Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

BEFORE: New session. Paste your context. Explain your project again. Repeat tomorrow. AFTER: New session. Opus 5.5 reads its memory. Picks up exactly where you left off. Plain markdown files. Version checks so nothing gets overwritten. Tests so you know it works. How to build it from scratch

CyrilXBT

197,284 subscribers

23,351 views • 2 days ago •via X (Twitter)

Gaming Education Science & Technology

Anya Rossi• Live Now

Private livecam show

6 Comments

Wen2 days ago

做过 agent 的人都懂，memory 难的从来不是写入，是三个月后它翻出一份早就作废的决策，还一本正经地接着用——version check 这个细节算是抓到点子上了。

小莓聊车小故事2 days ago

Nice work

小云聊车分享记2 days ago

walked away from the first with a random wag and now I cannot take the show sad one back

UKong2 days ago

终于不用每天重新喂上下文了这才是真效率

Yaki2 days ago

markdown files in git is elite

Sibyl Labs2 days ago

if you're interested in a serious solution let us know!!

Related Videos

If you have a file called context.txt that you paste into Claude every morning, read this. You already built a memory system. A manual one. That you update by hand. And forget to update. Let Opus 5.5 do it. It writes what it learns. Cleans it up. Loads only what each task needs. Your context.txt, but automatic. How to build it

If you have a file called context.txt that you paste into Claude every morning, read this. You already built a memory system. A manual one. That you update by hand. And forget to update. Let Opus 5.5 do it. It writes what it learns. Cleans it up. Loads only what each task needs. Your context.txt, but automatic. How to build it

CyrilXBT

23,641 views • 3 days ago

Boris Cherny gave Claude Code a memory that has no database, no indexing, no setup, and 80% of Anthropic runs on it a CLAUDE.md "gets automatically read into context at the start of every session" 12:48 - CLAUDE.md loads into memory every session 13:15 - Keep memory short or it burns your context 15:48 - Memory stacks from you to your whole company 17:30 - The memory command shows what Claude remembers 19:33 - Hit pound and Claude remembers it forever the whole memory layer is files → drop a CLAUDE.md in your repo and Claude reads it in on every session → nest more in subfolders and it pulls them only where it works → stack them personal to project to enterprise so one config serves the whole team → hit the pound key mid task and Claude writes the new fact into memory itself the best coding agent's memory was never a database, it was markdown you can open and edit

Boris Cherny gave Claude Code a memory that has no database, no indexing, no setup, and 80% of Anthropic runs on it a CLAUDE.md "gets automatically read into context at the start of every session" 12:48 - CLAUDE.md loads into memory every session 13:15 - Keep memory short or it burns your context 15:48 - Memory stacks from you to your whole company 17:30 - The memory command shows what Claude remembers 19:33 - Hit pound and Claude remembers it forever the whole memory layer is files → drop a CLAUDE.md in your repo and Claude reads it in on every session → nest more in subfolders and it pulls them only where it works → stack them personal to project to enterprise so one config serves the whole team → hit the pound key mid task and Claude writes the new fact into memory itself the best coding agent's memory was never a database, it was markdown you can open and edit

NO1ennn

13,015 views • 2 months ago

Reverse engineering with REA + Opus 5.5 takes any program apart piece by piece, and it's f**king unreal websites, games, apps what took years, REA now takes apart in one evening and copies everything without you writing a single line of code this file turns Claude into a system that knows everything: 1,143 files, 7,449 functions, every button first it reads the whole thing, over 1,000 files and thousands of pieces of code, and marks everything that still needs checking Opus 5.5 plans, Sonnet 5.5 does most of the work and a cheap Haiku finds the right spots in the code, so the expensive model only switches on when it has to think then REA opens the program itself, clicks the buttons and watches what happens on screen a second Opus 5.5 checks every conclusion and only lets through what's proven findings are saved between sessions, so tomorrow it picks up right where it stopped REA doesn't copy anyone's code. it explains how every button works, helps you write your own version from scratch and tests it against the original save it before you pay for your next tool ↓

Reverse engineering with REA + Opus 5.5 takes any program apart piece by piece, and it's f**king unreal websites, games, apps what took years, REA now takes apart in one evening and copies everything without you writing a single line of code this file turns Claude into a system that knows everything: 1,143 files, 7,449 functions, every button first it reads the whole thing, over 1,000 files and thousands of pieces of code, and marks everything that still needs checking Opus 5.5 plans, Sonnet 5.5 does most of the work and a cheap Haiku finds the right spots in the code, so the expensive model only switches on when it has to think then REA opens the program itself, clicks the buttons and watches what happens on screen a second Opus 5.5 checks every conclusion and only lets through what's proven findings are saved between sessions, so tomorrow it picks up right where it stopped REA doesn't copy anyone's code. it explains how every button works, helps you write your own version from scratch and tests it against the original save it before you pay for your next tool ↓

Noisy

12,672 views • 6 days ago

If you're a student, build this before your next exam. Every lecture note goes into one folder. Opus 5.5 sorts it, links it and finds what connects across classes. Then you ask it: "What will probably be on the exam?" "Explain the 3 topics I understand least." "Quiz me on everything from week 4." It answers from YOUR notes, not the internet. How to set it up in under an hour 👇

If you're a student, build this before your next exam. Every lecture note goes into one folder. Opus 5.5 sorts it, links it and finds what connects across classes. Then you ask it: "What will probably be on the exam?" "Explain the 3 topics I understand least." "Quiz me on everything from week 4." It answers from YOUR notes, not the internet. How to set it up in under an hour 👇

CyrilXBT

13,772 views • 14 days ago

Introducing Cross-Project Referencing. Lovable can now pull in context from across your workspace (files, components, even chat history) so new projects don’t start from scratch. This means you don't have to rebuild the same components, rewrite the same context, or re-explain the same decisions.

Introducing Cross-Project Referencing. Lovable can now pull in context from across your workspace (files, components, even chat history) so new projects don’t start from scratch. This means you don't have to rebuild the same components, rewrite the same context, or re-explain the same decisions.

Lovable

70,613 views • 8 months ago

Have you tried forking chat sessions in VS Code? Create a new independent session that keeps the conversation history so you can explore alternative approaches without losing your original context. Watch the video to see how it works.

Have you tried forking chat sessions in VS Code? Create a new independent session that keeps the conversation history so you can explore alternative approaches without losing your original context. Watch the video to see how it works.

Visual Studio Code

28,488 views • 7 months ago

> downloaded Claude Code > thought it was just another AI chatbot > opened it up > found a chat bar > typed a prompt > nothing happened to my files > realized the difference > browser AI waits for you to bring your files > Claude Code goes into your folder > all 45 PDFs > all the transcripts > all the messy Excel sheets > reads them all > remembers them session after session > you never re-explain your project again > so before your next literature review, do this: > create a folder with all your papers > open it in Claude Code > write a CLAUDEmd with your role, field, and writing style > ask it to screen every PDF using your inclusion/exclusion criteria > let it rename, extract, and summarize while you think > the AI didn't replace your scholarship > it eliminated the 6 hours of grunt work before it starts

> downloaded Claude Code > thought it was just another AI chatbot > opened it up > found a chat bar > typed a prompt > nothing happened to my files > realized the difference > browser AI waits for you to bring your files > Claude Code goes into your folder > all 45 PDFs > all the transcripts > all the messy Excel sheets > reads them all > remembers them session after session > you never re-explain your project again > so before your next literature review, do this: > create a folder with all your papers > open it in Claude Code > write a CLAUDEmd with your role, field, and writing style > ask it to screen every PDF using your inclusion/exclusion criteria > let it rename, extract, and summarize while you think > the AI didn't replace your scholarship > it eliminated the 6 hours of grunt work before it starts

Blum

19,698 views • 5 months ago

How to cut your AI bill by 60% switching to Opus 5.5: Most people will swap the model name, save 24%, and stop there. The other 37% is sitting in your settings. Here's the real before and after on an example agent workload. One month: 100M input tokens (80M cached reads, 5M cache writes, 15M uncached) 10M output tokens BEFORE: Opus 5 Cache reads: 80M × $0.50 = $40 Cache writes: 5M × $6.25 = $31.25 Uncached input: 15M × $5 = $75 Output: 10M × $25 = $250 Total: $396.25 STEP 1. Just switch the model. Same tokens, new prices. Cache reads $0.20. Writes $5. Input $4. Output $20. $16 + $25 + $60 + $200 = $301 24% cheaper. That's what everyone's screenshotting. STEP 2. Let it write less. Box measured Opus 5.5 at 40% less verbose with no drop in accuracy. 10M output tokens becomes 6M. Output drops from $200 to $120. Total: $221. Now you're at 44% off. STEP 3. Stop running every call at max effort. Thinking can't be switched off on 5.5 anymore, so effort is your lever. Classifying, routing, formatting, summarizing? Drop the effort. Save the high settings for the calls that actually reason. Say that trims output another 25%, 6M down to 4.5M. Output: $90. Total: $191. 52% off. STEP 4. Cache the stuff you keep resending. This is the one nobody does. Cache reads went from $0.50 to $0.20. That's 60% off the cheapest line on your bill. Move your system prompt, tool definitions and repo context into the cache. Uncached input drops from 15M to 5M, cache reads go up to 90M. $18 + $25 + $20 + $90 = $153 AFTER: $153 From $396.25. 61% cheaper. Same work. The model gave you 24%. You gave yourself the other 37%. Anthropic's own number is 40% cheaper than Opus 5 on typical workloads. Your number depends on how much of your bill is output and how much of your prompt you're resending uncached every call. So check those two first. Bookmark this for when you migrate. follow CyrilXBT

How to cut your AI bill by 60% switching to Opus 5.5: Most people will swap the model name, save 24%, and stop there. The other 37% is sitting in your settings. Here's the real before and after on an example agent workload. One month: 100M input tokens (80M cached reads, 5M cache writes, 15M uncached) 10M output tokens BEFORE: Opus 5 Cache reads: 80M × $0.50 = $40 Cache writes: 5M × $6.25 = $31.25 Uncached input: 15M × $5 = $75 Output: 10M × $25 = $250 Total: $396.25 STEP 1. Just switch the model. Same tokens, new prices. Cache reads $0.20. Writes $5. Input $4. Output $20. $16 + $25 + $60 + $200 = $301 24% cheaper. That's what everyone's screenshotting. STEP 2. Let it write less. Box measured Opus 5.5 at 40% less verbose with no drop in accuracy. 10M output tokens becomes 6M. Output drops from $200 to $120. Total: $221. Now you're at 44% off. STEP 3. Stop running every call at max effort. Thinking can't be switched off on 5.5 anymore, so effort is your lever. Classifying, routing, formatting, summarizing? Drop the effort. Save the high settings for the calls that actually reason. Say that trims output another 25%, 6M down to 4.5M. Output: $90. Total: $191. 52% off. STEP 4. Cache the stuff you keep resending. This is the one nobody does. Cache reads went from $0.50 to $0.20. That's 60% off the cheapest line on your bill. Move your system prompt, tool definitions and repo context into the cache. Uncached input drops from 15M to 5M, cache reads go up to 90M. $18 + $25 + $20 + $90 = $153 AFTER: $153 From $396.25. 61% cheaper. Same work. The model gave you 24%. You gave yourself the other 37%. Anthropic's own number is 40% cheaper than Opus 5 on typical workloads. Your number depends on how much of your bill is output and how much of your prompt you're resending uncached every call. So check those two first. Bookmark this for when you migrate. follow CyrilXBT

CyrilXBT

15,803 views • 26 days ago

i asked opus 5.5 to read every note i wrote this year and tell me what i keep avoiding it took [real time] it said: "[paste its actual answer]" it was right. that's not a notes app anymore. full course on how to build it below

i asked opus 5.5 to read every note i wrote this year and tell me what i keep avoiding it took [real time] it said: "[paste its actual answer]" it was right. that's not a notes app anymore. full course on how to build it below

CyrilXBT

14,327 views • 14 days ago

Codex and ChatGPT can now understand the context of your recent work. Opt into Computer History to give ChatGPT richer context, so it can pick up where you left off, understand patterns in your work, and suggest skills or scheduled tasks for work you repeat.

Codex and ChatGPT can now understand the context of your recent work. Opt into Computer History to give ChatGPT richer context, so it can pick up where you left off, understand patterns in your work, and suggest skills or scheduled tasks for work you repeat.

OpenAI Developers

500,231 views • 2 months ago

Claude Code just shipped a new /skill-doctor command It shows which skills are loaded in your session, which ones haven’t been used, and how many context tokens each is costing you Now you know exactly what to remove

Claude Code just shipped a new /skill-doctor command It shows which skills are loaded in your session, which ones haven’t been used, and how many context tokens each is costing you Now you know exactly what to remove

Daniel San

70,188 views • 1 month ago

BEFORE: 2,000 notes. Zero I could find. Inbox full of guilt. AFTER: Inbox empty every morning. Every note linked. Weekly review written for me. Same notes. Different brain. The difference is Opus 5.5 and one CLAUDE. md file. How to set it up from scratch 👇

BEFORE: 2,000 notes. Zero I could find. Inbox full of guilt. AFTER: Inbox empty every morning. Every note linked. Weekly review written for me. Same notes. Different brain. The difference is Opus 5.5 and one CLAUDE. md file. How to set it up from scratch 👇

CyrilXBT

31,206 views • 15 days ago

I am Open-Sourcing PyMolAI! Meet PyMolAI, an AI agent that can talk to your protein structures. Built on top of PyMOL, PyMolAI lets you interact with your structures in plain language. Whether you're: - Analyzing protein structures - Aligning complexes - Creating publication-ready figures - Or running design workflows PyMolAI interprets your request, executes the necessary PyMOL commands, and manages the workflow for you. It integrates with OpenBio APIs, giving you access to tools like Boltz, ProteinMPNN, and BoltzGen — directly from your PyMOL session. It has local chat history with session syncing, so you can pick up exactly where you left off.

I am Open-Sourcing PyMolAI! Meet PyMolAI, an AI agent that can talk to your protein structures. Built on top of PyMOL, PyMolAI lets you interact with your structures in plain language. Whether you're: - Analyzing protein structures - Aligning complexes - Creating publication-ready figures - Or running design workflows PyMolAI interprets your request, executes the necessary PyMOL commands, and manages the workflow for you. It integrates with OpenBio APIs, giving you access to tools like Boltz, ProteinMPNN, and BoltzGen — directly from your PyMOL session. It has local chat history with session syncing, so you can pick up exactly where you left off.

Ravi Sharma

91,056 views • 8 months ago

Someone gave Claude Code permanent memory. → 95% fewer tokens per session → Never hits context limits → Picks up exactly where you left off → 46,000 GitHub stars in 48 hours One command: `npx claude-mem install` Completely free. Community-built. Not Anthropic. Claude Code is exceptional. It just kept forgetting everything. This fixes that.

Someone gave Claude Code permanent memory. → 95% fewer tokens per session → Never hits context limits → Picks up exactly where you left off → 46,000 GitHub stars in 48 hours One command: `npx claude-mem install` Completely free. Community-built. Not Anthropic. Claude Code is exceptional. It just kept forgetting everything. This fixes that.

Charlie Hills

120,661 views • 6 months ago

Obsidian trading journal = a system that studies every trade you make. Works while you sleep. Every pattern identified. Every edge validated. Every mistake documented so it never costs you twice. After 30 days it knows your trading better than you do. After 90 days it is preventing losses your past positions already predicted. The people who build this tonight will never trade the same way again. Read this and Bookmark it now.

Obsidian trading journal = a system that studies every trade you make. Works while you sleep. Every pattern identified. Every edge validated. Every mistake documented so it never costs you twice. After 30 days it knows your trading better than you do. After 90 days it is preventing losses your past positions already predicted. The people who build this tonight will never trade the same way again. Read this and Bookmark it now.

CyrilXBT

205,924 views • 4 months ago

I made Apple's App Store is obsolete. I built an Claude Code for your iPhone, so you can build and run whatever app you want all on your device! Just describe an app and it writes the Swift, runs it, reads its own errors, and fixes them. Then it lands on your app's home screen. Here's exactly how it works 👇

I made Apple's App Store is obsolete. I built an Claude Code for your iPhone, so you can build and run whatever app you want all on your device! Just describe an app and it writes the Swift, runs it, reads its own errors, and fixes them. Then it lands on your app's home screen. Here's exactly how it works 👇

JC Builds

34,514 views • 3 months ago

One of the best applications for agentic loops is SEO. Here's why it works: the outcome is "black or white." You moved up this month, you moved down, or you stayed the same. That's the entire requirement for a loop — a metric that grades the work for you. Build it in five parts: 1) Access. - Connect it to your real data, not your description of your data. - Google Search Console has an API. - DataForSEO shows you the four pages ranking above yours. 2) Action. - Let it audit and fix before it experiments. - Meta tags, JSON-LD, a missing sitemap. - Quick wins on most sites. 3) Memory. - A markdown file where it logs every change it made and why. 4) Judgment. - Every two weeks it reopens that file and checks its own work: I rewrote that description — did it rank us higher or lower? 5) Reversal. - A change that took you from position 20 to 30 gets undone. - Nothing here is set in stone. Then put it on a routine so it picks up where it left off without you. The catch: this takes months, not days. Months of nothing, then month four, bam bam bam, page one. So find the number in your business that can't be argued with. Build the loop around that one.

One of the best applications for agentic loops is SEO. Here's why it works: the outcome is "black or white." You moved up this month, you moved down, or you stayed the same. That's the entire requirement for a loop — a metric that grades the work for you. Build it in five parts: 1) Access. - Connect it to your real data, not your description of your data. - Google Search Console has an API. - DataForSEO shows you the four pages ranking above yours. 2) Action. - Let it audit and fix before it experiments. - Meta tags, JSON-LD, a missing sitemap. - Quick wins on most sites. 3) Memory. - A markdown file where it logs every change it made and why. 4) Judgment. - Every two weeks it reopens that file and checks its own work: I rewrote that description — did it rank us higher or lower? 5) Reversal. - A change that took you from position 20 to 30 gets undone. - Nothing here is set in stone. Then put it on a routine so it picks up where it left off without you. The catch: this takes months, not days. Months of nothing, then month four, bam bam bam, page one. So find the number in your business that can't be argued with. Build the loop around that one.

The Startup Ideas Podcast (SIP) 🧃

11,505 views • 2 months ago

Anthropic just revealed a trick to improve how Claude Opus 5.5 works. The idea is that Opus does the work while Fable 5.1 acts as a second opinion. All you have to do is open Claude Code and run: /advisor fable When Opus consults Fable, it gets the full context of the session, including the tools used and their results. It can be especially useful: → Before creating a large plan → When the same error keeps happening → Before marking a task as finished This gives you a second review without having to copy context between conversations. The feature is still experimental, requires access to Fable, and uses additional tokens. 🔖 Save this for your next session.

Anthropic just revealed a trick to improve how Claude Opus 5.5 works. The idea is that Opus does the work while Fable 5.1 acts as a second opinion. All you have to do is open Claude Code and run: /advisor fable When Opus consults Fable, it gets the full context of the session, including the tools used and their results. It can be especially useful: → Before creating a large plan → When the same error keeps happening → Before marking a task as finished This gives you a second review without having to copy context between conversations. The feature is still experimental, requires access to Fable, and uses additional tokens. 🔖 Save this for your next session.

Dean W. Perkins

56,146 views • 16 days ago

The reason Claude keeps disappointing you is not Claude - it’s you. Every session - explain yourself from scratch. Every session - it forgets you exist. Every session - same mistakes. Same corrections. Same waste. That’s not AI. That’s a goldfish with a keyboard. One file broke that loop permanently. 21 lines. That’s it. Now Claude knows his voice. His rules. His decisions. What failed. What worked. Where he left off. He doesn’t explain himself. He doesn’t repeat himself. He just builds. While you’re re-explaining your job for the 200th time - he already shipped. Tony Stark didn’t prompt J.A.R.V.I.S. He talked to it like it knew him. Because it did. This guy has that. In a wine bar. On a laptop. With a text file. 82,000 GitHub stars. #1 trending. And you still don’t have it set up. Bookmark this before you forget. Because you will forget. And tomorrow you’ll be explaining yourself to Claude again.

The reason Claude keeps disappointing you is not Claude - it’s you. Every session - explain yourself from scratch. Every session - it forgets you exist. Every session - same mistakes. Same corrections. Same waste. That’s not AI. That’s a goldfish with a keyboard. One file broke that loop permanently. 21 lines. That’s it. Now Claude knows his voice. His rules. His decisions. What failed. What worked. Where he left off. He doesn’t explain himself. He doesn’t repeat himself. He just builds. While you’re re-explaining your job for the 200th time - he already shipped. Tony Stark didn’t prompt J.A.R.V.I.S. He talked to it like it knew him. Because it did. This guy has that. In a wine bar. On a laptop. With a text file. 82,000 GitHub stars. #1 trending. And you still don’t have it set up. Bookmark this before you forget. Because you will forget. And tomorrow you’ll be explaining yourself to Claude again.

leopardracer

10,804 views • 5 months ago

ANTHROPIC JUST KILLED A DOZEN AI MEMORY STARTUPS WITH ONE BUILT IN TOOL people raised millions to give AI long term memory. Anthropic just shipped it straight into the API. free, no framework, no vector database. it is a folder of files Claude reads and writes. that is it. the trick: it is client side. Claude never touches your data directly. it only requests a file operation, and your app runs it against storage you control. before every task Claude checks its /memories directory, pulls back what it learned last time, and keeps going. when it learns something new, it writes it down. six plain operations do all of it: > view, create, str_replace, insert, delete, rename why it actually wins: > just in time retrieval, it reads on demand instead of loading everything up front > it survives a context reset, so a long job resumes instead of starting over > every path is locked to /memories, so it cannot wander your filesystem > works on every Claude 4 model, no beta flag, generally available Anthropic even pairs it with context editing and compaction, so the live window stays tiny while the important stuff lives on disk. the whole thing in one line: Claude keeps the notes, you keep the storage. bookmark this before you pay for a memory layer you no longer need.

ANTHROPIC JUST KILLED A DOZEN AI MEMORY STARTUPS WITH ONE BUILT IN TOOL people raised millions to give AI long term memory. Anthropic just shipped it straight into the API. free, no framework, no vector database. it is a folder of files Claude reads and writes. that is it. the trick: it is client side. Claude never touches your data directly. it only requests a file operation, and your app runs it against storage you control. before every task Claude checks its /memories directory, pulls back what it learned last time, and keeps going. when it learns something new, it writes it down. six plain operations do all of it: > view, create, str_replace, insert, delete, rename why it actually wins: > just in time retrieval, it reads on demand instead of loading everything up front > it survives a context reset, so a long job resumes instead of starting over > every path is locked to /memories, so it cannot wander your filesystem > works on every Claude 4 model, no beta flag, generally available Anthropic even pairs it with context editing and compaction, so the live window stays tiny while the important stuff lives on disk. the whole thing in one line: Claude keeps the notes, you keep the storage. bookmark this before you pay for a memory layer you no longer need.

NO1ennn

42,946 views • 2 months ago