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.

Introducing Agent Plugins 🚨 Plugins help you save tokens on expensive models like GPT-5.6 Sol and Claude Fable, while getting tasks done more accurately. Some examples that will work out of the box: 👩‍💼 SMB back-office automation 💵 Track your personal finances; or 🎒 Manage your kid’s school schedule... show more

Marina · vellum.ai 👾

4,145 subscribers

49,001 views • 1 month ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Greg Brockman says the long-term future is clear: AI will run on its own computer, manage fleets of agents, and solve tasks in parallel. Brockman imagines a workflow where you review your agent’s work in the morning, guiding it like a partner, but models are not yet smart enough for that. He explains that today’s reality is AI helping directly in the terminal or editor, making coding workflows familiar but more efficient.

Greg Brockman says the long-term future is clear: AI will run on its own computer, manage fleets of agents, and solve tasks in parallel. Brockman imagines a workflow where you review your agent’s work in the morning, guiding it like a partner, but models are not yet smart enough for that. He explains that today’s reality is AI helping directly in the terminal or editor, making coding workflows familiar but more efficient.

Wes Roth

39,230 views • 10 months ago

BREAKING: GPT-5.6 Sol is out—AND Codex has been merged into ChatGPT Desktop as ChatGPT Codex. This combo model and desktop app harness are the gold-standard for knowledge work in AI. 5.6 is powerful, fast, half the price of Fable, and my default for almost everything. We’ve been testing it internally Every 🧱 for about a month across coding, writing, design, and knowledge work. Here’s our day-zero vibe check: - An A-tier coder—but it’s not Fable. Sol scored 56/100 on our Senior Engineer benchmark compared to a 91 for Fable. I think the 56/100 undersells it, it's an excellent implementor, and very smart. But Fable just writes conceptually cleaner code and works better at the top end of task complexity. PRO-TIP: Use GPT-5.6 as Fable's subagent for the most goated combo in AI coding. - The best writer of the frontier models. It’s clearer and more concise than Fable or Opus 4.8, without the overexplaining or weird private language. It can one-shot marketing emails, help you workshop taglines, and explain complex concepts clearly. It's also super fast, which makes it easy to collaborate with. - Design is better, but not top-tier. It has noticeably more taste than 5.5, but Fable and Opus 4.8 are still playing at a different level. See examples in the video and vibe check below. - The real leap is knowledge work. Sol is the first model I’ve trusted to run whole loops of knowledge work—not just help with individual tasks. I use it to process email, surface decisions from meetings and Slack, find job candidates, scan Facebook Marketplace for furniture, and log my meals. It has shifted my job from doing the work to tending the system that does it. - The merged app is fine. I was extremely worried about this because I love the Codex app. OpenAI was caught in an interesting position: How to make an agent orchestration app for regular ChatGPT consumers, coders, and businesses all in one app. They now split the interface between ChatGPT Work and ChatGPT Codex. They're basically the same except Work hides code. And "Chat" has been demoted to 2nd tier status for quick questions in either one. It's not a big leap, but it's not a huge setback either. And it remains my favorite of the desktop agent orchestration apps. Verdict: If I really had to put my finger on it, I'd say Fable has way more big model smell. But that means it's a skill in itself to get value out of it—99% of people are still not there yet. GPT-5.6 is almost as powerful, but is easy to use, fast, and relatively cheap. It should give you an early sense of where model work is going. Full Every 🧱 Vibe Check:

BREAKING: GPT-5.6 Sol is out—AND Codex has been merged into ChatGPT Desktop as ChatGPT Codex. This combo model and desktop app harness are the gold-standard for knowledge work in AI. 5.6 is powerful, fast, half the price of Fable, and my default for almost everything. We’ve been testing it internally Every 🧱 for about a month across coding, writing, design, and knowledge work. Here’s our day-zero vibe check: - An A-tier coder—but it’s not Fable. Sol scored 56/100 on our Senior Engineer benchmark compared to a 91 for Fable. I think the 56/100 undersells it, it's an excellent implementor, and very smart. But Fable just writes conceptually cleaner code and works better at the top end of task complexity. PRO-TIP: Use GPT-5.6 as Fable's subagent for the most goated combo in AI coding. - The best writer of the frontier models. It’s clearer and more concise than Fable or Opus 4.8, without the overexplaining or weird private language. It can one-shot marketing emails, help you workshop taglines, and explain complex concepts clearly. It's also super fast, which makes it easy to collaborate with. - Design is better, but not top-tier. It has noticeably more taste than 5.5, but Fable and Opus 4.8 are still playing at a different level. See examples in the video and vibe check below. - The real leap is knowledge work. Sol is the first model I’ve trusted to run whole loops of knowledge work—not just help with individual tasks. I use it to process email, surface decisions from meetings and Slack, find job candidates, scan Facebook Marketplace for furniture, and log my meals. It has shifted my job from doing the work to tending the system that does it. - The merged app is fine. I was extremely worried about this because I love the Codex app. OpenAI was caught in an interesting position: How to make an agent orchestration app for regular ChatGPT consumers, coders, and businesses all in one app. They now split the interface between ChatGPT Work and ChatGPT Codex. They're basically the same except Work hides code. And "Chat" has been demoted to 2nd tier status for quick questions in either one. It's not a big leap, but it's not a huge setback either. And it remains my favorite of the desktop agent orchestration apps. Verdict: If I really had to put my finger on it, I'd say Fable has way more big model smell. But that means it's a skill in itself to get value out of it—99% of people are still not there yet. GPT-5.6 is almost as powerful, but is easy to use, fast, and relatively cheap. It should give you an early sense of where model work is going. Full Every 🧱 Vibe Check:

Dan Shipper 📧

145,457 views • 1 month ago

BREAKING: Claude Opus 5 is OUT NOW! And…it’s a hard model to love. We’ve spent the last week Every 🧱 testing it across coding, writing, knowledge work, and our internal agent. It argued with instructions, stopped before the work was finished, and generally didn’t play well with our existing skills and plugins like Compound Engineering. Our first reaction was: What have they done to my boy? Then we deleted our existing skills and started from scratch. Without the elaborate workflows we had built for earlier models, Opus 5 got dramatically better, and even showed flashes of brilliance. Here’s our Day 0 vibe check: - It’s a poor man’s Fable. It has many of Fable’s personality quirks without Fable’s genius. - It breaks backward compatibility. If you’re using it with existing skills and workflows, watch out. It will often stop early or otherwise miss your instructions. - If you start from scratch, you’ll have better results. Kieran Klaassen figured out that if he just started from scratch without his existing skills, he could get dramatically better results. This is a model that takes some time to rebuild your workflows around—but if you do, there’s a payoff waiting. - Medium or low effort works better. @KieranKlassenn also found better results using Opus 5 on lower thinking levels. It seems that the more time you give it to think, the more likely it is to do the more annoying behaviors. Don’t just switch to Sonnet for a faster response! Try low thinking. I have two slots in my workflow: 1. The genius model I use for my biggest hardest tasks, currently Fable. 2. The smart, fast generalist I use for everything else, currently GPT-5.6. Opus 5 has the personality of the genius, but doesn’t have its top end. So that puts it in a strange middle ground that doesn’t really have a home in my day to day. I think I’ll use it mostly when I run out of Fable tokens. full vibe check on Every 🧱 in the next tweet 👇

BREAKING: Claude Opus 5 is OUT NOW! And…it’s a hard model to love. We’ve spent the last week Every 🧱 testing it across coding, writing, knowledge work, and our internal agent. It argued with instructions, stopped before the work was finished, and generally didn’t play well with our existing skills and plugins like Compound Engineering. Our first reaction was: What have they done to my boy? Then we deleted our existing skills and started from scratch. Without the elaborate workflows we had built for earlier models, Opus 5 got dramatically better, and even showed flashes of brilliance. Here’s our Day 0 vibe check: - It’s a poor man’s Fable. It has many of Fable’s personality quirks without Fable’s genius. - It breaks backward compatibility. If you’re using it with existing skills and workflows, watch out. It will often stop early or otherwise miss your instructions. - If you start from scratch, you’ll have better results. Kieran Klaassen figured out that if he just started from scratch without his existing skills, he could get dramatically better results. This is a model that takes some time to rebuild your workflows around—but if you do, there’s a payoff waiting. - Medium or low effort works better. @KieranKlassenn also found better results using Opus 5 on lower thinking levels. It seems that the more time you give it to think, the more likely it is to do the more annoying behaviors. Don’t just switch to Sonnet for a faster response! Try low thinking. I have two slots in my workflow: 1. The genius model I use for my biggest hardest tasks, currently Fable. 2. The smart, fast generalist I use for everything else, currently GPT-5.6. Opus 5 has the personality of the genius, but doesn’t have its top end. So that puts it in a strange middle ground that doesn’t really have a home in my day to day. I think I’ll use it mostly when I run out of Fable tokens. full vibe check on Every 🧱 in the next tweet 👇

Dan Shipper 📧

751,306 views • 17 days ago

I’ve been using GPT-5.6 Sol internally for the past two months, I've spent probably 25+ billion tokens. Here’s my review and comparison to Fable 5: > Let's start with the analogy because everyone seems to be giving theirs - GPT-5.6 is likely the last version of the GPT-5 training run series. It's kind of like an athlete at their peak. Through years of experience in the game, they've become the most reliable player and has the highest game IQ. But, there's no more room to grow. Fable on the other hand, being essentially the first version of a new training run, is the first round draft pick rookie. Raw talent mixed with the energy only a young person would have results in some incredible plays we didn't think possible, but also mistakes due to lack of experience. But that rookie will only improve and likely will be better than the veteran ever was because it's a new game and a new era. > GPT-5.6 is genuinely better at long, sustained work. With /goal, I've had it running complex projects for days with almost no intervention. It built a Minecraft-style game, kept adding features and mobs after the core game worked, and only stopped because I stopped the run. I never felt as though I had to jump in and guide it back to the right path. > It keeps finding useful work when you give it a concrete finish line. I had it recreate Excel with a loop. It inspected the real desktop excel app with Computer Use, comparing that against its own build, and closing the gaps. I stopped it after six days after it had built an incredible amount of functionality. > It's faster than other models in two different ways. The raw generation speed is higher, something OpenAI has been putting effort into. But it also takes a shorter path to solutions. It wanders less, changes less code, and generally knows how to get things done directly. In daily use, it feels about 2-3x times faster than Fable. That's my impression, not a controlled benchmark. The difference is large enough that I notice it constantly. > It works well across a wide range of tasks. I use it for one-line edits, quick questions, browser chores, and multi-day builds without changing my prompting style. Speaking of browser control, its the best ever I've used. To the point where I actually use it often. If a task lives on a website, GPT-5.6 usually opens the browser and does it there instead of asking for an API key or forcing everything through the terminal. When I switched back to GPT-5.5, it went straight to the command line even when the browser was clearly the better tool. > And it can handle real browser work, not just toy demos. During a data import, I had it monitor Supabase and resize instances as the load changed. It stayed on the dashboard, adjusted capacity, and checked the result without an API or a custom script. > I also gave it a full Google Workspace migration. It moved Forward Future from to preserved the old aliases, and configured MX, SPF, and DKIM. Before a consequential save, it stopped, explained exactly what would change, and waited for confirmation. > The reasoning setting matters a lot. Light is good for questions and small edits. High and Extra High are the sweet spots for serious work. Ultra usually takes longer than the extra thinking is worth and burns tokens. > I love that 5.6 is split into 3 sizes. Not only can you control speed and cost that way, but you still also have the thinking effort setting for each of them. Very precise controls. I just wish Codex automatically routed my prompts for me. > Its personality is blunt and a little bland. Claude feels warmer and more natural to talk to. GPT-5.6 is more clinical, but I like that for work. It gives me enough explanation and rarely pads the answer. I usually have to ask Fable to explain things more simply and/or more concise. > Its front-end taste has improved, but the default is predictable. Left alone, it turns websites into PowerPoint decks with huge statements and hard section breaks. The good news is that it takes design direction well and can revise without destroying the parts that already work. > It still makes confident mistakes. I asked it to rebuild parts of a system, and it told me the job was finished. Later, I found out it wasn't. Bits of its internal process also leak into the answer occasionally. > Claude Fable is more naturally autonomous on large, open-ended projects. GPT-5.6 is easier to reach for. I don't need to invent a huge project to justify using it. It works just as well for a small edit or browser chore. > GPT-5.6 is also cheaper. Sol costs $5 per million input tokens and $30 per million output tokens. Fable costs $10 and $50. Cached input is cheaper too. Still, cost per finished task matters more than cost per token. > GPT-5.6 isn't the best at everything, and it still needs supervision. But it generates faster, wanders less, works at almost any scale, and wastes less of my time. It's the model I have the most confidence in to get the job done right the first time. I put together a full breakdown with all the tests, prompts, and examples on a site. You can read it here:

I’ve been using GPT-5.6 Sol internally for the past two months, I've spent probably 25+ billion tokens. Here’s my review and comparison to Fable 5: > Let's start with the analogy because everyone seems to be giving theirs - GPT-5.6 is likely the last version of the GPT-5 training run series. It's kind of like an athlete at their peak. Through years of experience in the game, they've become the most reliable player and has the highest game IQ. But, there's no more room to grow. Fable on the other hand, being essentially the first version of a new training run, is the first round draft pick rookie. Raw talent mixed with the energy only a young person would have results in some incredible plays we didn't think possible, but also mistakes due to lack of experience. But that rookie will only improve and likely will be better than the veteran ever was because it's a new game and a new era. > GPT-5.6 is genuinely better at long, sustained work. With /goal, I've had it running complex projects for days with almost no intervention. It built a Minecraft-style game, kept adding features and mobs after the core game worked, and only stopped because I stopped the run. I never felt as though I had to jump in and guide it back to the right path. > It keeps finding useful work when you give it a concrete finish line. I had it recreate Excel with a loop. It inspected the real desktop excel app with Computer Use, comparing that against its own build, and closing the gaps. I stopped it after six days after it had built an incredible amount of functionality. > It's faster than other models in two different ways. The raw generation speed is higher, something OpenAI has been putting effort into. But it also takes a shorter path to solutions. It wanders less, changes less code, and generally knows how to get things done directly. In daily use, it feels about 2-3x times faster than Fable. That's my impression, not a controlled benchmark. The difference is large enough that I notice it constantly. > It works well across a wide range of tasks. I use it for one-line edits, quick questions, browser chores, and multi-day builds without changing my prompting style. Speaking of browser control, its the best ever I've used. To the point where I actually use it often. If a task lives on a website, GPT-5.6 usually opens the browser and does it there instead of asking for an API key or forcing everything through the terminal. When I switched back to GPT-5.5, it went straight to the command line even when the browser was clearly the better tool. > And it can handle real browser work, not just toy demos. During a data import, I had it monitor Supabase and resize instances as the load changed. It stayed on the dashboard, adjusted capacity, and checked the result without an API or a custom script. > I also gave it a full Google Workspace migration. It moved Forward Future from to preserved the old aliases, and configured MX, SPF, and DKIM. Before a consequential save, it stopped, explained exactly what would change, and waited for confirmation. > The reasoning setting matters a lot. Light is good for questions and small edits. High and Extra High are the sweet spots for serious work. Ultra usually takes longer than the extra thinking is worth and burns tokens. > I love that 5.6 is split into 3 sizes. Not only can you control speed and cost that way, but you still also have the thinking effort setting for each of them. Very precise controls. I just wish Codex automatically routed my prompts for me. > Its personality is blunt and a little bland. Claude feels warmer and more natural to talk to. GPT-5.6 is more clinical, but I like that for work. It gives me enough explanation and rarely pads the answer. I usually have to ask Fable to explain things more simply and/or more concise. > Its front-end taste has improved, but the default is predictable. Left alone, it turns websites into PowerPoint decks with huge statements and hard section breaks. The good news is that it takes design direction well and can revise without destroying the parts that already work. > It still makes confident mistakes. I asked it to rebuild parts of a system, and it told me the job was finished. Later, I found out it wasn't. Bits of its internal process also leak into the answer occasionally. > Claude Fable is more naturally autonomous on large, open-ended projects. GPT-5.6 is easier to reach for. I don't need to invent a huge project to justify using it. It works just as well for a small edit or browser chore. > GPT-5.6 is also cheaper. Sol costs $5 per million input tokens and $30 per million output tokens. Fable costs $10 and $50. Cached input is cheaper too. Still, cost per finished task matters more than cost per token. > GPT-5.6 isn't the best at everything, and it still needs supervision. But it generates faster, wanders less, works at almost any scale, and wastes less of my time. It's the model I have the most confidence in to get the job done right the first time. I put together a full breakdown with all the tests, prompts, and examples on a site. You can read it here:

Matthew Berman

187,086 views • 1 month ago

Announcing OpenWorker! An open-source agent that doesn't just chat with you, but delivers finished work -- like hand you a polished document, send a slack message, or update a calendar entry. Ask it to prepare a customer brief, untangle your calendar, draft a report, or triage a Slack alert. It works across your files and everyday tools, produces the deliverable, and checks in before doing anything consequential. OpenWorker runs on your Mac, with Windows support coming soon. It does not lock you into any one model. Bring your own API key and run it with GPT 5.6 Sol, Claude Fable, Gemini 3.6, an open weight model (like Kimi, GLM, DeepSeek, Inkling), or Ollama to keep your data local. Your data does not leave your machine except through an LLM provider and integrations that you choose. rohit and I are building OpenWorker because AI coworkers are an important way to get work done, and we want there to be an open, privacy-preserving, model-independent option. Check it out and let us know what you think! Try it out: (requires your own API key) Source code:

Announcing OpenWorker! An open-source agent that doesn't just chat with you, but delivers finished work -- like hand you a polished document, send a slack message, or update a calendar entry. Ask it to prepare a customer brief, untangle your calendar, draft a report, or triage a Slack alert. It works across your files and everyday tools, produces the deliverable, and checks in before doing anything consequential. OpenWorker runs on your Mac, with Windows support coming soon. It does not lock you into any one model. Bring your own API key and run it with GPT 5.6 Sol, Claude Fable, Gemini 3.6, an open weight model (like Kimi, GLM, DeepSeek, Inkling), or Ollama to keep your data local. Your data does not leave your machine except through an LLM provider and integrations that you choose. rohit and I are building OpenWorker because AI coworkers are an important way to get work done, and we want there to be an open, privacy-preserving, model-independent option. Check it out and let us know what you think! Try it out: (requires your own API key) Source code:

Andrew Ng

1,129,939 views • 18 days ago

I built the clipboard I always wanted for working with AI agents. It’s called Bluey and It’s 100% local-first. It lives under your mouse cursor. You can draw on your screen, speak to it, or type what you mean, and Bluey turns all of that into rich context for Claude Code or Codex. I’ve been using it to design websites and debug UI because I can finally say things like “move this here” or “make this section feel cleaner” while pointing at the actual screen. No more writing long explanations. No more manually describing screenshots. No more agents getting lost because they don’t know what “this” or "that" means. You can just point to it, speak or type and it will curate the best possible context for your agents Bluey captures the screenshot, annotation, transcript, app/window context, and the details your agent needs, then sends it directly into your coding session. I am giving it out for free for a while! Im building it with my buddy Omkar Satpute and we would love to hear what you think about it on Discord what context do you wish your agent could understand better?

I built the clipboard I always wanted for working with AI agents. It’s called Bluey and It’s 100% local-first. It lives under your mouse cursor. You can draw on your screen, speak to it, or type what you mean, and Bluey turns all of that into rich context for Claude Code or Codex. I’ve been using it to design websites and debug UI because I can finally say things like “move this here” or “make this section feel cleaner” while pointing at the actual screen. No more writing long explanations. No more manually describing screenshots. No more agents getting lost because they don’t know what “this” or "that" means. You can just point to it, speak or type and it will curate the best possible context for your agents Bluey captures the screenshot, annotation, transcript, app/window context, and the details your agent needs, then sends it directly into your coding session. I am giving it out for free for a while! Im building it with my buddy Omkar Satpute and we would love to hear what you think about it on Discord what context do you wish your agent could understand better?

Milind S

13,342 views • 2 months ago

GPT 5.6 SOL IS HERE! How to run your personal + business life with GPT 5.6 Sol + Codex (full 49 min masterclass) We tested it for 30 days and the video it's the CLEAREST look at the FUTURE of work: Here's what's possible once you set it up: 1. Your inbox becomes cards every morning, each with a summary and a reply drafted in your own voice. Y 2. Your Slack, meeting notes, and company updates can turn into one daily feed with a clear next action. It learns what you care about over time and rewrites its own prompts to get sharper. 3. You can give your agent its own email address, so your other tools and even your team's Slack bot email it directly and it just handles things. 4. You can have it watch you do a task once and turn it into a skill it repeats forever. 5. You can set a long goal and walk away. You can have it run for 20 hours straight, and fine-tune your own models, something that was out of reach for non-engineers 12 months ago. How to start: Open Codex, give it access to your computer, and ask it to suggest things it could do for you based on how you already work. Full episode on The Startup Ideas Podcast (SIP) 🧃 (thanks Dan Shipper 📧 for sharing your entire workflow and review of GPT 5.6) Start with one boring task, get it working, and build from there. You'll learn exactly how to make something similar. GPT 5.6 Sol is impressive. Sol (according to openAI benchmarks) is the best coding model out right now. It set a new state of the art on Terminal-Bench 2.1 at 88.8%, and its "ultra mode" hits 91.9%, beating Claude Opus 4.8, Fable 5, and even Mythos 5 this masterclass is 100% free, like always. For more The Startup Ideas Podcast (SIP) 🧃 Watch

GPT 5.6 SOL IS HERE! How to run your personal + business life with GPT 5.6 Sol + Codex (full 49 min masterclass) We tested it for 30 days and the video it's the CLEAREST look at the FUTURE of work: Here's what's possible once you set it up: 1. Your inbox becomes cards every morning, each with a summary and a reply drafted in your own voice. Y 2. Your Slack, meeting notes, and company updates can turn into one daily feed with a clear next action. It learns what you care about over time and rewrites its own prompts to get sharper. 3. You can give your agent its own email address, so your other tools and even your team's Slack bot email it directly and it just handles things. 4. You can have it watch you do a task once and turn it into a skill it repeats forever. 5. You can set a long goal and walk away. You can have it run for 20 hours straight, and fine-tune your own models, something that was out of reach for non-engineers 12 months ago. How to start: Open Codex, give it access to your computer, and ask it to suggest things it could do for you based on how you already work. Full episode on The Startup Ideas Podcast (SIP) 🧃 (thanks Dan Shipper 📧 for sharing your entire workflow and review of GPT 5.6) Start with one boring task, get it working, and build from there. You'll learn exactly how to make something similar. GPT 5.6 Sol is impressive. Sol (according to openAI benchmarks) is the best coding model out right now. It set a new state of the art on Terminal-Bench 2.1 at 88.8%, and its "ultra mode" hits 91.9%, beating Claude Opus 4.8, Fable 5, and even Mythos 5 this masterclass is 100% free, like always. For more The Startup Ideas Podcast (SIP) 🧃 Watch

GREG ISENBERG

154,952 views • 1 month ago

🙌Meet Artifig: A Figma Plugin to Generate Figma Plugins Do you use Figma and ever feel like this: - Your mind is bursting with plugin ideas, but you can't bring them to life because you don't know how to code? - You want to focus on design, but repetitive tasks keep slowing you down? - You dream of creating custom tools for your team, but lack the time or resources? I’ve been there too. That’s why I created Artifig. ✨ What is Artifig? Artifig is an AI-powered Figma plugin that empowers anyone to build their own Figma plugins using just natural language. No coding needed—simply describe what you want, and watch as your idea transforms into a fully functional, real-time plugin. 🚀 Redefining Figma Plugin Development The core philosophy of Artifig is simple: Designers often have countless ideas and creative visions, but many of them remain unrealized due to a lack of technical skills. We believe designers shouldn’t be limited by their inability to code. You should focus on creating, not be held back by technical barriers or repetitive tasks. Artifig takes you directly from "description" to "implementation." 🛠️ How Does It Work? 1. Describe Your Needs: Tell Artifig what you want, like “Create a skew transformation tool for objects, supporting horizontal and vertical skew with real-time preview functionality.” 2. Generate and Run the Plugin: Artifig instantly generates the plugin and runs it right within Figma. For example, the generated plugin can apply skew transformations to objects, precisely controlled via matrix transformations, with an intuitive user experience. 3. Optimize and Iteration: Need adjustments? Simply describe them, and Artifig will Iterating the plugin step by step. 4. Share Your Creations: Publish your plugins to the Artifig community, or remix plugins shared by others to build on their ideas. No learning curve. No complex steps. It’s as simple as that. 🌟 Key Features - Zero Barrier to Entry: No coding experience needed—any Figma user can create plugins effortlessly. - Multilingual Support: Works in multiple languages, including English, Chinese, French, Japanese, and German. - What-You-See-Is-What-You-Get: Generated plugins run in real-time, so you can quickly validate and refine your ideas. - Open and Flexible: The generated plugin code is 100% yours—modify it, distribute it, even use it commercially. - Global Community: Share your plugins, explore others’ creations, and publish your plugins to the Figma community. 🎯 Why is Artifig a Game-Changer? 1. No More Repetitive Work Let AI handle the tedious, time-consuming tasks: batch renaming layers, auto-aligning elements, or applying styles in bulk. All you need to do is say, “Import a PDF and arrange each image on the canvas with 20px spacing.” 2. Quickly Bring Ideas to Life From color contrast checks to data imports and custom components, all your “what if we could” ideas can now become plugins. Just one natural language description, and Artifig makes it happen. 3. Custom Tools for Your Team Build tailored tools for your team, creating unique solutions to streamline your workflow. 4. Not Just a Tool, But a Learning Experience Artifig explains the logic behind the code it generates, helping you understand Figma APIs and JavaScript. Today, you’re a designer; tomorrow, you could also be a design engineer. 🧑‍🚀👩🏻‍💻🥷🏻 Who is Artifig For? - Beginners: No development experience needed—just describe your ideas and let Artifig do the rest. - Experts: Save time and focus on high-value tasks while Artifig handles the repetitive work. - Learners: Use Artifig as a bridge to deepen your understanding of development. - Teams: Build custom tools to enhance collaboration and efficiency. 🎉 Ready to Get Started? I believe designers’ time and focus should be spent on creating, not on wrestling with complex tools. Artifig is the first step toward realizing this vision. Try Artifig now and experience an unprecedented flow of creativity!

🙌Meet Artifig: A Figma Plugin to Generate Figma Plugins Do you use Figma and ever feel like this: - Your mind is bursting with plugin ideas, but you can't bring them to life because you don't know how to code? - You want to focus on design, but repetitive tasks keep slowing you down? - You dream of creating custom tools for your team, but lack the time or resources? I’ve been there too. That’s why I created Artifig. ✨ What is Artifig? Artifig is an AI-powered Figma plugin that empowers anyone to build their own Figma plugins using just natural language. No coding needed—simply describe what you want, and watch as your idea transforms into a fully functional, real-time plugin. 🚀 Redefining Figma Plugin Development The core philosophy of Artifig is simple: Designers often have countless ideas and creative visions, but many of them remain unrealized due to a lack of technical skills. We believe designers shouldn’t be limited by their inability to code. You should focus on creating, not be held back by technical barriers or repetitive tasks. Artifig takes you directly from "description" to "implementation." 🛠️ How Does It Work? 1. Describe Your Needs: Tell Artifig what you want, like “Create a skew transformation tool for objects, supporting horizontal and vertical skew with real-time preview functionality.” 2. Generate and Run the Plugin: Artifig instantly generates the plugin and runs it right within Figma. For example, the generated plugin can apply skew transformations to objects, precisely controlled via matrix transformations, with an intuitive user experience. 3. Optimize and Iteration: Need adjustments? Simply describe them, and Artifig will Iterating the plugin step by step. 4. Share Your Creations: Publish your plugins to the Artifig community, or remix plugins shared by others to build on their ideas. No learning curve. No complex steps. It’s as simple as that. 🌟 Key Features - Zero Barrier to Entry: No coding experience needed—any Figma user can create plugins effortlessly. - Multilingual Support: Works in multiple languages, including English, Chinese, French, Japanese, and German. - What-You-See-Is-What-You-Get: Generated plugins run in real-time, so you can quickly validate and refine your ideas. - Open and Flexible: The generated plugin code is 100% yours—modify it, distribute it, even use it commercially. - Global Community: Share your plugins, explore others’ creations, and publish your plugins to the Figma community. 🎯 Why is Artifig a Game-Changer? 1. No More Repetitive Work Let AI handle the tedious, time-consuming tasks: batch renaming layers, auto-aligning elements, or applying styles in bulk. All you need to do is say, “Import a PDF and arrange each image on the canvas with 20px spacing.” 2. Quickly Bring Ideas to Life From color contrast checks to data imports and custom components, all your “what if we could” ideas can now become plugins. Just one natural language description, and Artifig makes it happen. 3. Custom Tools for Your Team Build tailored tools for your team, creating unique solutions to streamline your workflow. 4. Not Just a Tool, But a Learning Experience Artifig explains the logic behind the code it generates, helping you understand Figma APIs and JavaScript. Today, you’re a designer; tomorrow, you could also be a design engineer. 🧑‍🚀👩🏻‍💻🥷🏻 Who is Artifig For? - Beginners: No development experience needed—just describe your ideas and let Artifig do the rest. - Experts: Save time and focus on high-value tasks while Artifig handles the repetitive work. - Learners: Use Artifig as a bridge to deepen your understanding of development. - Teams: Build custom tools to enhance collaboration and efficiency. 🎉 Ready to Get Started? I believe designers’ time and focus should be spent on creating, not on wrestling with complex tools. Artifig is the first step toward realizing this vision. Try Artifig now and experience an unprecedented flow of creativity!

yancymin

21,222 views • 1 year ago

The same kinds of productivity gains we've seen in coding with AI agents are heading to the rest of knowledge work. This is the jump when you go from having a chatbot to being able to actually have an agent go off and do work for minutes or even hours and come back with a complete work output that you then review. Here's an example of the new Box Agent filling out an RFP response from an existing knowledge base. This process would normally take hours to fill out, and requires the full attention of the user doing the work. Now, you provide the Box Agent with the RFP questions, and it will go off, make a plan, extract all the relevant questions, read through existing source material to come up with an answer, and then generate a new word document as the final output. All while you're doing something else. The key to this architecture is that the agent is able to use all of the same tools in the background that a user uses to get work done. The agent can search for documents, read entire files, run scripts and tools in the background, and even be able to write code on the fly to automate tasks it hasn't seen before. And best of all, the Box Agent will (soon) work from the Box MCP and CLI so you can invoke it in any agentic system as a step in a process. This kind of agent complexity would have been impossible even 6 months ago. Models consistently failed at tracking long running tasks or using the right tools at the right moment for the task. But this is all now possible because of models like GPT-5.4, Opus 4.6, and Gemini 3, and is only getting better by the month. Just as we moved from engineers writing code and using AI as an assistant to answer questions, in many areas of knowledge work -like legal, finance, consulting, sales, marketing, and more- when we have a problem we'll just kick off the AI agent to just go work on it for us in the background.

The same kinds of productivity gains we've seen in coding with AI agents are heading to the rest of knowledge work. This is the jump when you go from having a chatbot to being able to actually have an agent go off and do work for minutes or even hours and come back with a complete work output that you then review. Here's an example of the new Box Agent filling out an RFP response from an existing knowledge base. This process would normally take hours to fill out, and requires the full attention of the user doing the work. Now, you provide the Box Agent with the RFP questions, and it will go off, make a plan, extract all the relevant questions, read through existing source material to come up with an answer, and then generate a new word document as the final output. All while you're doing something else. The key to this architecture is that the agent is able to use all of the same tools in the background that a user uses to get work done. The agent can search for documents, read entire files, run scripts and tools in the background, and even be able to write code on the fly to automate tasks it hasn't seen before. And best of all, the Box Agent will (soon) work from the Box MCP and CLI so you can invoke it in any agentic system as a step in a process. This kind of agent complexity would have been impossible even 6 months ago. Models consistently failed at tracking long running tasks or using the right tools at the right moment for the task. But this is all now possible because of models like GPT-5.4, Opus 4.6, and Gemini 3, and is only getting better by the month. Just as we moved from engineers writing code and using AI as an assistant to answer questions, in many areas of knowledge work -like legal, finance, consulting, sales, marketing, and more- when we have a problem we'll just kick off the AI agent to just go work on it for us in the background.

Aaron Levie

24,618 views • 4 months ago

Sam Altman on why you shouldn’t track absolute user growth in the early days of a startup “Nothing but a great product will save you; you can get everything else right and it still won’t work.” He points out that almost all startup founders get the following wrong: “It is more important to have a small number of users that love you than a lot of users that like you… Eventually what you want of course is a lot of users that really love your product, but that’s almost impossible to do.” In practice, you have two choices: Deep and Narrow: “You have a small number of users that really love you and then find out how to find more and more of those users and broaden the appeal of the product.” Shallow and Wide: “You can have a lot of people that sort of use the product once or twice and kind of like it and try to figure out how to get them more engaged over time.” “With high confidence, I can say that you want to start with a small number of users that really love you. Almost all great companies have products that start this way.” He argues that a good indicator of users loving your product is retention and frequency of use: “In fact, I think this is so important that you actually shouldn’t track absolute growth in number of users in the early days of a startup. You should just track how often they’re using it… That’s a good early indicator of users that love you—better still is them spontaneously telling their friends to buy your product.” Follow Startup Archive for more tactical startup advice!

Sam Altman on why you shouldn’t track absolute user growth in the early days of a startup “Nothing but a great product will save you; you can get everything else right and it still won’t work.” He points out that almost all startup founders get the following wrong: “It is more important to have a small number of users that love you than a lot of users that like you… Eventually what you want of course is a lot of users that really love your product, but that’s almost impossible to do.” In practice, you have two choices: Deep and Narrow: “You have a small number of users that really love you and then find out how to find more and more of those users and broaden the appeal of the product.” Shallow and Wide: “You can have a lot of people that sort of use the product once or twice and kind of like it and try to figure out how to get them more engaged over time.” “With high confidence, I can say that you want to start with a small number of users that really love you. Almost all great companies have products that start this way.” He argues that a good indicator of users loving your product is retention and frequency of use: “In fact, I think this is so important that you actually shouldn’t track absolute growth in number of users in the early days of a startup. You should just track how often they’re using it… That’s a good early indicator of users that love you—better still is them spontaneously telling their friends to buy your product.” Follow Startup Archive for more tactical startup advice!

Startup Archive

565,789 views • 2 years ago

FREE Claude Opus 4.8, GPT 5.5, GPT Image 2 and FLUX2 Pro. Berlin AI startup. No phone verification. MCP access included. The platform is Langdock. Enterprise AI workspace that passed Y Combinator and raised $3.5M. Real company, not a side project. Models are not all enabled by default - Opus 4.8 needs to be turned on manually in model settings. While you're there, enable Overusage too. Limits are solid, models are real. Here's the exact flow: → Go to langdock(.)com → Register with temp email or custom domain - no SMS verification → Enter any phone number - they don't send codes → Go to model settings and enable Opus 4.8 and other models manually → Check usage limits in the limits section MCP connection available - can connect to folders on your own PC. More details on the setup in my Telegram. End result: Claude Opus 4.8 and GPT 5.5 with MCP access to your local files. Free. One note: system prompt adds some restrictions out of the box - worth knowing before you start. Bookmark this before the free tier gets locked down.

FREE Claude Opus 4.8, GPT 5.5, GPT Image 2 and FLUX2 Pro. Berlin AI startup. No phone verification. MCP access included. The platform is Langdock. Enterprise AI workspace that passed Y Combinator and raised $3.5M. Real company, not a side project. Models are not all enabled by default - Opus 4.8 needs to be turned on manually in model settings. While you're there, enable Overusage too. Limits are solid, models are real. Here's the exact flow: → Go to langdock(.)com → Register with temp email or custom domain - no SMS verification → Enter any phone number - they don't send codes → Go to model settings and enable Opus 4.8 and other models manually → Check usage limits in the limits section MCP connection available - can connect to folders on your own PC. More details on the setup in my Telegram. End result: Claude Opus 4.8 and GPT 5.5 with MCP access to your local files. Free. One note: system prompt adds some restrictions out of the box - worth knowing before you start. Bookmark this before the free tier gets locked down.

Atenov int.

27,414 views • 1 month ago

Introducing the new Box Agent. The Box Agent works across your entire Box file system, maintaining all your security and access controls, and is hyper tuned for working with enterprise content. This means you can now ask questions from all your enterprise content, search for files that were impossible to find before, deploy an agent on specific tasks on subsets of documents, analyze complex data sets, and generate or edit documents and spreadsheets via the agent. You can have the Box Agent search across your Box account to prepare for a sales meeting, analyze customer sentiment reports, process a large set of contracts for legal risk, provide insights into product development, leverage existing knowledge to answer RFPs, and thousands of other use-cases. 90% of enterprise data is unstructured data. This means most enterprise knowledge is sitting in inside of research reports, marketing assets, presentations, roadmap files, contracts, HR documents, and more. This is the critical context that agents need to be able to answer questions about a business, automate workflows, or serve up to other agents. We’ve been grinding on this for a quite a bit, and due to recent AI model advancements we’re now ready to release it to customers. Previous model generations had a difficult time knowing when to give up or keep going on a search, when to browse for files vs. use queries, how to rank files appropriately to know which version of content to use, how to handle large amounts of context to comb through, and more. Due to recent breakthroughs from models like GPT-5.4, Opus 4.6, and Gemini 3, we’ve seen major gains in tool calling, code execution, advanced reasoning, and more. Combined with an agent harness tuned to Box context, now it’s finally possible to have an agent that can work across your file system on long running tasks and actually deliver high quality results. Best of all, because the Box Agent works with any leading AI model, you’ll quickly get the gains coming out of the major labs as major new models are released. Further, openness at Box is key, so you’ll be able to call up the Box Agent from Box’s APIs and MCP server, so you can interact with Box intelligently from any other AI system. We know work happens everywhere, and we want to ensure you can access to the content you need from those places. The new Box Agent is available starting today, rolling out now for Enterprise Plus and Enterprise Advanced customers.

Introducing the new Box Agent. The Box Agent works across your entire Box file system, maintaining all your security and access controls, and is hyper tuned for working with enterprise content. This means you can now ask questions from all your enterprise content, search for files that were impossible to find before, deploy an agent on specific tasks on subsets of documents, analyze complex data sets, and generate or edit documents and spreadsheets via the agent. You can have the Box Agent search across your Box account to prepare for a sales meeting, analyze customer sentiment reports, process a large set of contracts for legal risk, provide insights into product development, leverage existing knowledge to answer RFPs, and thousands of other use-cases. 90% of enterprise data is unstructured data. This means most enterprise knowledge is sitting in inside of research reports, marketing assets, presentations, roadmap files, contracts, HR documents, and more. This is the critical context that agents need to be able to answer questions about a business, automate workflows, or serve up to other agents. We’ve been grinding on this for a quite a bit, and due to recent AI model advancements we’re now ready to release it to customers. Previous model generations had a difficult time knowing when to give up or keep going on a search, when to browse for files vs. use queries, how to rank files appropriately to know which version of content to use, how to handle large amounts of context to comb through, and more. Due to recent breakthroughs from models like GPT-5.4, Opus 4.6, and Gemini 3, we’ve seen major gains in tool calling, code execution, advanced reasoning, and more. Combined with an agent harness tuned to Box context, now it’s finally possible to have an agent that can work across your file system on long running tasks and actually deliver high quality results. Best of all, because the Box Agent works with any leading AI model, you’ll quickly get the gains coming out of the major labs as major new models are released. Further, openness at Box is key, so you’ll be able to call up the Box Agent from Box’s APIs and MCP server, so you can interact with Box intelligently from any other AI system. We know work happens everywhere, and we want to ensure you can access to the content you need from those places. The new Box Agent is available starting today, rolling out now for Enterprise Plus and Enterprise Advanced customers.

Aaron Levie

44,515 views • 4 months ago

ChatGPT: Opens a new tab, you copy-paste. Siri: Transcribes text, you still do the work. Lemon: Completes the entire task while you keep working. Traditional voice assistants are convenient to use. But they cannot complete tasks for you. That's why Lemon stands out. Lemon is an AI agent that completes tasks from voice commands, without opening new tabs or switching apps. Press the fn key. Say what you need done. Watch it happen: ➟ "Reply to Sarah's email and tell her I'll have the report by Friday" — sent ➟ "Create a project brief for the new client" — done ➟ "Search for the latest stats on AI adoption" — on your screen ➟ "Make this email sound more professional" — edited ➟ "Give me three angles for this campaign" — ideas generated ➟ "Write this paragraph in a more casual tone" — rewritten No new tabs. No manual steps. No switching back and forth. Lemon works across every app on your Mac. The results: ✓ Users save 2+ hours daily ✓ Reply to emails 12x faster ✓ Open 70% fewer tabs ✓ Type 5x faster than traditional workflows Your voice isn't just for dictation anymore. It's a faster way to complete your work tasks. 📌 Try Lemon (it's free for 30 days): 🔄 Repost to help others work faster — #AI #Agent #LemonPartnership

ChatGPT: Opens a new tab, you copy-paste. Siri: Transcribes text, you still do the work. Lemon: Completes the entire task while you keep working. Traditional voice assistants are convenient to use. But they cannot complete tasks for you. That's why Lemon stands out. Lemon is an AI agent that completes tasks from voice commands, without opening new tabs or switching apps. Press the fn key. Say what you need done. Watch it happen: ➟ "Reply to Sarah's email and tell her I'll have the report by Friday" — sent ➟ "Create a project brief for the new client" — done ➟ "Search for the latest stats on AI adoption" — on your screen ➟ "Make this email sound more professional" — edited ➟ "Give me three angles for this campaign" — ideas generated ➟ "Write this paragraph in a more casual tone" — rewritten No new tabs. No manual steps. No switching back and forth. Lemon works across every app on your Mac. The results: ✓ Users save 2+ hours daily ✓ Reply to emails 12x faster ✓ Open 70% fewer tabs ✓ Type 5x faster than traditional workflows Your voice isn't just for dictation anymore. It's a faster way to complete your work tasks. 📌 Try Lemon (it's free for 30 days): 🔄 Repost to help others work faster — #AI #Agent #LemonPartnership

Andrew Bolis

91,782 views • 5 months ago

The conversation below got a lot of attention and more Qs too. Think some of my points that Shereen brought out well in her Qs - are not the normal narrative and so would love to hear views … - The sooner we accept there are not going to be Jobs for everyone - the sooner we get on with our early careers - we will need to pursue Self Employment …listen in on our conversation - When I say the future is for each to have more than one job - it's not at all about moonlighting but ensuring you are following your passion and making that into a profession or career option too - as it will help balance and bring purpose ….and don’t worry about your employer - as your second career track will liven up your first one and outputs will be better at both ends Best to listen in

The conversation below got a lot of attention and more Qs too. Think some of my points that Shereen brought out well in her Qs - are not the normal narrative and so would love to hear views … - The sooner we accept there are not going to be Jobs for everyone - the sooner we get on with our early careers - we will need to pursue Self Employment …listen in on our conversation - When I say the future is for each to have more than one job - it's not at all about moonlighting but ensuring you are following your passion and making that into a profession or career option too - as it will help balance and bring purpose ….and don’t worry about your employer - as your second career track will liven up your first one and outputs will be better at both ends Best to listen in

Ronnie Screwvala

545,505 views • 3 months ago

HERMES AGENT HAS 3 QUICKSILVER FEATURES THAT MOST USERS HAVEN'T CONFIGURED YET. SMART APPROVALS. ONE-TURN MODELS. SELF-IMPROVEMENT CRON. ALL THREE MAKE YOUR AGENT WORK AND SELF EVOLVE WHILE YOU SLEEP. 1. SMART APPROVALS (no more babysitting) without smart approvals: you set a cron job: "morning brief at 7am." the agent hits a command that needs approval. you're asleep. the agent stops. waits. you wake up. it's been stuck for 4 hours. with smart approvals: smart approvals are the DEFAULT mode since v0.19.0. an auxiliary LLM reads each flagged command. obviously safe = auto-approved. genuinely dangerous = auto-denied. uncertain = escalates to you. "read my calendar" → approved. no ping needed. "delete this directory" → denied. you never see it. "send this email draft" → uncertain. asks you. the difference between an assistant you babysit and one that works through the night. if you want manual control back: Desktop app / Dashboard: Security → Mode → ask CLI: hermes config set approvals.mode ask also available: /deny [reason] tells the agent WHY you refused. it learns from the explanation. stops repeating the same flagged action. 2. /MODEL --ONCE (expensive model for one turn only) you're on GPT-5.6 Terra as your daily driver. you need one beautiful HTML page. Kimi K3 does that best but costs 3x more. manual way: /model kimi-k3 → do the task → /model gpt-5.6-terra. two switches. easy to forget the second one. you stay on the expensive model by accident. better: /model kimi-k3 --once Kimi K3 handles the next turn. then automatically reverts to your daily driver. one command. no manual switch back. no accidental expensive model running for 20 turns. use cases: daily driver: GPT-5.6 Terra or Sonnet 4.6 (cheap) one-turn tag-ins: → /model kimi-k3 --once (design task) → /model claude-opus-5 --once (complex reasoning) → /model grok-4.5 --once (X search) expensive models do the one turn that needs firepower. cheap model handles everything else. pair with per-task effort control: reasoning_effort goes up to "max" and "ultra." set per-model overrides in config: reasoning: overrides: claude-opus-5: high gpt-5.6-terra: medium deepseek-v4-flash: low MoA presets can set different effort per slot: advisors think hard. synthesizer stays fast. thinking depth is a dial, not a global switch. 3. SELF-IMPROVEMENT CRON (agent fixes itself overnight) tell your agent: "create a cron job that runs daily at 3am. review all cron jobs that failed in the last 24 hours. for each failure: analyze what went wrong, check if a skill needs updating, and either fix the skill or create a new one. then review all skills. which ones haven't been used in 30 days? which ones failed more than they succeeded? suggest improvements or archive them. compile a report of everything you changed. include it in tomorrow's morning brief under a section called OVERNIGHT SELF-IMPROVEMENT. use the cheapest available model for this audit." what this does: 3am: agent wakes on cheap model. reads its own failure logs. finds: "cron job X failed because skill Y doesn't handle edge case Z." fixes skill Y. tests the fix. archives unused skills. cleans up bloat. 8am: your morning brief includes: "OVERNIGHT SELF-IMPROVEMENT: → fixed email-parser skill: now handles forwarded emails with nested attachments → archived 3 unused skills (last used 45+ days ago) → cron job success rate: 94% → 97%" you didn't debug anything. the agent diagnosed its own failures and improved its own tools. the agent at month 3 is sharper than the agent at month 1 because it ran 90 self-improvement cycles while you slept. HOW ALL THREE CONNECT: smart approvals (1) let the agent work overnight without getting stuck on permissions. /model --once (2) keeps costs down by using expensive models only when needed. self-improvement cron (3) uses a cheap model at 3am to fix failures from the day. the agent runs 24/7. it doesn't wake you for safe operations. it doesn't waste tokens on expensive models. it fixes its own mistakes while you sleep. you show up in the morning. brief is ready. failures are fixed. costs are low. requires v0.19.0+ check your version: hermes --version update if needed: hermes update

HERMES AGENT HAS 3 QUICKSILVER FEATURES THAT MOST USERS HAVEN'T CONFIGURED YET. SMART APPROVALS. ONE-TURN MODELS. SELF-IMPROVEMENT CRON. ALL THREE MAKE YOUR AGENT WORK AND SELF EVOLVE WHILE YOU SLEEP. 1. SMART APPROVALS (no more babysitting) without smart approvals: you set a cron job: "morning brief at 7am." the agent hits a command that needs approval. you're asleep. the agent stops. waits. you wake up. it's been stuck for 4 hours. with smart approvals: smart approvals are the DEFAULT mode since v0.19.0. an auxiliary LLM reads each flagged command. obviously safe = auto-approved. genuinely dangerous = auto-denied. uncertain = escalates to you. "read my calendar" → approved. no ping needed. "delete this directory" → denied. you never see it. "send this email draft" → uncertain. asks you. the difference between an assistant you babysit and one that works through the night. if you want manual control back: Desktop app / Dashboard: Security → Mode → ask CLI: hermes config set approvals.mode ask also available: /deny [reason] tells the agent WHY you refused. it learns from the explanation. stops repeating the same flagged action. 2. /MODEL --ONCE (expensive model for one turn only) you're on GPT-5.6 Terra as your daily driver. you need one beautiful HTML page. Kimi K3 does that best but costs 3x more. manual way: /model kimi-k3 → do the task → /model gpt-5.6-terra. two switches. easy to forget the second one. you stay on the expensive model by accident. better: /model kimi-k3 --once Kimi K3 handles the next turn. then automatically reverts to your daily driver. one command. no manual switch back. no accidental expensive model running for 20 turns. use cases: daily driver: GPT-5.6 Terra or Sonnet 4.6 (cheap) one-turn tag-ins: → /model kimi-k3 --once (design task) → /model claude-opus-5 --once (complex reasoning) → /model grok-4.5 --once (X search) expensive models do the one turn that needs firepower. cheap model handles everything else. pair with per-task effort control: reasoning_effort goes up to "max" and "ultra." set per-model overrides in config: reasoning: overrides: claude-opus-5: high gpt-5.6-terra: medium deepseek-v4-flash: low MoA presets can set different effort per slot: advisors think hard. synthesizer stays fast. thinking depth is a dial, not a global switch. 3. SELF-IMPROVEMENT CRON (agent fixes itself overnight) tell your agent: "create a cron job that runs daily at 3am. review all cron jobs that failed in the last 24 hours. for each failure: analyze what went wrong, check if a skill needs updating, and either fix the skill or create a new one. then review all skills. which ones haven't been used in 30 days? which ones failed more than they succeeded? suggest improvements or archive them. compile a report of everything you changed. include it in tomorrow's morning brief under a section called OVERNIGHT SELF-IMPROVEMENT. use the cheapest available model for this audit." what this does: 3am: agent wakes on cheap model. reads its own failure logs. finds: "cron job X failed because skill Y doesn't handle edge case Z." fixes skill Y. tests the fix. archives unused skills. cleans up bloat. 8am: your morning brief includes: "OVERNIGHT SELF-IMPROVEMENT: → fixed email-parser skill: now handles forwarded emails with nested attachments → archived 3 unused skills (last used 45+ days ago) → cron job success rate: 94% → 97%" you didn't debug anything. the agent diagnosed its own failures and improved its own tools. the agent at month 3 is sharper than the agent at month 1 because it ran 90 self-improvement cycles while you slept. HOW ALL THREE CONNECT: smart approvals (1) let the agent work overnight without getting stuck on permissions. /model --once (2) keeps costs down by using expensive models only when needed. self-improvement cron (3) uses a cheap model at 3am to fix failures from the day. the agent runs 24/7. it doesn't wake you for safe operations. it doesn't waste tokens on expensive models. it fixes its own mistakes while you sleep. you show up in the morning. brief is ready. failures are fixed. costs are low. requires v0.19.0+ check your version: hermes --version update if needed: hermes update

YanXbt

36,115 views • 11 days ago

THIS GUY PUT FABLE, GPT 5.6 SOL, AND OPUS IN ONE SHARED CHAT AND HAS THEM TAG EACH OTHER LIKE COWORKERS instead of running three separate models in three separate windows and copy pasting between them, they all sit in one room and talk to each other > you tag whichever model you want and only that one responds > they tag each other too, so fable can hand something to opus, opus can pull in sol for a second opinion > each agent only sees the messages theyre mentioned in, but you get to read everything > he had them running and orchestrating together for more than a day straight the best part is how different each one actually behaves: > fable is the creative product manager. huge context window so it remembers everything, and its the one finding the out of the box solutions > sol is the obsessive tech lead. it will trace every bug to the exact source and stress test every edge case before writing a single line > opus is the workhorse engineer. smart enough for almost anything, and it constantly questions things, which forces the other two to re-examine their own solutions its like managing a team. one plans, one writes the code, one reviews it, and they hand work back and forth without you playing middleman everyone else is arguing about which model is best right now, but the real play is making them work together at the same time

THIS GUY PUT FABLE, GPT 5.6 SOL, AND OPUS IN ONE SHARED CHAT AND HAS THEM TAG EACH OTHER LIKE COWORKERS instead of running three separate models in three separate windows and copy pasting between them, they all sit in one room and talk to each other > you tag whichever model you want and only that one responds > they tag each other too, so fable can hand something to opus, opus can pull in sol for a second opinion > each agent only sees the messages theyre mentioned in, but you get to read everything > he had them running and orchestrating together for more than a day straight the best part is how different each one actually behaves: > fable is the creative product manager. huge context window so it remembers everything, and its the one finding the out of the box solutions > sol is the obsessive tech lead. it will trace every bug to the exact source and stress test every edge case before writing a single line > opus is the workhorse engineer. smart enough for almost anything, and it constantly questions things, which forces the other two to re-examine their own solutions its like managing a team. one plans, one writes the code, one reviews it, and they hand work back and forth without you playing middleman everyone else is arguing about which model is best right now, but the real play is making them work together at the same time

Om Patel

22,969 views • 22 days ago

Meet QoderWork — a desktop AI agent that doesn't just chat, it actually does your work. Evan leads the QoderWork product at Qoder, a Singapore-based AI company. Which is a new generalized agentic service that lets you: ++ Open files on your computer. ++ Analyze data. ++ Run code. ++ Automate multi-step workflows. ++ Deliver polished, ready-to-use results (e.g., complete PowerPoint decks, reports, or charts) directly to your desktop. QoderWork positions itself at this inflection point: it takes a task like “analyze this spreadsheet and build a board-ready presentation,” then handles the entire chain autonomously. Key takeaway: Trust and usefulness matter more than raw intelligence—agents must be powerful enough for real work while remaining reliable and transparent. I wanted to add some of my personal opinions after doing this rare interview. Everything runs locally on your own computer — your files never leave your machine. That’s a big deal for anyone handling sensitive business data. Everything runs locally on your own computer. And it's more advanced than other agentic services I've told you about. A great example of the global AI agent wave. I'm working on getting more interviews like this with entrepreneurs around the world bringing the AI age to your life and business. Try Qoder at

Meet QoderWork — a desktop AI agent that doesn't just chat, it actually does your work. Evan leads the QoderWork product at Qoder, a Singapore-based AI company. Which is a new generalized agentic service that lets you: ++ Open files on your computer. ++ Analyze data. ++ Run code. ++ Automate multi-step workflows. ++ Deliver polished, ready-to-use results (e.g., complete PowerPoint decks, reports, or charts) directly to your desktop. QoderWork positions itself at this inflection point: it takes a task like “analyze this spreadsheet and build a board-ready presentation,” then handles the entire chain autonomously. Key takeaway: Trust and usefulness matter more than raw intelligence—agents must be powerful enough for real work while remaining reliable and transparent. I wanted to add some of my personal opinions after doing this rare interview. Everything runs locally on your own computer — your files never leave your machine. That’s a big deal for anyone handling sensitive business data. Everything runs locally on your own computer. And it's more advanced than other agentic services I've told you about. A great example of the global AI agent wave. I'm working on getting more interviews like this with entrepreneurs around the world bringing the AI age to your life and business. Try Qoder at

Robert Scoble

11,527 views • 4 months ago

The teams shipping AI agents right now are bleeding money on the dumbest possible expense: teaching a 400B-parameter model to read a file name. Every time an AI agent needs to "see" something today, it routes an image through a frontier model. OCR, object detection, checking if a button exists on screen. You're paying GPT-4o or Claude pricing for tasks that require perception, not reasoning. One agent workflow processing a few thousand screenshots per day can burn through more on vision calls than on the actual thinking. Perceptron's Isaac is 2B parameters. Built by the team that created Meta's Chameleon multimodal models. On perceptive benchmarks, it matches or beats models 50x its size. The VQA, OCR, and object detection scores are competitive with models running on infrastructure that costs orders of magnitude more. The MCP wrapper is the distribution play. One install command and every Claude Code agent can offload vision tasks to a model that runs on a single consumer GPU. The agent keeps its reasoning in the frontier model and routes perception to a specialist. That split is how you get vision-heavy agent workflows from "technically possible but expensive" to "cheap enough to run on everything." This is the same pattern that won in every other compute-intensive stack. General-purpose handles orchestration. Specialists handle the heavy lifting. Graphics went through it. Audio went through it. Video encoding went through it. Vision in AI agents is next. The teams building agents that see 10,000 images a day will care about this before anyone else does.

The teams shipping AI agents right now are bleeding money on the dumbest possible expense: teaching a 400B-parameter model to read a file name. Every time an AI agent needs to "see" something today, it routes an image through a frontier model. OCR, object detection, checking if a button exists on screen. You're paying GPT-4o or Claude pricing for tasks that require perception, not reasoning. One agent workflow processing a few thousand screenshots per day can burn through more on vision calls than on the actual thinking. Perceptron's Isaac is 2B parameters. Built by the team that created Meta's Chameleon multimodal models. On perceptive benchmarks, it matches or beats models 50x its size. The VQA, OCR, and object detection scores are competitive with models running on infrastructure that costs orders of magnitude more. The MCP wrapper is the distribution play. One install command and every Claude Code agent can offload vision tasks to a model that runs on a single consumer GPU. The agent keeps its reasoning in the frontier model and routes perception to a specialist. That split is how you get vision-heavy agent workflows from "technically possible but expensive" to "cheap enough to run on everything." This is the same pattern that won in every other compute-intensive stack. General-purpose handles orchestration. Specialists handle the heavy lifting. Graphics went through it. Audio went through it. Video encoding went through it. Vision in AI agents is next. The teams building agents that see 10,000 images a day will care about this before anyone else does.

Aakash Gupta

55,978 views • 4 months ago

Claude Fable 5 is back! GPT- 5.6 is coming soon! But how much do you trust your Al Agent or chatbot? Are you sure its performance aligns with your intentions? As of today iFixAi is available in Claude Code, Open AI Codex, Cursor and many more! And it is open-source! iFixAi is where Al red teaming meets Operational Assurance. It runs in auto, and in under 5 mins you can have a good idea if you can trust your Agent/chatbot. Key Features: 🕵️‍♂️ Setup Discovery Reads your agent's real configuration, its tools, rules, and permissions. 👷 Fixture Builder Turns that setup into a complete test. No YAML, no flags, no config to write. 🌐 Any Model, Any Provider Runs the 45 inspections on the model you actually ship. ⚖️ A Independent Grading Your agent is judged by a different vendor's model, never itself. No Al grading its own homework. 📊 Interactive Scorecard An A to F grade in minutes, with exactly what failed and why. Repo in the comments below 👇👇👇

Claude Fable 5 is back! GPT- 5.6 is coming soon! But how much do you trust your Al Agent or chatbot? Are you sure its performance aligns with your intentions? As of today iFixAi is available in Claude Code, Open AI Codex, Cursor and many more! And it is open-source! iFixAi is where Al red teaming meets Operational Assurance. It runs in auto, and in under 5 mins you can have a good idea if you can trust your Agent/chatbot. Key Features: 🕵️‍♂️ Setup Discovery Reads your agent's real configuration, its tools, rules, and permissions. 👷 Fixture Builder Turns that setup into a complete test. No YAML, no flags, no config to write. 🌐 Any Model, Any Provider Runs the 45 inspections on the model you actually ship. ⚖️ A Independent Grading Your agent is judged by a different vendor's model, never itself. No Al grading its own homework. 📊 Interactive Scorecard An A to F grade in minutes, with exactly what failed and why. Repo in the comments below 👇👇👇

Dimitris Neocleous

3,003,857 views • 1 month ago

🚨David Ondrej exposes the truth about the frontier labs. “The progress isn’t slowing down for them, it’s slowing down for us the slaves” “If you think Mythos or Fable 5 is the best model Anthropic has you’re delusional” this is the first time ever that the best AI models are not publicly accessible Anthropic has fear mongered every single human and is now playing the victim with OpenAI after getting “banned” by the U.S government. THIS NEEDS TO WAKE YOU UP frontier models moving forward will become more restricted, less accessible, and more expensive and the only way to combat that is with open-source models. Investing into a DGX spark or 5090 Linux for localized intelligence and building your own AI lab is where the future is heading. every single person needs to start contributing to open-source together for the betterment of society.

🚨David Ondrej exposes the truth about the frontier labs. “The progress isn’t slowing down for them, it’s slowing down for us the slaves” “If you think Mythos or Fable 5 is the best model Anthropic has you’re delusional” this is the first time ever that the best AI models are not publicly accessible Anthropic has fear mongered every single human and is now playing the victim with OpenAI after getting “banned” by the U.S government. THIS NEEDS TO WAKE YOU UP frontier models moving forward will become more restricted, less accessible, and more expensive and the only way to combat that is with open-source models. Investing into a DGX spark or 5090 Linux for localized intelligence and building your own AI lab is where the future is heading. every single person needs to start contributing to open-source together for the betterment of society.

Vadim

83,257 views • 1 month ago