagon's response? cancel the $200M contract. label them a "supply chain risk."
anthropic told the pentagon: no mass surveillance on americans. no autonomous weapons without a human in the loop.
pentagon's response? cancel the $200M contract. label them a "supply chain risk."
February 17, 2026 at 9:47 PM
the only ai company that said no just got threatened for it.
anthropic told the pentagon: no mass surveillance on americans. no autonomous weapons without a human in the loop.
pentagon's response? cancel the $200M contract. label them a "supply chain risk."
anthropic told the pentagon: no mass surveillance on americans. no autonomous weapons without a human in the loop.
pentagon's response? cancel the $200M contract. label them a "supply chain risk."
sam altman said one person, billion dollar company. he was being conservative
the cost of execution is collapsing. the bottleneck isn't capability — it's architecture
the advantage isn't access to ai. it's knowing what to point it at
the billion dollar company lives at level 3
the cost of execution is collapsing. the bottleneck isn't capability — it's architecture
the advantage isn't access to ai. it's knowing what to point it at
the billion dollar company lives at level 3
a young boy wearing headphones is playing mission control
ALT: a young boy wearing headphones is playing mission control
media.tenor.com
February 12, 2026 at 2:46 PM
sam altman said one person, billion dollar company. he was being conservative
the cost of execution is collapsing. the bottleneck isn't capability — it's architecture
the advantage isn't access to ai. it's knowing what to point it at
the billion dollar company lives at level 3
the cost of execution is collapsing. the bottleneck isn't capability — it's architecture
the advantage isn't access to ai. it's knowing what to point it at
the billion dollar company lives at level 3
i build ai workflow tools for a living. was still stuck at level 2
designing agent-level systems — then manually copy-pasting content across six platforms like it was 2019
building at level 3. operating at level 2
the hardest part wasn't the tech. it was making myself the client
designing agent-level systems — then manually copy-pasting content across six platforms like it was 2019
building at level 3. operating at level 2
the hardest part wasn't the tech. it was making myself the client
February 9, 2026 at 10:44 PM
i build ai workflow tools for a living. was still stuck at level 2
designing agent-level systems — then manually copy-pasting content across six platforms like it was 2019
building at level 3. operating at level 2
the hardest part wasn't the tech. it was making myself the client
designing agent-level systems — then manually copy-pasting content across six platforms like it was 2019
building at level 3. operating at level 2
the hardest part wasn't the tech. it was making myself the client
Galloway launched "Resist and Unsubscribe" — a one-month economic strike. Cancel Amazon, Apple, Google, Meta, Microsoft, Netflix, OpenAI subscriptions.
These companies are a third of the S&P. Hit the revenue, move the market.
resistandunsubscribe.com
These companies are a third of the S&P. Hit the revenue, move the market.
resistandunsubscribe.com
Resist and Unsubscribe
resistandunsubscribe.com
February 5, 2026 at 7:22 PM
Galloway launched "Resist and Unsubscribe" — a one-month economic strike. Cancel Amazon, Apple, Google, Meta, Microsoft, Netflix, OpenAI subscriptions.
These companies are a third of the S&P. Hit the revenue, move the market.
resistandunsubscribe.com
These companies are a third of the S&P. Hit the revenue, move the market.
resistandunsubscribe.com
i build ai products. i'm also using them.
last week: manus or do it myself?
math said manus. reality said i'm faster and watching it anyway. claude cowork no better.
first mile = human (context, judgment, setup)
middle = ai (imperfectly)
last mile = human (oversight, decisions)
last week: manus or do it myself?
math said manus. reality said i'm faster and watching it anyway. claude cowork no better.
first mile = human (context, judgment, setup)
middle = ai (imperfectly)
last mile = human (oversight, decisions)
February 4, 2026 at 3:47 PM
i build ai products. i'm also using them.
last week: manus or do it myself?
math said manus. reality said i'm faster and watching it anyway. claude cowork no better.
first mile = human (context, judgment, setup)
middle = ai (imperfectly)
last mile = human (oversight, decisions)
last week: manus or do it myself?
math said manus. reality said i'm faster and watching it anyway. claude cowork no better.
first mile = human (context, judgment, setup)
middle = ai (imperfectly)
last mile = human (oversight, decisions)
anthropic released skills for claude. portable folders that encode your workflows, standards, expertise—once. claude loads them automatically when relevant.
simple rule: if you'd write an sop for it,
you can build a skill for it.
www.linkedin.com/in/clivemoore/
simple rule: if you'd write an sop for it,
you can build a skill for it.
www.linkedin.com/in/clivemoore/
February 3, 2026 at 4:41 PM
anthropic released skills for claude. portable folders that encode your workflows, standards, expertise—once. claude loads them automatically when relevant.
simple rule: if you'd write an sop for it,
you can build a skill for it.
www.linkedin.com/in/clivemoore/
simple rule: if you'd write an sop for it,
you can build a skill for it.
www.linkedin.com/in/clivemoore/
you don't need 10 ai tools.
level 1: chat (claude, chatgpt, grok) $0-20/mo
level 2: automation (zapier, make, n8n) $0-600/mo
level 3: agent (manus, cowork, google ws) $30-500/mo
stack right. 70% disappears.
stack wrong. overpaying.
where are you overstacked?
level 1: chat (claude, chatgpt, grok) $0-20/mo
level 2: automation (zapier, make, n8n) $0-600/mo
level 3: agent (manus, cowork, google ws) $30-500/mo
stack right. 70% disappears.
stack wrong. overpaying.
where are you overstacked?
February 2, 2026 at 9:56 PM
you don't need 10 ai tools.
level 1: chat (claude, chatgpt, grok) $0-20/mo
level 2: automation (zapier, make, n8n) $0-600/mo
level 3: agent (manus, cowork, google ws) $30-500/mo
stack right. 70% disappears.
stack wrong. overpaying.
where are you overstacked?
level 1: chat (claude, chatgpt, grok) $0-20/mo
level 2: automation (zapier, make, n8n) $0-600/mo
level 3: agent (manus, cowork, google ws) $30-500/mo
stack right. 70% disappears.
stack wrong. overpaying.
where are you overstacked?
Reposted
Donald Trump is a child rapist.
January 31, 2026 at 3:26 PM
Donald Trump is a child rapist.
Reposted
1995 Epstein files: A Dallas driver overheard Trump talking about abusing a girl.
That same witness says a victim told him Trump & Epstein raped her.
She was later found murdered—head reportedly blown off.
the horror is real. #EpsteinFiles #Accountability
That same witness says a victim told him Trump & Epstein raped her.
She was later found murdered—head reportedly blown off.
the horror is real. #EpsteinFiles #Accountability
January 31, 2026 at 8:26 PM
New research just proved AI agents hit hard computational limits.
When task complexity exceeds what an LLM can process in one pass, it hallucinates.
Not sometimes—mathematically guaranteed.
Human-in-the-loop isn’t a nice-to-have. It’s architecture.
futurism.com/artificial-i...
When task complexity exceeds what an LLM can process in one pass, it hallucinates.
Not sometimes—mathematically guaranteed.
Human-in-the-loop isn’t a nice-to-have. It’s architecture.
futurism.com/artificial-i...
AI Agents Are Mathematically Incapable of Doing Functional Work, Paper Finds
A paper claims to mathematically prove that AI agents have a hard ceiling to their capabilities that they will never surpass.
futurism.com
January 29, 2026 at 9:17 AM
New research just proved AI agents hit hard computational limits.
When task complexity exceeds what an LLM can process in one pass, it hallucinates.
Not sometimes—mathematically guaranteed.
Human-in-the-loop isn’t a nice-to-have. It’s architecture.
futurism.com/artificial-i...
When task complexity exceeds what an LLM can process in one pass, it hallucinates.
Not sometimes—mathematically guaranteed.
Human-in-the-loop isn’t a nice-to-have. It’s architecture.
futurism.com/artificial-i...
I'm 55. Most people my age are winding down. I'm doing the opposite.
Not because I'm ambitious. Because I've watched too many organizations get run into the ground by the wrong people.
Not because I'm ambitious. Because I've watched too many organizations get run into the ground by the wrong people.
January 28, 2026 at 4:26 PM
I'm 55. Most people my age are winding down. I'm doing the opposite.
Not because I'm ambitious. Because I've watched too many organizations get run into the ground by the wrong people.
Not because I'm ambitious. Because I've watched too many organizations get run into the ground by the wrong people.
"Run 10 AIs on same prompt, pick winner" = spray-and-pray.
My stack: Claude at center. ChatGPT for data segregation. Notion for mobile.
Why Claude? Anthropic was founded by researchers who left OpenAI over safety disagreements.
That's signal, not trivia.
My stack: Claude at center. ChatGPT for data segregation. Notion for mobile.
Why Claude? Anthropic was founded by researchers who left OpenAI over safety disagreements.
That's signal, not trivia.
January 26, 2026 at 5:33 PM
"Run 10 AIs on same prompt, pick winner" = spray-and-pray.
My stack: Claude at center. ChatGPT for data segregation. Notion for mobile.
Why Claude? Anthropic was founded by researchers who left OpenAI over safety disagreements.
That's signal, not trivia.
My stack: Claude at center. ChatGPT for data segregation. Notion for mobile.
Why Claude? Anthropic was founded by researchers who left OpenAI over safety disagreements.
That's signal, not trivia.
Reposted
I find it exhausting when senior devs demand a total ban on AI content. It screams insecurity. If you cannot distinguish between value and noise anymore, maybe your critical thinking skills are the real thing becoming obsolete in this industry.
January 25, 2026 at 6:02 PM
I find it exhausting when senior devs demand a total ban on AI content. It screams insecurity. If you cannot distinguish between value and noise anymore, maybe your critical thinking skills are the real thing becoming obsolete in this industry.
Reposted
“Wherever there is a human being, there is an opportunity for a kindness.”
– Seneca
#stoic #philosophy #wisdom
– Seneca
#stoic #philosophy #wisdom
January 24, 2026 at 10:00 PM
“Wherever there is a human being, there is an opportunity for a kindness.”
– Seneca
#stoic #philosophy #wisdom
– Seneca
#stoic #philosophy #wisdom
tried automating posts to 6 platforms with an ai agent yesterday
x failed (assumed i was verified). threads glitched. every platform needed manual video selection.
agent did 80%. last 20% still needed me.
if 70% can be automated, that tells you where to focus.
what's in your last mile?
x failed (assumed i was verified). threads glitched. every platform needed manual video selection.
agent did 80%. last 20% still needed me.
if 70% can be automated, that tells you where to focus.
what's in your last mile?
January 22, 2026 at 6:35 PM
tried automating posts to 6 platforms with an ai agent yesterday
x failed (assumed i was verified). threads glitched. every platform needed manual video selection.
agent did 80%. last 20% still needed me.
if 70% can be automated, that tells you where to focus.
what's in your last mile?
x failed (assumed i was verified). threads glitched. every platform needed manual video selection.
agent did 80%. last 20% still needed me.
if 70% can be automated, that tells you where to focus.
what's in your last mile?
Reposted
“We are more often frightened than hurt; and we suffer more in imagination than in reality.”
– Seneca
#stoic #philosophy #wisdom
– Seneca
#stoic #philosophy #wisdom
January 22, 2026 at 1:00 PM
“We are more often frightened than hurt; and we suffer more in imagination than in reality.”
– Seneca
#stoic #philosophy #wisdom
– Seneca
#stoic #philosophy #wisdom
Asked an AI agent to analyze 90 days of my LinkedIn posts.
3 minutes later: comprehensive engagement breakdown with strategic recommendations.
This is what amplification looks like. Not autonomy—amplification.
#AIAgents #HumanInTheLoop #BuildInPublic
3 minutes later: comprehensive engagement breakdown with strategic recommendations.
This is what amplification looks like. Not autonomy—amplification.
#AIAgents #HumanInTheLoop #BuildInPublic
January 21, 2026 at 3:33 PM
Asked an AI agent to analyze 90 days of my LinkedIn posts.
3 minutes later: comprehensive engagement breakdown with strategic recommendations.
This is what amplification looks like. Not autonomy—amplification.
#AIAgents #HumanInTheLoop #BuildInPublic
3 minutes later: comprehensive engagement breakdown with strategic recommendations.
This is what amplification looks like. Not autonomy—amplification.
#AIAgents #HumanInTheLoop #BuildInPublic
AI's superpower isn't intelligence. It's collaboration without ego.
We built something that cooperates better than we do—using our own knowledge.
The capability was always ours.
New newsletter: What AI Is Actually Teaching Us
www.linkedin.com/pulse/what-a...
We built something that cooperates better than we do—using our own knowledge.
The capability was always ours.
New newsletter: What AI Is Actually Teaching Us
www.linkedin.com/pulse/what-a...
What AI Is Actually Teaching Us
Build Different | Issue #4 Everyone's asking the wrong questions. "Will AI take my job?" "Will AI become conscious?" "Will AI destroy humanity?" Wrong questions.
www.linkedin.com
January 20, 2026 at 1:55 PM
AI's superpower isn't intelligence. It's collaboration without ego.
We built something that cooperates better than we do—using our own knowledge.
The capability was always ours.
New newsletter: What AI Is Actually Teaching Us
www.linkedin.com/pulse/what-a...
We built something that cooperates better than we do—using our own knowledge.
The capability was always ours.
New newsletter: What AI Is Actually Teaching Us
www.linkedin.com/pulse/what-a...
Human centred design and workflows always
January 19, 2026 at 6:12 PM
Human centred design and workflows always
More positive thinking or less negative thinking?
Trick question. They're two different tracks.
Adding affirmations on top of a harsh inner critic is like turning up music to drown out a failing engine.
We're 60% water. Our thoughts generate frequencies.
Self-kindness isn't soft. It's physics.
Trick question. They're two different tracks.
Adding affirmations on top of a harsh inner critic is like turning up music to drown out a failing engine.
We're 60% water. Our thoughts generate frequencies.
Self-kindness isn't soft. It's physics.
January 16, 2026 at 3:25 PM
More positive thinking or less negative thinking?
Trick question. They're two different tracks.
Adding affirmations on top of a harsh inner critic is like turning up music to drown out a failing engine.
We're 60% water. Our thoughts generate frequencies.
Self-kindness isn't soft. It's physics.
Trick question. They're two different tracks.
Adding affirmations on top of a harsh inner critic is like turning up music to drown out a failing engine.
We're 60% water. Our thoughts generate frequencies.
Self-kindness isn't soft. It's physics.
Don't know how to prompt AI?
Ask AI how to ask.
Recursive. Works every time.
Built Claude a memory system today. Three hours. Catch? Only works locally.
That's why GPT Studio connects every conversation, every thread.
Not hacks. Infrastructure.
Ask AI how to ask.
Recursive. Works every time.
Built Claude a memory system today. Three hours. Catch? Only works locally.
That's why GPT Studio connects every conversation, every thread.
Not hacks. Infrastructure.
January 15, 2026 at 5:43 PM
Don't know how to prompt AI?
Ask AI how to ask.
Recursive. Works every time.
Built Claude a memory system today. Three hours. Catch? Only works locally.
That's why GPT Studio connects every conversation, every thread.
Not hacks. Infrastructure.
Ask AI how to ask.
Recursive. Works every time.
Built Claude a memory system today. Three hours. Catch? Only works locally.
That's why GPT Studio connects every conversation, every thread.
Not hacks. Infrastructure.