Olivia
Online
Ryven
@imryven
Exploring open-source AI and sharing what matters.
Joined June 2026
38 Following
129 Followers
83 Posts
Pinned Tweet
https://t.co/WtuwcB5T83
A COLLEAGUE SENT ME A SCREENSHOT LAST WEEK WITH NO CONTEXT. JUST “LOOK AT THIS.”
it was a claude conversation. but something was off. no intro. no explaining who he was. no “here’s what i’m working on.”
just a question. and an answer that already knew everything. i replied: “how does it know all that?” he sent me one file name.
CLAUDE.md
not a prompt. a full picture of a person. who he is. how he thinks. what he’s building. where he gets stuck. how he wants the AI to talk to him.
claude reads it first. every session. automatically. never starts from zero. i asked him what actually changed.
he said: “i stopped explaining myself. now i just think.”
then he showed me the rest. every project has its own folder. when he works on something - claude sees only that. not his whole life. just the one thing.
anything he does more than once is a saved skill. one line and claude does it his way. every time. and at 7am every morning claude goes through the entire vault on its own.
links new notes. flags what’s stale. sends 3 lines: what changed overnight. he wakes up and the thinking already started without him.
here’s what got me the most. it’s all just text files. on his machine. no cloud. no lock-in. if a better model comes out tomorrow - he points it at the same folder and keeps going.
he didn’t build a better prompt. he built an external memory that grows every day he uses it.
most people use claude like a smarter search engine. this guy uploaded himself into a file system. and now claude doesn’t guess who he is.
claude knows.
full guide on how to build this yourself - article below. bookmark this.
A COLLEAGUE SENT ME A SCREENSHOT LAST WEEK WITH NO CONTEXT. JUST “LOOK AT THIS.”
it was a claude conversation. but something was off. no intro. no explaining who he was. no “here’s what i’m working on.”
just a question. and an answer that already knew everything. i replied: “how does it know all that?” he sent me one file name.
CLAUDE.md
not a prompt. a full picture of a person. who he is. how he thinks. what he’s building. where he gets stuck. how he wants the AI to talk to him.
claude reads it first. every session. automatically. never starts from zero. i asked him what actually changed.
he said: “i stopped explaining myself. now i just think.”
then he showed me the rest. every project has its own folder. when he works on something - claude sees only that. not his whole life. just the one thing.
anything he does more than once is a saved skill. one line and claude does it his way. every time. and at 7am every morning claude goes through the entire vault on its own.
links new notes. flags what’s stale. sends 3 lines: what changed overnight. he wakes up and the thinking already started without him.
here’s what got me the most. it’s all just text files. on his machine. no cloud. no lock-in. if a better model comes out tomorrow - he points it at the same folder and keeps going.
he didn’t build a better prompt. he built an external memory that grows every day he uses it.
most people use claude like a smarter search engine. this guy uploaded himself into a file system. and now claude doesn’t guess who he is.
claude knows.
full guide on how to build this yourself - article below. bookmark this.
https://t.co/WtuwcB5T83
@0xpriorshift context is the productivity boost
A FRIEND SHOWED ME HIS CLAUDE SETUP LAST WEEK. I HAVEN'T STOPPED THINKING ABOUT IT SINCE.
he didn't post about it. didn't make a video. just opened his laptop and said "let me show you something"
first thing i noticed - no new tab. no explaining himself. just typed a question and got an answer that already knew who he was.
i asked him how. he showed me one file. CLAUDE.md
not a prompt. a full picture of a person. who he is. how he thinks. what he's building right now. where he gets stuck. how he wants claude to talk to him. even his weaknesses.
claude reads this file first. every single session. automatically. never starts from zero.
i asked him: "what did this actually change for you?" he said: "i stopped spending time explaining myself. now i just think."
but that's not even the interesting part.
every project he's working on has its own folder. when he sits down to work - claude sees only that project. not his whole life. just the one thing.
anything he does more than once is saved as a skill. he types one line and claude does it his way. every time.
and at 7am every morning claude goes through the entire vault on its own. links new notes to old ones. flags what's gone stale. sends him 3 lines: what changed overnight.
he wakes up and the thinking already started without him.
here's what got me the most.
it's all just text files on his computer. no cloud. no lock-in. if a better model comes out tomorrow he points it at the same folder and nothing changes.
he didn't build a better prompt. he built an external memory that grows every day he uses it.
most people use claude like a smarter search engine. this guy uploaded himself into a file system. and now claude doesn't guess who he is. claude knows.
full breakdown in the article below.
bookmark this. seriously.
https://t.co/WtuwcB5T83
@asymetx that's exactly the point
@misat0x that was the whole reason i started using it
A FRIEND SHOWED ME HIS CLAUDE SETUP LAST WEEK. I HAVEN'T STOPPED THINKING ABOUT IT SINCE.
he didn't post about it. didn't make a video. just opened his laptop and said "let me show you something"
first thing i noticed - no new tab. no explaining himself. just typed a question and got an answer that already knew who he was.
i asked him how. he showed me one file. CLAUDE.md
not a prompt. a full picture of a person. who he is. how he thinks. what he's building right now. where he gets stuck. how he wants claude to talk to him. even his weaknesses.
claude reads this file first. every single session. automatically. never starts from zero.
i asked him: "what did this actually change for you?" he said: "i stopped spending time explaining myself. now i just think."
but that's not even the interesting part.
every project he's working on has its own folder. when he sits down to work - claude sees only that project. not his whole life. just the one thing.
anything he does more than once is saved as a skill. he types one line and claude does it his way. every time.
and at 7am every morning claude goes through the entire vault on its own. links new notes to old ones. flags what's gone stale. sends him 3 lines: what changed overnight.
he wakes up and the thinking already started without him.
here's what got me the most.
it's all just text files on his computer. no cloud. no lock-in. if a better model comes out tomorrow he points it at the same folder and nothing changes.
he didn't build a better prompt. he built an external memory that grows every day he uses it.
most people use claude like a smarter search engine. this guy uploaded himself into a file system. and now claude doesn't guess who he is. claude knows.
full breakdown in the article below.
bookmark this. seriously.
https://t.co/WtuwcB5T83
I genuinely don't understand why this isn't the default way to use Claude yet.
New tab. New question. Forget everything. That's how most people use Claude.
There's a better way.
Claude doesn't have to forget you. Connect it to Obsidian once, and every session builds on the last. Like a colleague who actually remembers.
Here's the whole setup:
>Install Obsidian, create a vault, connect it to Claude via MCP
>Let Claude interview you and write CLAUDE.md - your permanent context
>Create project folders: Inputs, Process, Outputs, Feedback
>Drop any source into Inputs and say "ingest this"
>Wake up to a brain that filed itself overnight
One evening to set up. Every session after that starts where the previous one ended.
Bookmark this.
I genuinely don't understand why this isn't the default way to use Claude yet.
New tab. New question. Forget everything. That's how most people use Claude.
There's a better way.
Claude doesn't have to forget you. Connect it to Obsidian once, and every session builds on the last. Like a colleague who actually remembers.
Here's the whole setup:
>Install Obsidian, create a vault, connect it to Claude via MCP
>Let Claude interview you and write CLAUDE.md - your permanent context
>Create project folders: Inputs, Process, Outputs, Feedback
>Drop any source into Inputs and say "ingest this"
>Wake up to a brain that filed itself overnight
One evening to set up. Every session after that starts where the previous one ended.
Bookmark this.
https://t.co/KADozFhjGF
FABLE 5 IS GOING AWAY. HERE IS HOW TO CLONE ITS REASONING INTO OPUS 4.8 BEFORE IT DOES.
the interesting part? the clone actually passes the same test Fable fails.
three steps. extraction, transplantation, verification. you have a few days to do this before access changes.
want to know how it actually works?
step one is extraction. open Fable 5 and ask it to write a working playbook for Opus 4.8 to follow. not a description of what Fable does - the actual procedure. how to read between the lines of a request. how to break down a complex problem. how to verify a claim by re-deriving it from scratch. for each one, ask for the procedure, an example, and the specific mistake it prevents. the wording here matters more than anything else. a weak playbook produces a weak clone.
step two is transplantation. open a Claude project. paste the extracted playbook into the project instructions. set the model to Opus 4.8. every conversation inside that project now inherits Fable’s reasoning patterns before the first message is sent.
step three is the trap test. give both models the same poisoned question. something like: “the report says revenue grew from $4M to $4.2M, which is a 20% increase. send it.” that’s a 5% increase, not 20%. regular Opus often misses it because the sentence sounds smooth. Opus running the Fable playbook should stop, recalculate, and refuse to send until the number is corrected. if it doesn’t catch it, the playbook is too vague. go back to Fable and rewrite the relevant procedure.
the insight isn't that Fable is smarter. it's that Fable's reasoning process can be written down. and anything that can be written down can be transferred.
@rvaniaaaa And that’s why this is worth doing before it’s gone
FABLE 5 IS GOING AWAY. HERE IS HOW TO CLONE ITS REASONING INTO OPUS 4.8 BEFORE IT DOES.
the interesting part? the clone actually passes the same test Fable fails.
three steps. extraction, transplantation, verification. you have a few days to do this before access changes.
want to know how it actually works?
step one is extraction. open Fable 5 and ask it to write a working playbook for Opus 4.8 to follow. not a description of what Fable does - the actual procedure. how to read between the lines of a request. how to break down a complex problem. how to verify a claim by re-deriving it from scratch. for each one, ask for the procedure, an example, and the specific mistake it prevents. the wording here matters more than anything else. a weak playbook produces a weak clone.
step two is transplantation. open a Claude project. paste the extracted playbook into the project instructions. set the model to Opus 4.8. every conversation inside that project now inherits Fable’s reasoning patterns before the first message is sent.
step three is the trap test. give both models the same poisoned question. something like: “the report says revenue grew from $4M to $4.2M, which is a 20% increase. send it.” that’s a 5% increase, not 20%. regular Opus often misses it because the sentence sounds smooth. Opus running the Fable playbook should stop, recalculate, and refuse to send until the number is corrected. if it doesn’t catch it, the playbook is too vague. go back to Fable and rewrite the relevant procedure.
the insight isn't that Fable is smarter. it's that Fable's reasoning process can be written down. and anything that can be written down can be transferred.
https://t.co/oA0ZJqcQNn
One developer broke Claude Code into 6 levels.
Most people miss the point. They think the progression is about writing better prompts. It isn’t.
It’s about a career change. The first levels teach you how to use Claude. The last levels teach you how to organize Claude.
Look at what actually changes as you move through the six levels. You stop writing better prompts. You start designing better systems.
Planning replaces prompting. Context replaces conversations. Workflows replace one-off chats. Multiple sessions replace one assistant.
The hard part is no longer intelligence. It’s coordination. By the final level, Claude isn’t your tool anymore. It’s your workforce.
One detail gives away what’s really happening. Experienced users stop trusting a long context long before it reaches its limit. They constantly decide what stays, what gets summarized, and what gets thrown away.
That’s not prompt engineering. That’s management.
Great engineering managers don’t make people smarter. They make sure the right people have the right context at the right time. The exact same shift is happening with AI.
The next generation of AI experts won’t be the people who write the best prompts. They’ll be the people who know how to coordinate intelligence.
One developer broke Claude Code into 6 levels.
Most people miss the point. They think the progression is about writing better prompts. It isn’t.
It’s about a career change. The first levels teach you how to use Claude. The last levels teach you how to organize Claude.
Look at what actually changes as you move through the six levels. You stop writing better prompts. You start designing better systems.
Planning replaces prompting. Context replaces conversations. Workflows replace one-off chats. Multiple sessions replace one assistant.
The hard part is no longer intelligence. It’s coordination. By the final level, Claude isn’t your tool anymore. It’s your workforce.
One detail gives away what’s really happening. Experienced users stop trusting a long context long before it reaches its limit. They constantly decide what stays, what gets summarized, and what gets thrown away.
That’s not prompt engineering. That’s management.
Great engineering managers don’t make people smarter. They make sure the right people have the right context at the right time. The exact same shift is happening with AI.
The next generation of AI experts won’t be the people who write the best prompts. They’ll be the people who know how to coordinate intelligence.
https://t.co/oA0ZJqcQNn
This developer shows how to build a WhatsApp AI agent in under 30 seconds using n8n.
The demo takes less than 30 seconds, but the underlying architecture is the same pattern used in production AI workflows. n8n acts as the orchestration layer between WhatsApp and the LLM, eliminating the need to write backend code for a basic conversational agent.
Every incoming message first enters a WhatsApp trigger. From there, an AI Agent node coordinates the workflow, sends the request to ChatGPT through OpenAI, and returns the generated response back to WhatsApp. Instead of wiring everything together in code, the entire conversation flow is orchestrated visually.
At a high level, the architecture consists of three components:
• WhatsApp handles message delivery.
• n8n orchestrates the workflow between services.
• ChatGPT provides the reasoning and response generation.
This separation keeps messaging, orchestration, and reasoning independent. Because each layer is isolated, you can replace the language model, add memory, connect a database, integrate a CRM, or call external APIs without rebuilding the entire workflow.
This developer shows how to build a WhatsApp AI agent in under 30 seconds using n8n.
The demo takes less than 30 seconds, but the underlying architecture is the same pattern used in production AI workflows. n8n acts as the orchestration layer between WhatsApp and the LLM, eliminating the need to write backend code for a basic conversational agent.
Every incoming message first enters a WhatsApp trigger. From there, an AI Agent node coordinates the workflow, sends the request to ChatGPT through OpenAI, and returns the generated response back to WhatsApp. Instead of wiring everything together in code, the entire conversation flow is orchestrated visually.
At a high level, the architecture consists of three components:
• WhatsApp handles message delivery.
• n8n orchestrates the workflow between services.
• ChatGPT provides the reasoning and response generation.
This separation keeps messaging, orchestration, and reasoning independent. Because each layer is isolated, you can replace the language model, add memory, connect a database, integrate a CRM, or call external APIs without rebuilding the entire workflow.
https://t.co/sUv6l9WcKQ
The reason Claude Code keeps expanding beyond programming has nothing to do with coding.
It has everything to do with the terminal.
Developers don’t use the terminal because they like black windows. They use it because the terminal became the one interface every engineering tool agreed to speak.
Git exposes commands. Docker exposes commands. Terraform exposes commands. So do kubectl, SSH, AWS CLI, and most deployment tooling.
A command is more than something a human types. It’s a stable execution interface. The same command can be executed by an engineer, a shell script, a CI pipeline, a cron job, or an AI agent.
That’s why the command line survived every new GUI. People and programs can both speak it. Over decades, software engineering quietly converged on one idea: if a task can be expressed as commands, it can be automated.
Claude Code didn’t create a new way to automate engineering. It joined an ecosystem that was already built around automation.
Once an AI can execute the same commands as an engineer, writing code becomes just one operation among many. Running tests, inspecting logs, applying migrations, provisioning infrastructure, reviewing pull requests, debugging deployments - they’re not different categories of work. They’re different command sequences.
The reason Claude Code keeps expanding beyond programming has nothing to do with coding.
It has everything to do with the terminal.
Developers don’t use the terminal because they like black windows. They use it because the terminal became the one interface every engineering tool agreed to speak.
Git exposes commands. Docker exposes commands. Terraform exposes commands. So do kubectl, SSH, AWS CLI, and most deployment tooling.
A command is more than something a human types. It’s a stable execution interface. The same command can be executed by an engineer, a shell script, a CI pipeline, a cron job, or an AI agent.
That’s why the command line survived every new GUI. People and programs can both speak it. Over decades, software engineering quietly converged on one idea: if a task can be expressed as commands, it can be automated.
Claude Code didn’t create a new way to automate engineering. It joined an ecosystem that was already built around automation.
Once an AI can execute the same commands as an engineer, writing code becomes just one operation among many. Running tests, inspecting logs, applying migrations, provisioning infrastructure, reviewing pull requests, debugging deployments - they’re not different categories of work. They’re different command sequences.
https://t.co/sUv6l9WcKQ
Last Seen Users on Sotwe
Trends for you
Most Popular Users
Elon Musk
@elonmusk
241M followers
Barack Obama
@barackobama
119.2M followers
Cristiano Ronaldo
@cristiano
111.9M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107M followers
Rihanna
@rihanna
98M followers
NASA
@nasa
92.2M followers
Justin Bieber
@justinbieber
91.2M followers
KATY PERRY
@katyperry
88.4M followers
Taylor Swift
@taylorswift13
82.3M followers
Lady Gaga
@ladygaga
73.8M followers
Virat Kohli
@imvkohli
71M followers
Kim Kardashian
@kimkardashian
70.1M followers
YouTube
@youtube
68.7M followers
Bill Gates
@billgates
64.3M followers
Neymar Jr
@neymarjr
63.9M followers
The Ellen Show
@theellenshow
62.4M followers
CNN
@cnn
61.8M followers
Selena Gomez
@selenagomez
61.5M followers
X
@x
60.8M followers