正在加载视频...

视频加载失败

返回首页

Introducing: Granola CLI/Claude Code Skill/OpenClaw and Hermes skill from the Printing Press printed by Damien Stevens . - Cross-meeting SQLite search - MEMO pipeline runner - Attendee timelines - Stop the MCP logged-out pain Really excited about this one. I can't live without Granola I may have told Damien...

mvanhorn's profile picture

Matt Van Horn

36,399 subscribers

34,851 次观看 • 2 个月前 •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Introducing: Tesla CLI/Claude Code Skill/OpenClaw and Hermes skill from the Printing Press. - "Unlock the car" and "turn on dog mode" as one-line commands, callable from your phone or laptop - Agent: "during winter school days, defrost my car at 7:50 every weekday before school dropoff" - Charging cost ledger - Supercharger queue watcher pageable from an agent - Your signing key stays on YOUR host - Much more Fun fact: when I got my first ClawdBot, Tesla was one of the first skills I made. But I could only get it to work with my older Tesla. Now that I have the Printing Press, I was able to build what I wanted soup to nuts and now it's here.
0:41

Introducing: Tesla CLI/Claude Code Skill/OpenClaw and Hermes skill from the Printing Press. - "Unlock the car" and "turn on dog mode" as one-line commands, callable from your phone or laptop - Agent: "during winter school days, defrost my car at 7:50 every weekday before school dropoff" - Charging cost ledger - Supercharger queue watcher pageable from an agent - Your signing key stays on YOUR host - Much more Fun fact: when I got my first ClawdBot, Tesla was one of the first skills I made. But I could only get it to work with my older Tesla. Now that I have the Printing Press, I was able to build what I wanted soup to nuts and now it's here.

Matt Van Horn

128,553 次观看 • 2 个月前

"I am a Claude Code maximalist. I try to live in it all day if I can." Carl Vellotti 🥞 has gone deeper on Claude Code than any PM I know. In my new episode, he showed me how to connect Claude Code to: → Google Workspace: Prep for meetings → Linear: Create tickets from PRDs → Slack: Share status updates → Reddit: Monitor subreddit discussions He does all of the above using skills and MCPs in his personal Claude Code OS. Some quotes from Carl: "Connecting Claude Code to the apps that you use at work makes it way more powerful." "Why would you need to go to all these websites if you can just get the info right here?" "I use this consult-the-council skill to get Claude, ChatGPT, and Grok to all review my work in one shot." 📌 Watch now: Thanks to our sponsors: Granola - The best AI meeting notes app I've ever used: Replit ⠕ - Create beautiful prototypes and full stack apps:
0:41

"I am a Claude Code maximalist. I try to live in it all day if I can." Carl Vellotti 🥞 has gone deeper on Claude Code than any PM I know. In my new episode, he showed me how to connect Claude Code to: → Google Workspace: Prep for meetings → Linear: Create tickets from PRDs → Slack: Share status updates → Reddit: Monitor subreddit discussions He does all of the above using skills and MCPs in his personal Claude Code OS. Some quotes from Carl: "Connecting Claude Code to the apps that you use at work makes it way more powerful." "Why would you need to go to all these websites if you can just get the info right here?" "I use this consult-the-council skill to get Claude, ChatGPT, and Grok to all review my work in one shot." 📌 Watch now: Thanks to our sponsors: Granola - The best AI meeting notes app I've ever used: Replit ⠕ - Create beautiful prototypes and full stack apps:

Peter Yang

142,115 次观看 • 5 个月前

I sat down with matt van horn (Matt Van Horn) and watched him turn claude code into a real-time research engine with his /last30days claude code skill he "fixes" claude code in 30 seconds. this skill pulls what’s actually working right now from x, reddit, and the web, then feeds that context straight into your prompts so you stop building off stale advice. we went from trending rap songs → cold email frameworks → researching clawdbot → planning and building a competitor live, with almost zero hand-written code. pretty nifty little claude code skill share this with a friend / full ep available on The Startup Ideas Podcast (SIP) 🧃 where i will give you ideas/tools/tutorials to make your dreams a reality i will not hold back any alpha and this claude skill is alpha forsure you can install this claude code skill in 30 seconds dream big my friends
25:21

I sat down with matt van horn (Matt Van Horn) and watched him turn claude code into a real-time research engine with his /last30days claude code skill he "fixes" claude code in 30 seconds. this skill pulls what’s actually working right now from x, reddit, and the web, then feeds that context straight into your prompts so you stop building off stale advice. we went from trending rap songs → cold email frameworks → researching clawdbot → planning and building a competitor live, with almost zero hand-written code. pretty nifty little claude code skill share this with a friend / full ep available on The Startup Ideas Podcast (SIP) 🧃 where i will give you ideas/tools/tutorials to make your dreams a reality i will not hold back any alpha and this claude skill is alpha forsure you can install this claude code skill in 30 seconds dream big my friends

GREG ISENBERG

176,694 次观看 • 6 个月前

This guy literally tested Hermes Agent vs OpenClaw side by side. Hermes won. And it wasn't close. Here's what made the difference: → Hermes built the website faster → Hermes gave live updates the whole time. OpenClaw went silent. → Hermes designed a cleaner page using the exact same AI model → And then Hermes did something OpenClaw can't do at all. It learned from the task. Every mistake gets logged. Every win gets saved as a reusable skill. The more you use it, the smarter it gets. OpenClaw doesn't do this unless you manually set up a learning file. Hermes does it automatically. Every single time. Other wins: → Runs free with Ollama + MiniMax → Works on Telegram, Discord, Slack, WhatsApp, Signal, and Email → Supports 200+ AI models → Built-in migration tool from OpenClaw — one command to switch 11,600 GitHub stars. 142 contributors. If you're already on OpenClaw, you don't have to start over. Just ask Hermes to import everything.
10:34

This guy literally tested Hermes Agent vs OpenClaw side by side. Hermes won. And it wasn't close. Here's what made the difference: → Hermes built the website faster → Hermes gave live updates the whole time. OpenClaw went silent. → Hermes designed a cleaner page using the exact same AI model → And then Hermes did something OpenClaw can't do at all. It learned from the task. Every mistake gets logged. Every win gets saved as a reusable skill. The more you use it, the smarter it gets. OpenClaw doesn't do this unless you manually set up a learning file. Hermes does it automatically. Every single time. Other wins: → Runs free with Ollama + MiniMax → Works on Telegram, Discord, Slack, WhatsApp, Signal, and Email → Supports 200+ AI models → Built-in migration tool from OpenClaw — one command to switch 11,600 GitHub stars. 142 contributors. If you're already on OpenClaw, you don't have to start over. Just ask Hermes to import everything.

Julian Goldie SEO

40,877 次观看 • 4 个月前

Michael Cohen's former legal adviser says Cohen told him "I swear to God I don't have anything on Donald Trump." "I can't sit by and let this guy [lie] to the jurors, when he told us the opposite story at a time when he was desperately seeking a way out."
1:25

Michael Cohen's former legal adviser says Cohen told him "I swear to God I don't have anything on Donald Trump." "I can't sit by and let this guy [lie] to the jurors, when he told us the opposite story at a time when he was desperately seeking a way out."

MAGA War Room

67,637 次观看 • 2 年前

Peter Steinberger Quietly Started A Shift That Makes Claude 35x Cheaper To Run The way Claude Code talks to tools matters more than people think. Most use MCP, and it quietly eats tokens, it loads everything into context every time, even tools you never call. One benchmark: MCP burned 35x more tokens than a CLI on the same task. Peter Steinberger, the OpenClaw guy, got annoyed by this and started building lean CLIs himself. That kicked off a tool called Printing Press. You point Claude Code at any website, even ones with no API like ESPN or Craigslist, and it builds a small command tool for it in about 10 minutes. In one demo, a request pulled 132,000 tokens of raw data, but the tool processed it locally and only handed Claude a 2,000-token summary. The rest never touched the context window. It also comes with ~50 ready-made tools you can grab right away. To start, point Claude Code at the links from and ask it to set it up. Bookmark this.
14:45

Peter Steinberger Quietly Started A Shift That Makes Claude 35x Cheaper To Run The way Claude Code talks to tools matters more than people think. Most use MCP, and it quietly eats tokens, it loads everything into context every time, even tools you never call. One benchmark: MCP burned 35x more tokens than a CLI on the same task. Peter Steinberger, the OpenClaw guy, got annoyed by this and started building lean CLIs himself. That kicked off a tool called Printing Press. You point Claude Code at any website, even ones with no API like ESPN or Craigslist, and it builds a small command tool for it in about 10 minutes. In one demo, a request pulled 132,000 tokens of raw data, but the tool processed it locally and only handed Claude a 2,000-token summary. The rest never touched the context window. It also comes with ~50 ready-made tools you can grab right away. To start, point Claude Code at the links from and ask it to set it up. Bookmark this.

Ridark

32,351 次观看 • 1 个月前

One thing I really loved about this showing of Speed Racer is they shown a little video before hand and explaining very briefly the history of the film! I managed to record most of it. Only missing the intro and stopping when I thought it was over, missing him introducing the cast. I would love to have this in full for archives in the museum!
3:31

One thing I really loved about this showing of Speed Racer is they shown a little video before hand and explaining very briefly the history of the film! I managed to record most of it. Only missing the intro and stopping when I thought it was over, missing him introducing the cast. I would love to have this in full for archives in the museum!

The Mach 5 Museum

30,959 次观看 • 3 个月前

🚨HOLY SHIT: Trump chickens out about Epstein and tells the press he won’t be answering questions: “I really don't want to soil it up by answering questions, even questions that are very fair questions. I think we have to just stop right here.” COWARD!
0:41

🚨HOLY SHIT: Trump chickens out about Epstein and tells the press he won’t be answering questions: “I really don't want to soil it up by answering questions, even questions that are very fair questions. I think we have to just stop right here.” COWARD!

CALL TO ACTIVISM

541,411 次观看 • 7 个月前

Everyone keeps talking about the possibilities of making money with Openclaw, but nobody actually shows you how So, I made a skill for your Openclaw that transforms it into a one-man AI real estate marketing machine. This skill took me 2 weeks, 50+ hours and ~$800 in API credits to make With this skill, your Openclaw will autonomously make videos like the one below by simply scraping images from listings No approval gate. Fully autonomous. Output decided by your CRONs cadence. It's exactly what mine does, and that's exactly what you should be doing with yours. Run the CRON once a day; if you sell at least one video, you should make $300-$800 each. Also: total cost per video: ~$5.00 I truly believe this skill is one of the most practical ways you can actually make money with Openclaw in the lowest effort way possible. Most people will read this, maybe bookmark the tweet, and never take action. All I hope is that one of you guys actually takes action and makes money with this. Because of how long this skill took me to make, I'm only giving the entire thing away to my paid Subscribers on Twitter. You'll also have direct access to me for any troubleshooting + any other Skills I make like this in the future (one more dropping this week)! You just have to feed it to your Openclaw and link a few API keys. This will be the first of many Openclaw skills/overpowered AI that I only offer to my Subs! Full Skill + a lot more is waiting for you here for $25/m:
0:30

Everyone keeps talking about the possibilities of making money with Openclaw, but nobody actually shows you how So, I made a skill for your Openclaw that transforms it into a one-man AI real estate marketing machine. This skill took me 2 weeks, 50+ hours and ~$800 in API credits to make With this skill, your Openclaw will autonomously make videos like the one below by simply scraping images from listings No approval gate. Fully autonomous. Output decided by your CRONs cadence. It's exactly what mine does, and that's exactly what you should be doing with yours. Run the CRON once a day; if you sell at least one video, you should make $300-$800 each. Also: total cost per video: ~$5.00 I truly believe this skill is one of the most practical ways you can actually make money with Openclaw in the lowest effort way possible. Most people will read this, maybe bookmark the tweet, and never take action. All I hope is that one of you guys actually takes action and makes money with this. Because of how long this skill took me to make, I'm only giving the entire thing away to my paid Subscribers on Twitter. You'll also have direct access to me for any troubleshooting + any other Skills I make like this in the future (one more dropping this week)! You just have to feed it to your Openclaw and link a few API keys. This will be the first of many Openclaw skills/overpowered AI that I only offer to my Subs! Full Skill + a lot more is waiting for you here for $25/m:

ashen

22,822 次观看 • 4 个月前

😂💪 Merab calls himself a 'decision machine' and Petr Yan will have to KO him to win: "This time I will be more technical. I will show more skill. He will be the aggressive one. He will try to knock me out. That's only way he can stop me. Because if we fight to a decision - I am the decision machine." 🎥 UFC
0:34

😂💪 Merab calls himself a 'decision machine' and Petr Yan will have to KO him to win: "This time I will be more technical. I will show more skill. He will be the aggressive one. He will try to knock me out. That's only way he can stop me. Because if we fight to a decision - I am the decision machine." 🎥 UFC

Home of Fight

60,094 次观看 • 8 个月前

Megyn Kelly says RFK Jr. being hated by the press — but loved by the people — shows he’s doing something right. Megyn Kelly: “No one gets it as bad as Bobby from the press.” “They hate Elon, they hate Pete Hegseth, but they really hate Bobby in a particular way because he was a Democrat and you’re not allowed to turn on them.” Cheryl Hines: “But by the way, when he was a Democrat, they did not embrace him.” Megyn Kelly: “When he’s in the news for whatever, all three — the Wall Street Journal, the New York Times, and the New York Post — are bad.” “It’s very rare to have universally bad press.” “But he’s the most popular cabinet member, and the least popular with the press — which I think says he’s doing something right.”
0:43

Megyn Kelly says RFK Jr. being hated by the press — but loved by the people — shows he’s doing something right. Megyn Kelly: “No one gets it as bad as Bobby from the press.” “They hate Elon, they hate Pete Hegseth, but they really hate Bobby in a particular way because he was a Democrat and you’re not allowed to turn on them.” Cheryl Hines: “But by the way, when he was a Democrat, they did not embrace him.” Megyn Kelly: “When he’s in the news for whatever, all three — the Wall Street Journal, the New York Times, and the New York Post — are bad.” “It’s very rare to have universally bad press.” “But he’s the most popular cabinet member, and the least popular with the press — which I think says he’s doing something right.”

MAHA Action

385,611 次观看 • 9 个月前

Zendaya reveals how she freaked out when Tom Holland told her about starring in The Odyssey. "He had the script, and I may have taken a little peek. I had to know. He had a Christopher Nolan script in my house, what am I gonna do? So I looked" "Then he went off to his meeting, and when he came home, I was like, 'Okay! Tell me all about it! How was it? What was it like?' You know what I mean? I was so ready, so excited" "And he goes, 'Read this again.' And I was like, 'What? Why? We just read it last night.' He was like, 'Read it again, but with Athena in mind.' And I was like, 'Yo, stop it right there'" "I freaked out. I was so grateful. And the fact that he got to be the one to tell me, it was so special"
0:49

Zendaya reveals how she freaked out when Tom Holland told her about starring in The Odyssey. "He had the script, and I may have taken a little peek. I had to know. He had a Christopher Nolan script in my house, what am I gonna do? So I looked" "Then he went off to his meeting, and when he came home, I was like, 'Okay! Tell me all about it! How was it? What was it like?' You know what I mean? I was so ready, so excited" "And he goes, 'Read this again.' And I was like, 'What? Why? We just read it last night.' He was like, 'Read it again, but with Athena in mind.' And I was like, 'Yo, stop it right there'" "I freaked out. I was so grateful. And the fact that he got to be the one to tell me, it was so special"

Buggy

6,520,995 次观看 • 23 天前

MAY THE 4TH BE WITH YOUR AGENTS... starting today, TinyFish's `Web Search` and `Fetch` are FREE 🤯 Hand-rolling Playwright scripts and watching messy DOMs explode your context limits used to feel like battling the Dark Side. Today, TinyFish brings balance to the Force. Their Search and Fetch layer is built specifically for AI agents, and it's now free. It uses a real full-browser render to strip out raw HTML, meaning you stop paying token bills for cookie banners and heavy JavaScript 💥 Plus, a free API is pointless if you have to build a custom bridge... so TinyFish lives natively inside the stack you already use. You can literally use it anywhere: • REST API • MCP (Claude, Cursor, Codex) • Python + TS SDKs • CLI + use-tinyfish Skill (Claude Code, Cursor, OpenCode) • OpenClaw (ClawHub skill) • Hermes Agent (Nous Research) • n8n · Dify · Vercel Skills • LangChain · CrewAI If you are building the trending PrepToMe workflows or high-speed Hermes agents, this is the exact drop-in retrieval engine you need. → Search returns rank-stable, structured JSON → Fetch renders dynamic SPAs and returns clean markdown → Integrates natively via MCP, Python/TS SDKs, or direct API There are even some Star Wars Easter eggs hidden on their site today :) 4 Jedi-approved use cases 🧵 ↓ 1/ Codex Medical Case
0:37

MAY THE 4TH BE WITH YOUR AGENTS... starting today, TinyFish's `Web Search` and `Fetch` are FREE 🤯 Hand-rolling Playwright scripts and watching messy DOMs explode your context limits used to feel like battling the Dark Side. Today, TinyFish brings balance to the Force. Their Search and Fetch layer is built specifically for AI agents, and it's now free. It uses a real full-browser render to strip out raw HTML, meaning you stop paying token bills for cookie banners and heavy JavaScript 💥 Plus, a free API is pointless if you have to build a custom bridge... so TinyFish lives natively inside the stack you already use. You can literally use it anywhere: • REST API • MCP (Claude, Cursor, Codex) • Python + TS SDKs • CLI + use-tinyfish Skill (Claude Code, Cursor, OpenCode) • OpenClaw (ClawHub skill) • Hermes Agent (Nous Research) • n8n · Dify · Vercel Skills • LangChain · CrewAI If you are building the trending PrepToMe workflows or high-speed Hermes agents, this is the exact drop-in retrieval engine you need. → Search returns rank-stable, structured JSON → Fetch renders dynamic SPAs and returns clean markdown → Integrates natively via MCP, Python/TS SDKs, or direct API There are even some Star Wars Easter eggs hidden on their site today :) 4 Jedi-approved use cases 🧵 ↓ 1/ Codex Medical Case

Charly Wargnier

12,844 次观看 • 3 个月前

CLAUDE CODE CREATOR: "BUILDING SOFTWARE IS GOING TO BE A SKILL LIKE KNOWING HOW TO SEND A TEXT MESSAGE." Boris Cherny, the creator of Claude Code, just made the most important prediction about software in years. His parallel: the printing press. Before the printing press, 10% of Europeans were literate. Their entire job was reading and writing for people who couldn't. In the 50 years after it was invented, more literature was published than in the thousand years before. Book costs dropped 100x. Literacy went from 10% to 70%. He believes the same thing is about to happen to software. Faster than 50 years. The best person to write accounting software isn't an engineer. It's a great accountant. Because knowing the domain is the hard part. Coding is just the tool. AI just made that tool available to everyone. Doctors will build medical software. Teachers will build education tools. Accountants will build accounting systems. Without a single engineer involved. We're not watching software get easier. We're watching it get democratized the way literacy did 600 years ago. Bookmark this post. The full interview is in the video below.
1:01

CLAUDE CODE CREATOR: "BUILDING SOFTWARE IS GOING TO BE A SKILL LIKE KNOWING HOW TO SEND A TEXT MESSAGE." Boris Cherny, the creator of Claude Code, just made the most important prediction about software in years. His parallel: the printing press. Before the printing press, 10% of Europeans were literate. Their entire job was reading and writing for people who couldn't. In the 50 years after it was invented, more literature was published than in the thousand years before. Book costs dropped 100x. Literacy went from 10% to 70%. He believes the same thing is about to happen to software. Faster than 50 years. The best person to write accounting software isn't an engineer. It's a great accountant. Because knowing the domain is the hard part. Coding is just the tool. AI just made that tool available to everyone. Doctors will build medical software. Teachers will build education tools. Accountants will build accounting systems. Without a single engineer involved. We're not watching software get easier. We're watching it get democratized the way literacy did 600 years ago. Bookmark this post. The full interview is in the video below.

SCOTTY BEAM

10,786 次观看 • 1 个月前

HERMES AGENT LEARNS FROM ITS OWN MISTAKES. UPDATES ITS MEMORY. CREATES ITS OWN SKILLS. NO CLOUD. EVERYTHING STORED LOCALLY. THIS IS HOW THE SELF-IMPROVING LOOP WORKS. most agents start from zero every session. Hermes carries forward what it learned. THREE MEMORY SYSTEMS: 1. PROCEDURAL MEMORY (how to act) stored in ~/.hermes/skills/ as SKILL.md files. when the agent repeats a complex workflow, it saves the procedure as a reusable skill. next time the same task comes up, it follows the skill instead of figuring it out again. you can also create skills explicitly: "create a skill called video-prep that captures how I format my video scripts. spoken english, define jargon inline, no em-dashes, close with a catchphrase." the agent writes the SKILL.md. available as a slash command from that moment. Hermes ships with 90+ skills. the number grows the longer you use it. 2. SEMANTIC MEMORY (durable facts about you) stored in ~/.hermes/memory/memory.md the agent scans conversations for facts worth remembering. preferences, habits, corrections, project details. real example from the video: agent tried to scrape a YouTube channel. URL was wrong. it failed. it updated memory.md with the correct URL pattern so it never makes the same mistake again. you can also save explicitly: "save to memory that my favorite testing framework is pytest" the agent updates memory.md immediately. this file loads into context on every session. the agent knows you better every week. 3. EPISODIC MEMORY (chat history) stored in ~/.hermes/state.db (local SQLite). every conversation. every tool call. every result. searchable with FTS5 full-text search. "search our past sessions. what was the first thing I ever said to you?" the agent queries state.db and finds it. over time, auxiliary models consolidate episodic memory into semantic memory. distilling recurring patterns into durable facts. THE SELF-IMPROVING LOOP: every agent run follows this cycle: → you send a prompt → working memory loads: SOUL.md + memory.md + relevant skills + chat history → agent calls tools (terminal, browser, delegate_task) → agent completes the task, replies to you → AFTER the reply: agent checks "did I learn something worth saving?" → if yes: updates memory.md or creates a new skill → next session starts smarter than the last this happens automatically. you don't ask the agent to learn. it decides what to remember on its own. WHAT MAKES THIS DIFFERENT FROM CLAUDE CODE: Claude Code has memory too. but Hermes stores everything locally. no cloud. your data never leaves your machine. Claude Code doesn't auto-create skills from experience. Hermes turns repeated workflows into reusable procedures. Claude Code memory is instruction-based. Hermes memory is conversational and self-updating. over months of usage, Hermes builds a knowledge base of your preferences, your projects, your mistakes, and the procedures that work for your specific workflow. the agent that remembers your birthday also remembers why your last deploy failed. NO EMBEDDINGS. PLAIN TEXT. Hermes does not use embeddings or RAG for memory. skill and memory search runs on plain text keyword matching. simpler. faster. no vector database to maintain. works entirely offline on your local machine. DELEGATE TO CLAUDE CODE: Hermes can spawn a sub-agent that runs Claude Code in headless mode: "spawn a sub-agent using Claude CLI to build a Python script that fetches the top 5 Hacker News stories to markdown." Hermes delegates. Claude Code writes the code. result returns to Hermes. Hermes runs the script and delivers the output. use Hermes for orchestration. use Claude Code for heavy coding. both tools. not competitors. WHAT HERMES DOES NOT HAVE: no built-in eval or LMOps system. no LangSmith, no LangFuse integration out of the box. trajectory export and logs exist but there is no automated quality tracking. if you need eval, build it yourself or connect external tools. the loop is self-improving. measuring how well it improves is on you. comment LOOP and I'll send you the configs that control how fast Hermes learns and what it remembers. memory limits, skill auto-creation triggers, and the auxiliary model that runs the learning. Replace your entire team with 8 hermes agents👇
20:54

HERMES AGENT LEARNS FROM ITS OWN MISTAKES. UPDATES ITS MEMORY. CREATES ITS OWN SKILLS. NO CLOUD. EVERYTHING STORED LOCALLY. THIS IS HOW THE SELF-IMPROVING LOOP WORKS. most agents start from zero every session. Hermes carries forward what it learned. THREE MEMORY SYSTEMS: 1. PROCEDURAL MEMORY (how to act) stored in ~/.hermes/skills/ as SKILL.md files. when the agent repeats a complex workflow, it saves the procedure as a reusable skill. next time the same task comes up, it follows the skill instead of figuring it out again. you can also create skills explicitly: "create a skill called video-prep that captures how I format my video scripts. spoken english, define jargon inline, no em-dashes, close with a catchphrase." the agent writes the SKILL.md. available as a slash command from that moment. Hermes ships with 90+ skills. the number grows the longer you use it. 2. SEMANTIC MEMORY (durable facts about you) stored in ~/.hermes/memory/memory.md the agent scans conversations for facts worth remembering. preferences, habits, corrections, project details. real example from the video: agent tried to scrape a YouTube channel. URL was wrong. it failed. it updated memory.md with the correct URL pattern so it never makes the same mistake again. you can also save explicitly: "save to memory that my favorite testing framework is pytest" the agent updates memory.md immediately. this file loads into context on every session. the agent knows you better every week. 3. EPISODIC MEMORY (chat history) stored in ~/.hermes/state.db (local SQLite). every conversation. every tool call. every result. searchable with FTS5 full-text search. "search our past sessions. what was the first thing I ever said to you?" the agent queries state.db and finds it. over time, auxiliary models consolidate episodic memory into semantic memory. distilling recurring patterns into durable facts. THE SELF-IMPROVING LOOP: every agent run follows this cycle: → you send a prompt → working memory loads: SOUL.md + memory.md + relevant skills + chat history → agent calls tools (terminal, browser, delegate_task) → agent completes the task, replies to you → AFTER the reply: agent checks "did I learn something worth saving?" → if yes: updates memory.md or creates a new skill → next session starts smarter than the last this happens automatically. you don't ask the agent to learn. it decides what to remember on its own. WHAT MAKES THIS DIFFERENT FROM CLAUDE CODE: Claude Code has memory too. but Hermes stores everything locally. no cloud. your data never leaves your machine. Claude Code doesn't auto-create skills from experience. Hermes turns repeated workflows into reusable procedures. Claude Code memory is instruction-based. Hermes memory is conversational and self-updating. over months of usage, Hermes builds a knowledge base of your preferences, your projects, your mistakes, and the procedures that work for your specific workflow. the agent that remembers your birthday also remembers why your last deploy failed. NO EMBEDDINGS. PLAIN TEXT. Hermes does not use embeddings or RAG for memory. skill and memory search runs on plain text keyword matching. simpler. faster. no vector database to maintain. works entirely offline on your local machine. DELEGATE TO CLAUDE CODE: Hermes can spawn a sub-agent that runs Claude Code in headless mode: "spawn a sub-agent using Claude CLI to build a Python script that fetches the top 5 Hacker News stories to markdown." Hermes delegates. Claude Code writes the code. result returns to Hermes. Hermes runs the script and delivers the output. use Hermes for orchestration. use Claude Code for heavy coding. both tools. not competitors. WHAT HERMES DOES NOT HAVE: no built-in eval or LMOps system. no LangSmith, no LangFuse integration out of the box. trajectory export and logs exist but there is no automated quality tracking. if you need eval, build it yourself or connect external tools. the loop is self-improving. measuring how well it improves is on you. comment LOOP and I'll send you the configs that control how fast Hermes learns and what it remembers. memory limits, skill auto-creation triggers, and the auxiliary model that runs the learning. Replace your entire team with 8 hermes agents👇

YanXbt

22,720 次观看 • 1 个月前

When the first time I saw this, I really believed that we lost Kakashi Sensei, that goddamn PAIN has murdered our beloved Kakashi. 😭💔 But can't stop myself from praising his skills, strategy planning abilities during the fight. Not just that, he was ready to sacrifice himself to pass on the information to the higher ups.😫💔 TBH, I feel like PAIN had plot armour and got saved by the writers, otherwise Kakashi 100% deserved this kill.💯🔥
5:45

When the first time I saw this, I really believed that we lost Kakashi Sensei, that goddamn PAIN has murdered our beloved Kakashi. 😭💔 But can't stop myself from praising his skills, strategy planning abilities during the fight. Not just that, he was ready to sacrifice himself to pass on the information to the higher ups.😫💔 TBH, I feel like PAIN had plot armour and got saved by the writers, otherwise Kakashi 100% deserved this kill.💯🔥

Black Light

74,808 次观看 • 1 个月前

Chamath describes Donald Trump: "I've seen some of the most incredible survivors, they have that one skill." "Oh, he's really inspiring. I mean, I have to say here that in my limited interactions with him, he's really curious. He's really smart. His intuition is off the charts. I'll tell you a very specific example. Had a meeting with him, I had a complicated-" "What were you pitching him?" "It's not a pitch. I was just walking him through some structural issues in this specific case and electricity rates, and you know, the underlying reasons why, which have more to do with the utilities, data centers, water consumption, and then some structural decrease in Americans' electricity, right? It's like a 15-page deck, and he's one of these guys where he's listening, he'll scan it, he's scanning it while he listens to you, and then he goes right to the one thing. And I have a lot of respect for this because I've seen some of the most incredible survivors, the people that have that one skill. They can go in and go to that thing, that little detail, that thread and pull it. So my experience with him is quite impressive; when I anchor him to the other people I've been with, he's really up there."
1:34

Chamath describes Donald Trump: "I've seen some of the most incredible survivors, they have that one skill." "Oh, he's really inspiring. I mean, I have to say here that in my limited interactions with him, he's really curious. He's really smart. His intuition is off the charts. I'll tell you a very specific example. Had a meeting with him, I had a complicated-" "What were you pitching him?" "It's not a pitch. I was just walking him through some structural issues in this specific case and electricity rates, and you know, the underlying reasons why, which have more to do with the utilities, data centers, water consumption, and then some structural decrease in Americans' electricity, right? It's like a 15-page deck, and he's one of these guys where he's listening, he'll scan it, he's scanning it while he listens to you, and then he goes right to the one thing. And I have a lot of respect for this because I've seen some of the most incredible survivors, the people that have that one skill. They can go in and go to that thing, that little detail, that thread and pull it. So my experience with him is quite impressive; when I anchor him to the other people I've been with, he's really up there."

Tony Jacob | FindaClip.com

191,816 次观看 • 1 个月前

Vice President JD Vance was not mincing words in ending the White House press briefing by again ripping the press for their Minneapolis coverage... “And again, thank you all for — for listening. Thanks for taking questions. And I just ask you, look, this is politics and often Republicans and you know, get — get in — get in arguments with the press about things. And I understand that. I think it's really irresponsible for you guys to go out there and imply or tell the American people that a guy who defended himself from being rammed by an automobile is guilty of murder, be a little bit more careful. We want to talk about toning down the temperature, which I know the President wants to do, and I certainly want to do. One of the ways we tone down the temperature is to have a media that tells the truth. I encourage you all to do that. God bless you. [REPORTERS SHOUTING] Happy New Year, guys!”
0:46

Vice President JD Vance was not mincing words in ending the White House press briefing by again ripping the press for their Minneapolis coverage... “And again, thank you all for — for listening. Thanks for taking questions. And I just ask you, look, this is politics and often Republicans and you know, get — get in — get in arguments with the press about things. And I understand that. I think it's really irresponsible for you guys to go out there and imply or tell the American people that a guy who defended himself from being rammed by an automobile is guilty of murder, be a little bit more careful. We want to talk about toning down the temperature, which I know the President wants to do, and I certainly want to do. One of the ways we tone down the temperature is to have a media that tells the truth. I encourage you all to do that. God bless you. [REPORTERS SHOUTING] Happy New Year, guys!”

Curtis Houck

91,111 次观看 • 7 个月前